ruvnet-brain 4.4.1 → 4.5.0

package/plugin/scripts/package-cards.json

@@ -0,0 +1,797 @@
+{"schema":"ruvnet-brain.package-cards/1","tokenizer":"pkgrec-tok/1","derivedFrom":{"builtUtc":"2026-10-01T11:01:11.071Z","releaseTag":"v4.4.0","manifestPassages":1703,"readmePassages":2623,"stores":175},"grounding":"Every field is copied from a manifest passage of a public store; see source + sourceSha256. t/s are derived scoring tokens.","cards":[
+{"id":"@agentic-flow/byzantine-quic","family":"byzantine-quic","kind":"npm","store":"agentic-flow","version":"1.0.0","description":"Byzantine fault-tolerant consensus protocol over QUIC transport for real-time systems","keywords":["byzantine","consensus","distributed-systems","fault-tolerance","pbft","quic"],"source":"agentic-flow/packages/integrations/byzantine-quic/package.json","sourceSha256":"e9e56dc3347c92932fd885d9f9e621789e8271440e5046a15153516d51599fb8","variants":[],"readme":"Production-ready implementation of PBFT (Practical Byzantine Fault Tolerance) consensus protocol with QUIC transport for real-time distributed systems. - Byzantine Fault Tolerance: Tolerates up to f malicious nodes in 3f+1 configuration - Ultra-Low Latency: <10ms consensus latency (p95) with QUIC transport - High Throughput: 1000+ operations/second - Cryptographic Security: Ed25519 signatures on all messages - View Changes: Automatic primary election and failover - Checkpointing: Periodic stable checkpoints for garbage collection - Production Ready: Comprehensive tests, metrics, and monitoring For f Byzantine faults, need 3f+1 total nodes: - f=1: 4 nodes (tolerates 1 malicious) - f=2: 7 node","readmeSource":"agentic-flow/packages/integrations/byzantine-quic/README.md","product":"key:agentic-flow:byzantine-quic","canonical":"@agentic-flow/byzantine-quic","tier":"T1","t":["byzantine-quic","byzantine","quic","fault-tolerant","fault","tolerant","consensus","protocol","transport","real","distributed-systems","distribut","fault-tolerance","tolerance","pbft"],"s":["agentic-flow","flow","byzantine-quic","byzantine","quic","consensus","distributed-systems","distribut","fault-tolerance","fault","tolerance","pbft"]},
+{"id":"@agentic-flow/crdt-gossip","family":"crdt-gossip","kind":"npm","store":"agentic-flow","version":"1.0.0","description":"Conflict-free Replicated Data Types (CRDTs) with Gossip Protocol for decentralized state synchronization","keywords":["conflict-free","crdt","decentralized","distributed","eventual-consistency","gossip","replication"],"source":"agentic-flow/packages/integrations/crdt-gossip/package.json","sourceSha256":"b3222b3e0a1494f36866a523cbd9ef5ba6c7c7ae772fa735af5e6284dff41997","variants":[],"readme":"A production-ready implementation of Conflict-free Replicated Data Types (CRDTs) with epidemic gossip protocol for truly decentralized, eventually consistent applications. - 5 CRDT Types: G-Counter, PN-Counter, LWW-Set, OR-Set, RGA - Gossip Protocol: Push-pull hybrid with anti-entropy - Phi-Accrual Failure Detection: Adaptive failure detection (Hayashibara et al., 2004) - Vector Clocks: Causal ordering and happens-before relationships - O(log N) Message Complexity: Efficient state dissemination - <100ms Convergence: Fast convergence for 1000 nodes - Strong Eventual Consistency: All replicas eventually converge Conflict-free Replicated Data Types (CRDTs) are data structures that can be replic","readmeSource":"agentic-flow/packages/integrations/crdt-gossip/README.md","product":"key:agentic-flow:crdt-gossip","canonical":"@agentic-flow/crdt-gossip","tier":"T1","t":["crdt-gossip","crdt","gossip","conflict-free","conflict","free","replicat","type","protocol","decentraliz","state","synchronization","distribut","eventual-consistency","eventual","consistency","replication"],"s":["agentic-flow","flow","crdt-gossip","crdt","gossip","conflict-free","conflict","free","decentraliz","distribut","eventual-consistency","eventual","consistency","replication"]},
+{"id":"@agentic-flow/ephemeral-memory","family":"ephemeral-memory","kind":"npm","store":"agentic-flow","version":"1.0.0","description":"Ephemeral Agents + Persistent Memory - On-demand agent spawning with AgentDB memory persistence. Achieve 90%+ resource savings with <50ms spawn times.","keywords":["agent-lifecycle","agent-memory","agentdb","cost-savings","ephemeral-agents","federation","on-demand-spawning","persistent-memory","resource-optimization","vector-search"],"source":"agentic-flow/packages/integrations/ephemeral-memory/package.json","sourceSha256":"2030e44490bda00552bfcb02fb178e9a712256f048f41104b6a6ad2b08f029ed","variants":[],"readme":"Ephemeral Memory combines short-lived agents (ephemeral) with long-lived storage (persistent memory) to dramatically reduce infrastructure costs while maintaining continuity across agent lifecycles. - 90%+ Resource Savings: Only pay for agents when they're working - <50ms Spawn Time: Lightning-fast agent creation - 10K+ Spawns/Second: Scale to massive workloads - Persistent Context: Memory survives agent termination - Automatic Cleanup: No resource leaks - Semantic Search: Find relevant past memories Main orchestrator for the ephemeral agent system. Persistent storage with semantic search. Manages agent TTL and automatic cleanup.","readmeSource":"agentic-flow/packages/integrations/ephemeral-memory/README.md","product":"key:agentic-flow:ephemeral-memory","canonical":"@agentic-flow/ephemeral-memory","tier":"T1","t":["ephemeral-memory","ephemeral","memory","persistent","on-demand","demand","spawn","agentdb","persistence","achieve","resource","saving","50ms","agent-lifecycle","lifecycle","agent-memory","cost-savings","cost","ephemeral-agents","federation","on-demand-spawning","persistent-memory","resource-optimization","vector-search","vector","search"],"s":["agentic-flow","flow","ephemeral-memory","ephemeral","memory","agent-lifecycle","lifecycle","agent-memory","agentdb","cost-savings","cost","saving","ephemeral-agents","federation","on-demand-spawning","demand","spawn","persistent-memory","persistent","resource-optimization","resource","vector-search","vector","search"]},
+{"id":"@agentic-flow/integrations-shared","family":"integrations-shared","kind":"npm","store":"agentic-flow","version":"1.0.0","description":"Shared bridges package for agentic-flow integrations connecting agent-booster, reasoningbank, agentdb, and QUIC components","keywords":["agent-booster","agentdb","agentic-flow","bridges","integration","quic","reasoningbank"],"source":"agentic-flow/packages/integrations/shared/package.json","sourceSha256":"91f02433174c7d72fdcf64cc3105e154c585df30e2ebedf5b50bcd76fd7c37c2","variants":[],"readme":"TypeScript bridges connecting agent-booster, reasoningbank, agentdb, and QUIC components for seamless integration in agentic-flow patterns. - AgentBoosterBridge: Ultra-fast code editing (52x faster than Morph LLM) - ReasoningBankBridge: Reinforcement learning with 9 RL algorithms - AgentDBBridge: Frontier memory features with 150x faster vector search - QuicBridge: High-performance networking with connection pooling","readmeSource":"agentic-flow/packages/integrations/shared/README.md","product":"key:agentic-flow:integrations-shared","canonical":"@agentic-flow/integrations-shared","tier":"T1","t":["integrations-shared","integration","shar","bridge","agentic-flow","flow","connect","agent-booster","booster","reasoningbank","agentdb","quic","component"],"s":["agentic-flow","flow","integrations-shared","integration","shar","agent-booster","booster","agentdb","bridge","quic","reasoningbank"]},
+{"id":"@agentic-flow/self-improving-codegen","family":"self-improving-codegen","kind":"npm","store":"agentic-flow","version":"0.1.0","description":"Self-improving code generation system combining AgentBooster (352x faster) with ReasoningBank (adaptive learning)","keywords":["adaptive-learning","agent-booster","agentdb","ai-agents","code-generation","code-improvement","pattern-recognition","reasoningbank","self-learning","trajectory-learning"],"source":"agentic-flow/packages/integrations/self-improving-codegen/package.json","sourceSha256":"fb9afbc3198b78aa9fb70da7c6dfb20ad2332047f8b4e58bad1ebd65212f7e1e","variants":[],"readme":"Generate code 352x faster than traditional LLMs while continuously learning from experience to improve over time. - ⚡ Ultra-Fast Generation: Sub-millisecond code edits via AgentBooster (352x faster than LLM APIs) - 🧠 Adaptive Learning: Learns from successful/failed trajectories using ReasoningBank - 📈 Continuous Improvement: Gets better with every generation (+20% quality after 100 trajectories) - 🔍 Pattern Recognition: Automatically extracts reusable patterns from successful code - 📊 Quality Analytics: Comprehensive metrics for complexity, maintainability, security - 🗄️ Persistent Memory: Stores trajectories in AgentDB for fast similarity search (150x faster) Main orchestrator class th","readmeSource":"agentic-flow/packages/integrations/self-improving-codegen/README.md","product":"key:agentic-flow:self-improving-codegen","canonical":"@agentic-flow/self-improving-codegen","tier":"T1","t":["self-improving-codegen","self","improv","codegen","self-improving","code","generation","combin","agentbooster","352x","reasoningbank","adaptive","learn","adaptive-learning","agent-booster","booster","agentdb","ai-agents","code-generation","code-improvement","improvement","pattern-recognition","pattern","recognition","self-learning","trajectory-learning","trajectory"],"s":["agentic-flow","flow","self-improving-codegen","self","improv","codegen","adaptive-learning","adaptive","learn","agent-booster","booster","agentdb","ai-agents","code-generation","code","generation","code-improvement","improvement","pattern-recognition","pattern","recognition","reasoningbank","self-learning","trajectory-learning","trajectory"]},
+{"id":"@agentic-qe/opencode-bridge","family":"opencode-bridge","kind":"npm","store":"agentic-qe","version":"0.1.0","description":"Bridge between AQE MCP server and OpenCode — config templates, output compaction, tool mapping","keywords":["agentic-qe","bridge","mcp","opencode"],"source":"agentic-qe/packages/aqe-opencode-bridge/package.json","sourceSha256":"7ae2b0c3026a54876bfdff31155b120b4cbf3ee6c3440ba78292adf3a25d587a","variants":[],"product":"key:agentic-qe:opencode-bridge","canonical":"@agentic-qe/opencode-bridge","tier":"T1","t":["opencode-bridge","opencode","bridge","between","aqe","mcp","server","config","template","output","compaction","mapp","agentic-qe"],"s":["agentic-qe","opencode-bridge","opencode","bridge","mcp"]},
+{"id":"@agentic-qe/opencode-plugin","family":"opencode-plugin","kind":"npm","store":"agentic-qe","version":"0.1.0","description":"AQE plugin for OpenCode — lifecycle hooks, agent routing, and skill execution","keywords":[],"source":"agentic-qe/packages/aqe-opencode-plugin/package.json","sourceSha256":"c5e705503648f47b4e267e669d027bc0827e258bf5a00a82220d07a5ac7b78aa","variants":[],"product":"key:agentic-qe:opencode-plugin","canonical":"@agentic-qe/opencode-plugin","tier":"T1","t":["opencode-plugin","opencode","plugin","aqe","lifecycle","hook","rout","skill","execution"],"s":["agentic-qe","opencode-plugin","opencode","plugin"]},
+{"id":"@agentic-qe/opencode-types","family":"opencode","kind":"npm","store":"agentic-qe","version":"0.1.0","description":"Shared TypeScript types for AQE-OpenCode integration","keywords":["agentic-qe","mcp","opencode","types"],"source":"agentic-qe/packages/aqe-opencode-types/package.json","sourceSha256":"de9448c0fff74d765fc551ed7ce7081ead89038b877f3d8dca63fb2035d73f30","variants":[],"product":"key:agentic-qe:opencode","canonical":"@agentic-qe/opencode-types","tier":"T1","t":["opencode-types","opencode","type","shar","aqe-opencode","aqe","integration","agentic-qe","mcp"],"s":["agentic-qe","opencode-types","opencode","type","mcp"]},
+{"id":"@agentic-robotics/cli","family":"cli","kind":"npm","store":"agentic-robotics","version":"0.2.3","description":"CLI tools for agentic robotics framework","keywords":["cli","robotics","ros"],"source":"agentic-robotics/npm/cli/package.json","sourceSha256":"fa8d2a88ab56985f8314615c07e8ba09ebb1565557efd9309e2aca1afa64970b","variants":[],"readme":"Command-line tools for the agentic-robotics framework. - 🛠️ Test Framework - Validate node creation and communication - 📊 System Info - Display framework version and capabilities - ⚡ Quick Testing - Rapid prototyping and debugging - 🔍 Diagnostics - Built-in health checks - 🎯 Simple Interface - Easy-to-use commands Global installation (recommended): Or use with npx: Test node creation, publisher, and message publishing: Run comprehensive system diagnostics to check your environment: Verbose mode: Output: Enter interactive dialog mode to work with the robotics framework: Interactive commands: List available AI agents with optional filtering: Filter by category: Output:","readmeSource":"agentic-robotics/npm/cli/README.md","product":"repo:agentic-robotics","canonical":"agentic-robotics","tier":"T2","t":["robotic","ros"],"s":["agentic-robotics","robotic","ros"]},
+{"id":"@agentic-robotics/core","family":"core","kind":"npm","store":"agentic-robotics","version":"0.2.1","description":"High-performance agentic robotics framework - Core bindings","keywords":["agents","middleware","napi-rs","robotics","ros","ros2"],"source":"agentic-robotics/npm/core/package.json","sourceSha256":"9f3417147fa6615cc13a3591da8d38ce95da31a9d78afbcf6ee053a7ef16d853","variants":[],"readme":"Core native bindings for the agentic-robotics framework - High-performance robotics middleware with ROS2 compatibility. - 🚀 High Performance - Native Rust implementation via NAPI-RS - 🤖 ROS2 Compatible - Full ROS2 message compatibility - 📡 Pub/Sub Pattern - Efficient publisher/subscriber communication - 🔄 Real-time Capable - Low-latency message passing - 📊 Statistics Tracking - Built-in performance metrics - 🌐 Cross-platform - Linux (x64, ARM64), macOS (x64, ARM64) - 💪 Type-safe - Full TypeScript support - ⚡ Zero-copy - Efficient memory management Platform-specific binaries are automatically installed based on your system. Main class for creating publishers and subscribers. Creates a","readmeSource":"agentic-robotics/npm/core/README.md","product":"repo:agentic-robotics","canonical":"agentic-robotics","tier":"T2","t":["core","high","robotic","middleware","ros","ros2"],"s":["agentic-robotics","robotic","core","middleware","ros","ros2"]},
+{"id":"@agentic-robotics/mcp","family":"mcp","kind":"npm","store":"agentic-robotics","version":"0.2.2","description":"Model Context Protocol server for agentic robotics with AgentDB integration","keywords":["agentdb","agentic-flow","mcp","model-context-protocol","robotics"],"source":"agentic-robotics/npm/mcp/package.json","sourceSha256":"80075abf650ce7fa74909d2fa9d1de4ba736345a7a307f170e8a668711d2bc73","variants":[],"readme":"Model Context Protocol (MCP) server for agentic robotics with AgentDB memory and agentic-flow orchestration. - 🤖 21 MCP Tools - Complete AI-robot integration toolkit - 🧠 AgentDB Memory - 13,000x faster than CLI (reflexion, skills, causal reasoning) - 🌊 Agentic Flow - 66 AI agents + 213 MCP tools orchestration - ⚡ High Performance - 5,725 ops/sec storage throughput - 🔄 Multi-Robot Swarm - Coordinate multiple robots intelligently - 📊 Performance Tracking - Built-in metrics and benchmarking - 🎯 Production Ready - Optimized hybrid SQL implementation Or add to your project: Add to your Claude config (~/Library/Application Support/Claude/claude_desktop_config.json): 1. move_robot - Move robo","readmeSource":"agentic-robotics/npm/mcp/README.md","product":"key:agentic-robotics:mcp","canonical":"@agentic-robotics/mcp","tier":"T2","t":["mcp","model","context","protocol","server","robotic","agentdb","integration","agentic-flow","flow","model-context-protocol"],"s":["agentic-robotics","robotic","mcp","agentdb","agentic-flow","flow","model-context-protocol","model","context","protocol"]},
+{"id":"@aqe/qe-browser-scripts","family":"qe-browser-scripts","kind":"npm","store":"agentic-qe","version":"1.0.0","description":"Scoped package.json that keeps qe-browser helper scripts in CommonJS despite the repo root being ESM.","keywords":[],"source":"agentic-qe/assets/skills/qe-browser/scripts/package.json","sourceSha256":"ddac12de68119a59a84e35ecee67f2b6a14d9101bfde7556790125923989386f","variants":[],"product":"key:agentic-qe:qe-browser-scripts","canonical":"@aqe/qe-browser-scripts","tier":"T1","t":["qe-browser-scripts","browser","script","scop","package.json","qe-browser","helper","commonj","despite","repo","root","esm"],"s":["aqe","qe-browser-scripts","browser","script"]},
+{"id":"@claude-flow/aidefence","family":"aidefence","kind":"npm","store":"ruflo","version":"3.0.2","description":"AI Manipulation Defense System (AIMDS) with self-learning, prompt injection detection, and vector search integration","keywords":["ai-security","aimds","claude-flow","jailbreak-detection","llm-security","pii-detection","prompt-injection","self-learning","threat-detection","vector-search"],"source":"ruflo/v3/@claude-flow/aidefence/package.json","sourceSha256":"7f79843b4450354934d57c5ddf9e39a9f024315a1bfd89b93b0c096fc634ad24","variants":[],"readme":"AI Manipulation Defense System (AIMDS) - Protect your AI applications from prompt injection, jailbreak attempts, and sensitive data exposure with sub-millisecond detection. - [Introduction](#introduction) - [Features](#features) - [Installation](#installation) - [Quick Start](#quick-start) - [API Reference](#api-reference) - [Threat Types](#threat-types) - [PII Detection](#pii-detection) - [Self-Learning](#self-learning) - [CLI Integration](#cli-integration) - [MCP Tools](#mcp-tools) - [Performance](#performance) - [Advanced Usage](#advanced-usage) - [Contributing](#contributing) - [License](#license) @claude-flow/aidefence is a high-performance security library designed to protect AI/LLM ap","readmeSource":"ruflo/v3/@claude-flow/aidefence/README.md","product":"key:aidefence","canonical":"@claude-flow/aidefence","tier":"T0","t":["aidefence","manipulation","defense","aimd","self-learning","self","learn","prompt","injection","detection","vector","search","integration","ai-security","security","claude-flow","claude","flow","jailbreak-detection","jailbreak","llm-security","pii-detection","pii","prompt-injection","threat-detection","threat","vector-search"],"s":["claude-flow","claude","flow","aidefence","ai-security","security","aimd","jailbreak-detection","jailbreak","detection","llm-security","pii-detection","pii","prompt-injection","prompt","injection","self-learning","self","learn","threat-detection","threat","vector-search","vector","search"]},
+{"id":"@claude-flow/browser","family":"browser","kind":"npm","store":"ruflo","version":"3.0.0","description":"Browser automation for AI agents - integrates agent-browser with claude-flow swarms","keywords":["ai-agent","automation","browser","claude-flow","playwright","testing","web-scraping"],"source":"ruflo/v3/@claude-flow/browser/package.json","sourceSha256":"9d74b476bae44c9aa77b6bee54e0ba66a5e3a5c3a22a75cec49d68991f6bdaf6","variants":[],"readme":"AI-Optimized Browser Automation for Claude-Flow Swarms *Integrate [agent-browser](https://github.com/AugmentCode/agent-browser) with claude-flow for intelligent web automation, trajectory learning, security scanning, and multi-agent browser coordination.* @claude-flow/browser provides a comprehensive browser automation layer for AI agents, combining Vercel Labs' agent-browser CLI with claude-flow's learning, memory, and security capabilities. It enables agents to navigate websites, fill forms, extract data, and learn from successful interaction patterns. 🌐 59 MCP Browser Tools - Complete browser automation via MCP protocol with element refs, snapshots, forms, navigation, and JavaScript exec","readmeSource":"ruflo/v3/@claude-flow/browser/README.md","product":"key:ruflo:browser","canonical":"@claude-flow/browser","tier":"T0","t":["browser","automation","integrate","agent-browser","claude-flow","claude","flow","swarm","ai-agent","playwright","test","web-scraping","web","scrap"],"s":["claude-flow","claude","flow","browser","ai-agent","automation","playwright","test","web-scraping","web","scrap"]},
+{"id":"@claude-flow/claims","family":"claims","kind":"npm","store":"ruflo","version":"3.0.0","description":"Issue claiming and work coordination module for Claude Flow V3","keywords":["claims","claude-flow","issue-tracking","mcp","multi-agent","swarm","work-coordination"],"source":"ruflo/v3/@claude-flow/claims/package.json","sourceSha256":"ca215845dbe0b049c44ff9a398c2b485259faca13ca6d07f6fd4de4e138d5119","variants":[],"readme":"Issue claiming and work coordination for human-agent collaboration. This package provides a complete system for coordinating work between humans and AI agents: - Issue Claiming - Claim issues to work on, preventing conflicts - Handoffs - Transfer work between humans and agents seamlessly - Work Stealing - Automatically reassign stalled work - Load Balancing - Distribute work evenly across agents Or install via Claude Flow CLI: When work stalls, issues can be marked as stealable:","readmeSource":"ruflo/v3/@claude-flow/claims/README.md","product":"key:ruflo:claims","canonical":"@claude-flow/claims","tier":"T0","t":["claim","issue","coordination","claude","flow","claude-flow","issue-tracking","track","mcp","multi-agent","multi","swarm","work-coordination"],"s":["claude-flow","claude","flow","claim","issue-tracking","issue","track","mcp","multi-agent","multi","swarm","work-coordination","coordination"]},
+{"id":"@claude-flow/cli","family":"cli","kind":"npm","store":"ruflo","version":"3.49.0","description":"Ruflo CLI - Enterprise AI agent orchestration with 60+ specialized agents, swarm coordination, MCP server, self-learning hooks, and vector memory for Claude Code","keywords":["ai-agents","anthropic","automation","claude","claude-code","claude-flow","cli","coding-assistant","core","developer-tools","embeddings","enterprise","hooks","lazy-load","lite","llm","mcp","memory","model-context-protocol","multi-agent","orchestration","self-learning","swarm","vector-database"],"source":"ruflo/v3/@claude-flow/cli/package.json","sourceSha256":"a86600091d9c0bebd50689808a0671784b35cf056a791bac6d27711160c02672","variants":["@claude-flow/cli-core"],"readme":"An agent meta-harness for Claude Code and Codex. [📖 RuFlo Explained — Build an AI Team That Plans, Remembers, Tests, and Improves](docs/ruflo-explained.md) A 14-chapter guide: from the basic idea to a first useful task, then memory, agent teams, plugins, cost and verification.","readmeSource":"ruflo/v3/@claude-flow/cli/README.md","product":"repo:ruflo","canonical":"ruflo","tier":"T0","t":["ruflo","enterprise","orchestration","specializ","swarm","coordination","mcp","server","self-learning","self","learn","hook","vector","memory","claude","code","ai-agents","anthropic","automation","claude-code","claude-flow","flow","coding-assistant","coding","assistant","core","developer-tools","developer","embedding","lazy-load","lazy","load","lite","model-context-protocol","model","context","protocol","multi-agent","multi","vector-database","database"],"s":["claude-flow","claude","flow","ai-agents","anthropic","automation","claude-code","code","coding-assistant","coding","assistant","core","developer-tools","developer","embedding","enterprise","hook","lazy-load","lazy","load","lite","mcp","memory","model-context-protocol","model","context","protocol","multi-agent","multi","orchestration","self-learning","self","learn","swarm","vector-database","vector","database"]},
+{"id":"@claude-flow/codex","family":"codex","kind":"npm","store":"ruflo","version":"3.0.3","description":"Codex CLI integration for Ruflo (claude-flow) - OpenAI Codex platform adapter","keywords":["agentic-ai","agents","agents.md","claude-flow","codex","coflow","openai","skills"],"source":"ruflo/v3/@claude-flow/codex/package.json","sourceSha256":"e42d89395cf68ac93355bea87e60af20f385a30c3e67d78f2e3f205ca4d15832","variants":[],"readme":"Transform OpenAI Codex CLI into a self-improving AI development system. While Codex executes code, claude-flow orchestrates, coordinates, and learns from every interaction. Codex does the work. Claude-flow coordinates and learns. That's it! The MCP server is auto-registered, skills are installed, and your project is ready for self-learning development. When you run init --codex, the MCP server is automatically registered with Codex: If MCP is not present, add manually:","readmeSource":"ruflo/v3/@claude-flow/codex/README.md","product":"key:ruflo:codex","canonical":"@claude-flow/codex","tier":"T0","t":["codex","integration","ruflo","claude-flow","claude","flow","openai","adapter","agentic-ai","agents.md","coflow","skill"],"s":["claude-flow","claude","flow","codex","agentic-ai","agents.md","coflow","openai","skill"]},
+{"id":"@claude-flow/deployment","family":"deployment","kind":"npm","store":"ruflo","version":"3.0.0","description":"Deployment module - Release management, CI/CD, versioning","keywords":[],"source":"ruflo/v3/@claude-flow/deployment/package.json","sourceSha256":"b29ca37d74dced74ea17e4340d8786811027dd4b5b9035581f983c6d0f150ddf","variants":[],"readme":"- Version Bumping: Automatic version management (major, minor, patch, prerelease) - Changelog Generation: Generate changelogs from conventional commits - Git Integration: Automatic tagging and committing - NPM Publishing: Publish packages with tag support (alpha, beta, latest) - Pre-Release Validation: Lint, test, build, and dependency checks - Dry Run Mode: Test releases without making changes Generates changelog from conventional commits: Generated changelog:","readmeSource":"ruflo/v3/@claude-flow/deployment/README.md","product":"key:ruflo:deployment","canonical":"@claude-flow/deployment","tier":"T0","t":["deployment","release","management","version"],"s":["claude-flow","claude","flow","deployment"]},
+{"id":"@claude-flow/embeddings","family":"embeddings","kind":"npm","store":"ruflo","version":"3.0.0-alpha.18","description":"V3 Embedding Service - OpenAI, Transformers.js, Agentic-Flow (ONNX), Mock providers with hyperbolic embeddings, normalization, and chunking","keywords":["chunking","claude-flow","embeddings","hyperbolic","neural-substrate","normalization","openai","poincare","similarity","transformers","v3","vector"],"source":"ruflo/v3/@claude-flow/embeddings/package.json","sourceSha256":"98c6ae3ade4ee83f3c9b944e158561191d3134a94c57c30cf7559f367e86071c","variants":[],"readme":"- Multiple Providers - Agentic-Flow (ONNX), OpenAI, Transformers.js, and Mock - Auto-Install - Automatically installs agentic-flow when using provider: 'auto' - Smart Fallback - Graceful fallback chain: agentic-flow → transformers → mock - LRU + Disk Caching - In-memory LRU + SQLite persistent cache with TTL - Batch Processing - Efficient batch embedding with partial cache hits - Similarity Functions - Cosine, Euclidean, and dot product metrics - 75x Faster - Agentic-flow ONNX is 75x faster than Transformers.js - Document Chunking - Character, sentence, paragraph, and token-based chunking with overlap - Multiple Normalization - L2, L1, min-max, and z-score normalization - Hyperbolic Embeddin","readmeSource":"ruflo/v3/@claude-flow/embeddings/README.md","product":"key:ruflo:embeddings","canonical":"@claude-flow/embeddings","tier":"T0","t":["embedding","embedd","service","openai","transformers.j","agentic-flow","flow","onnx","mock","provider","hyperbolic","normalization","chunk","claude-flow","claude","neural-substrate","neural","substrate","poincare","similarity","transformer","vector"],"s":["claude-flow","claude","flow","embedding","chunk","hyperbolic","neural-substrate","neural","substrate","normalization","openai","poincare","similarity","transformer","vector"]},
+{"id":"@claude-flow/guidance","family":"guidance","kind":"npm","store":"ruflo","version":"3.0.0","description":"Guidance Control Plane - Compiles, retrieves, enforces, and evolves guidance rules for Claude Code sessions","keywords":["adversarial-defense","agent-governance","agent-orchestration","ai-agent","ai-governance","ai-safety","anthropic","autonomous-agent","claude","claude-code","claude-flow","claude-md","claude.md","code-generation","control-plane","enforcement-gates","guidance","llm-governance","long-horizon","mcp","memory-governance","model-context-protocol","multi-agent","policy-enforcement","prompt-injection","proof-chain","ruv","ruvbot","trust-system","wasm"],"source":"ruflo/v3/@claude-flow/guidance/package.json","sourceSha256":"070fcc39ca10fb6f86b82cc1e4b35b831da3af1a2105d068917ecb6d820bdbdd","variants":[],"readme":"Long-horizon governance for Claude Code agents. AI coding agents are powerful for short tasks, but they break down over long sessions. They forget rules, repeat mistakes, run in circles, corrupt their own memory, and eventually need a human to step in. The longer the session, the worse it gets. @claude-flow/guidance fixes this. It takes the memory files Claude Code already uses — CLAUDE.md and CLAUDE.local.md — and turns them into a structured control plane that compiles rules, enforces them through gates the agent cannot bypass, proves every decision cryptographically, and evolves the rule set over time based on what actually works. The result: agents that can operate for days instead of mi","readmeSource":"ruflo/v3/@claude-flow/guidance/README.md","product":"key:ruflo:guidance","canonical":"@claude-flow/guidance","tier":"T0","t":["guidance","control","plane","compile","retrieve","enforce","evolve","rule","claude","code","session","adversarial-defense","adversarial","defense","agent-governance","governance","agent-orchestration","orchestration","ai-agent","ai-governance","ai-safety","safety","anthropic","autonomous-agent","autonomous","claude-code","claude-flow","flow","claude-md","claude.md","code-generation","generation","control-plane","enforcement-gates","enforcement","gate","llm-governance","long-horizon","long","horizon","mcp","memory-governance","memory","model-context-protocol","model","context","protocol","multi-agent","multi","policy-enforcement","policy","prompt-injection","prompt","injection","proof-chain","proof","chain","ruvbot","trust-system","trust"],"s":["claude-flow","claude","flow","guidance","adversarial-defense","adversarial","defense","agent-governance","governance","agent-orchestration","orchestration","ai-agent","ai-governance","ai-safety","safety","anthropic","autonomous-agent","autonomous","claude-code","code","claude-md","claude.md","code-generation","generation","control-plane","control","plane","enforcement-gates","enforcement","gate","llm-governance","long-horizon","long","horizon","mcp","memory-governance","memory","model-context-protocol","model","context","protocol","multi-agent","multi","policy-enforcement","policy","prompt-injection","prompt","injection","proof-chain","proof","chain","ruvbot","trust-system","trust"]},
+{"id":"@claude-flow/hooks","family":"hooks","kind":"npm","store":"ruflo","version":"3.0.0","description":"V3 Hooks System - Event-driven lifecycle hooks with ReasoningBank learning integration","keywords":["agents","claude-flow","daemons","hooks","mcp","reasoningbank","statusline"],"source":"ruflo/v3/@claude-flow/hooks/package.json","sourceSha256":"bdd72539ce5ec22d40bd7dfbf800821075215a9d629eabc3432ad5be45421148","variants":[],"readme":"The @claude-flow/hooks package provides a comprehensive hooks system for intercepting and extending Claude Flow operations. It enables intelligent task routing, pattern learning, background metrics collection, and real-time statusline integration. - 🎣 Hook Registry - Priority-based hook registration with filtering and management - ⚡ Hook Executor - Timeout handling, error recovery, and result aggregation - 🤖 Background Daemons - Metrics collection, swarm monitoring, pattern learning - 👷 Background Workers - 12 specialized workers for analysis, optimization, and automation - 📊 Statusline Integration - Real-time status display for Claude Code - 🧠 ReasoningBank Learning - Intelligent task","readmeSource":"ruflo/v3/@claude-flow/hooks/README.md","product":"key:ruflo:hooks","canonical":"@claude-flow/hooks","tier":"T0","t":["hook","event-driven","event","driven","lifecycle","reasoningbank","learn","integration","claude-flow","claude","flow","daemon","mcp","statusline"],"s":["claude-flow","claude","flow","hook","daemon","mcp","reasoningbank","statusline"]},
+{"id":"@claude-flow/integration","family":"integration","kind":"npm","store":"ruflo","version":"3.0.0","description":"Integration module - agentic-flow@alpha deep integration, ADR-001 compliance, TokenOptimizer","keywords":["agent-booster","agentic-flow","anti-drift","claude-flow","reasoningbank","swarm","token-optimizer"],"source":"ruflo/v3/@claude-flow/integration/package.json","sourceSha256":"d4763ef05f84816b9c214c5d5d845149f47bf703f063fc1653dcf06f7e05e4cc","variants":[],"readme":"- ADR-001 Compliance - Eliminates 10,000+ duplicate lines by building on agentic-flow - SONA Adapter - Seamless integration with SONA learning systems - Flash Attention - 2.49x-7.47x speedup with attention coordination - SDK Bridge - Version negotiation and API compatibility layer - Feature Flags - Dynamic feature enabling/disabling - Runtime Detection - Auto-select optimal runtime (NAPI, WASM, JS) - Graceful Fallback - Works with or without agentic-flow installed","readmeSource":"ruflo/v3/@claude-flow/integration/README.md","product":"key:ruflo:integration","canonical":"@claude-flow/integration","tier":"T0","t":["integration","agentic-flow","flow","alpha","deep","adr-001","adr","001","compliance","tokenoptimizer","agent-booster","booster","anti-drift","anti","drift","claude-flow","claude","reasoningbank","swarm","token-optimizer","token","optimizer"],"s":["claude-flow","claude","flow","integration","agent-booster","booster","agentic-flow","anti-drift","anti","drift","reasoningbank","swarm","token-optimizer","token","optimizer"]},
+{"id":"@claude-flow/mcp","family":"mcp","kind":"npm","store":"ruflo","version":"3.0.0","description":"Standalone MCP (Model Context Protocol) server - stdio/http/websocket transports, connection pooling, tool registry","keywords":["ai-agents","claude","json-rpc","mcp","model-context-protocol","stdio","tool-registry","websocket"],"source":"ruflo/v3/@claude-flow/mcp/package.json","sourceSha256":"91c7f53f6d69b633f003ad1d882894dc9433bbf56efd622aed7308788da8794a","variants":[],"readme":"- MCP 2025-11-25 Compliant - Full specification support - High-Performance Server - <400ms startup time - Multiple Transports - stdio, HTTP, WebSocket, in-process - Resources - list, read, subscribe with caching - Prompts - templates with arguments and embedded resources - Tasks - Async operations with progress and cancellation - Pagination - Cursor-based pagination for large datasets - Connection Pooling - Max 10 connections, configurable - Fast Tool Registry - O(1) lookup, <10ms registration - Session Management - Timeout handling, authentication - Security - CORS, Helmet, auth tokens - Zero Dependencies - No @claude-flow/* dependencies","readmeSource":"ruflo/v3/@claude-flow/mcp/README.md","product":"key:ruflo:mcp","canonical":"@claude-flow/mcp","tier":"T0","t":["mcp","standalone","model","context","protocol","server","stdio","http","websocket","transport","connection","pool","registry","ai-agents","claude","json-rpc","json","rpc","model-context-protocol","tool-registry"],"s":["claude-flow","claude","flow","mcp","ai-agents","json-rpc","json","rpc","model-context-protocol","model","context","protocol","stdio","tool-registry","registry","websocket"]},
+{"id":"@claude-flow/memory","family":"memory","kind":"npm","store":"ruflo","version":"3.0.0","description":"Memory module - AgentDB unification, HNSW indexing, vector search, hybrid SQLite+AgentDB backend (ADR-009)","keywords":[],"source":"ruflo/v3/@claude-flow/memory/package.json","sourceSha256":"6ccdef0fb26751583b4dc2a77f402df53dce4f8c33980aa3057d8cae4c47def2","variants":[],"readme":"This release closes out [ADR-125 — Memory Consolidation](../../../v3/docs/adr/ADR-125-memory-consolidation.md). If you're upgrading from 3.0.0-alpha.17 or earlier, this is the headline: - One canonical entry point. MemoryService replaces the old UnifiedMemoryService name (both still exported — the old one is @deprecated and will be removed at 3.0.0-rc). See [Migration](#migrating-from-30-0-alpha-17-or-earlier) below. - A real hybrid default. createHybridService(...) finally returns a service whose backend is an actual HybridBackend (sql.js + AgentDB). Earlier releases silently downgraded this to AgentDB-only — ADR-009's promise wasn't delivered until now. - Persistent HNSW. Closing a service","readmeSource":"ruflo/v3/@claude-flow/memory/README.md","product":"key:ruflo:memory","canonical":"@claude-flow/memory","tier":"T0","t":["memory","agentdb","unification","hnsw","index","vector","search","hybrid","sqlite+agentdb","backend","adr-009","adr","009"],"s":["claude-flow","claude","flow","memory"]},
+{"id":"@claude-flow/migration","family":"migration","kind":"npm","store":"synaptic-mesh","version":"1.0.0","description":"Migration system for claude-flow projects","keywords":["automation","claude-flow","migration","optimization","prompts"],"source":"synaptic-mesh/src/js/claude-flow/src/migration/package.json","sourceSha256":"2ba998a60f8a6951b87f4548f140e9a46f4ce0f80e1a7e70ec3a9e8bd98db9a7","variants":[],"readme":"A comprehensive migration system for existing claude-flow projects to adopt optimized prompts and configurations. The migration system provides tools to: - Analyze existing projects for migration readiness - Migrate projects using different strategies (full, selective, merge) - Create automatic backups for safe rollback - Validate successful migrations - Handle edge cases and conflicts gracefully 1. MigrationAnalyzer - Analyzes project structure and detects conflicts 2. MigrationRunner - Executes migration strategies 3. RollbackManager - Handles backup creation and restoration 4. MigrationValidator - Validates successful migrations 5. ProgressReporter - Provides visual feedback during operat","readmeSource":"synaptic-mesh/src/js/claude-flow/src/migration/README.md","product":"key:synaptic-mesh:migration","canonical":"@claude-flow/migration","tier":"T1","t":["migration","claude-flow","claude","flow","project","automation","prompt"],"s":["claude-flow","claude","flow","migration","automation","prompt"]},
+{"id":"@claude-flow/neural","family":"neural","kind":"npm","store":"ruflo","version":"3.0.0","description":"Self-Optimizing Neural Architecture (SONA) for Claude Flow — adaptive learning, trajectory tracking, pattern reuse, 7 RL algorithms (PPO/A2C/DQN/Q-Learning/SARSA/Decision Transformer/Curiosity), Flash Attention, MoE routing, LoRA, EWC++ for continual learning. Reproducible via seedable PRNG; persistence via serialize/deserialize.","keywords":["a2c","adaptive-learning","advantage-actor-critic","agent-coordination","agent-orchestration","agentdb","agentic","agentic-ai","agentic-systems","agents","ai","ai-agents","anthropic","catastrophic-forgetting","claude","claude-code","claude-flow","continual-learning","cosine-similarity","curiosity-driven","decision-transformer","deep-q-network","deep-rl","dqn","elastic-weight-consolidation","embeddings","ensemble","esm","ewc","experience-replay","expert-routing","exploration","fast-attention","fine-tuning","flash-attention","hnsw","knowledge-distillation","lifelong-learning","llm","llm-agents","lora","low-rank-adaptation","machine-learning","mcp","memory","meta-learning","mixture-of-experts","ml","model-compression","model-context-protocol","moe","multi-agent","multi-agent-systems","neural","neural-network","neural-networks","online-learning","open-source","pattern-evolution","pattern-extraction","pattern-matching","pattern-recognition","policy-gradient","ppo","proximal-policy-optimization","q-learning","quantization","reasoning","reasoning-bank","reasoning-traces","reinforcement-learning","rl","ruflo","sarsa","self-consistency","self-improving","self-optimizing","semantic-search","sona","trajectory","trajectory-learning","transfer-learning","typescript","value-function","vector-embeddings","vector-search","wasm"],"source":"ruflo/v3/@claude-flow/neural/package.json","sourceSha256":"a12d304024d4d38a88a21dcd8d3568ddf13985f0b93bb9074701594184fa2363","variants":[],"readme":"A self-contained learning module that records agent execution trajectories, distills them into reusable patterns, retrieves matches for new tasks, and adapts via SONA + LoRA + EWC++. Designed to be the substrate that the Claude Flow CLI's intelligence layer composes onto — the package owns the algorithms, the CLI owns the orchestration. NeuralLearningSystem is the high-level entry point — it wires SONAManager, ReasoningBank, and PatternLearner together so callers don't have to: For callers that want to manage trajectories and patterns directly:","readmeSource":"ruflo/v3/@claude-flow/neural/README.md","product":"key:ruflo:neural","canonical":"@claude-flow/neural","tier":"T0","t":["neural","self-optimizing","self","optimiz","architecture","sona","claude","flow","adaptive","learn","trajectory","track","pattern","reuse","algorithm","ppo","a2c","dqn","q-learning","sarsa","decision","transformer","curiosity","flash","attention","moe","rout","lora","ewc","continual","reproducible","via","seedable","prng","persistence","serialize","deserialize","adaptive-learning","advantage-actor-critic","advantage","actor","critic","agent-coordination","coordination","agent-orchestration","orchestration","agentdb","agentic-ai","agentic-systems","ai-agents","anthropic","catastrophic-forgetting","catastrophic","forgett","claude-code","code","claude-flow","continual-learning","cosine-similarity","cosine","similarity","curiosity-driven","driven","decision-transformer","deep-q-network","deep","network","deep-rl","elastic-weight-consolidation","elastic","weight","consolidation","embedding","ensemble","esm","experience-replay","experience","replay","expert-routing","expert","exploration","fast-attention","fine-tuning","fine","tuning","flash-attention","hnsw","knowledge-distillation","knowledge","distillation","lifelong-learning","lifelong","llm-agents","low-rank-adaptation","low","rank","adaptation","machine-learning","machine","mcp","memory","meta-learning","meta","mixture-of-experts","mixture","model-compression","model","compression","model-context-protocol","context","protocol","multi-agent","multi","multi-agent-systems","neural-network","neural-networks","online-learning","online","open-source","open","source","pattern-evolution","evolution","pattern-extraction","extraction","pattern-matching","match","pattern-recognition","recognition","policy-gradient","policy","gradient","proximal-policy-optimization","proximal","quantization","reason","reasoning-bank","bank","reasoning-traces","trace","reinforcement-learning","reinforcement","ruflo","self-consistency","consistency","self-improving","improv","semantic-search","semantic","search","trajectory-learning","transfer-learning","transfer","value-function","value","function","vector-embeddings","vector","vector-search"],"s":["claude-flow","claude","flow","neural","a2c","adaptive-learning","adaptive","learn","advantage-actor-critic","advantage","actor","critic","agent-coordination","coordination","agent-orchestration","orchestration","agentdb","agentic-ai","agentic-systems","ai-agents","anthropic","catastrophic-forgetting","catastrophic","forgett","claude-code","code","continual-learning","continual","cosine-similarity","cosine","similarity","curiosity-driven","curiosity","driven","decision-transformer","decision","transformer","deep-q-network","deep","network","deep-rl","dqn","elastic-weight-consolidation","elastic","weight","consolidation","embedding","ensemble","esm","ewc","experience-replay","experience","replay","expert-routing","expert","rout","exploration","fast-attention","attention","fine-tuning","fine","tuning","flash-attention","flash","hnsw","knowledge-distillation","knowledge","distillation","lifelong-learning","lifelong","llm-agents","lora","low-rank-adaptation","low","rank","adaptation","machine-learning","machine","mcp","memory","meta-learning","meta","mixture-of-experts","mixture","model-compression","model","compression","model-context-protocol","context","protocol","moe","multi-agent","multi","multi-agent-systems","neural-network","neural-networks","online-learning","online","open-source","open","source","pattern-evolution","pattern","evolution","pattern-extraction","extraction","pattern-matching","match","pattern-recognition","recognition","policy-gradient","policy","gradient","ppo","proximal-policy-optimization","proximal","q-learning","quantization","reason","reasoning-bank","bank","reasoning-traces","trace","reinforcement-learning","reinforcement","ruflo","sarsa","self-consistency","self","consistency","self-improving","improv","self-optimizing","optimiz","semantic-search","semantic","search","sona","trajectory","trajectory-learning","transfer-learning","transfer","value-function","value","function","vector-embeddings","vector","vector-search"]},
+{"id":"@claude-flow/performance","family":"performance","kind":"npm","store":"ruflo","version":"3.0.0","description":"Performance module - benchmarking, Flash Attention validation, optimization","keywords":[],"source":"ruflo/v3/@claude-flow/performance/package.json","sourceSha256":"3eb838f95cf4d555e279d103b9bb5c405ee89f245a7504a4ec66613a42b0cbc5","variants":[],"readme":"- Statistical Benchmarking - Mean, median, P95, P99, standard deviation, outlier removal - Memory Tracking - Heap, RSS, external, and array buffer monitoring - Auto-Calibration - Automatically adjusts iterations for statistical significance - Regression Detection - Compare against baselines with significance testing - V3 Performance Targets - Built-in targets for CLI, memory, swarm, and attention - Flash Attention Validation - Validate 2.49x-7.47x speedup targets - Multiple Output Formats - Console, JSON, and programmatic access","readmeSource":"ruflo/v3/@claude-flow/performance/README.md","product":"key:ruflo:performance","canonical":"@claude-flow/performance","tier":"T0","t":["benchmark","flash","attention","validation"],"s":["claude-flow","claude","flow"]},
+{"id":"@claude-flow/plugin-agent-federation","family":"plugin-agent-federation","kind":"npm","store":"ruflo","version":"1.0.0","description":"Cross-installation agent federation with zero-trust security, PII-gated data flow, and compliance-grade audit trails.","keywords":["audit","compliance","federation","multi-agent","pii","security","trust"],"source":"ruflo/v3/@claude-flow/plugin-agent-federation/package.json","sourceSha256":"755dd5c7f3b238aa1b457213e1ea1af48cadcad92000eaa8812fd20a5c517363","variants":[],"readme":"Cross-installation agent federation with zero-trust security, PII-gated data flow, and compliance-grade audit trails. 325 unit tests covering audit, routing, discovery, plugin lifecycle. MIT — Claude Flow Team.","readmeSource":"ruflo/v3/@claude-flow/plugin-agent-federation/README.md","product":"key:ruflo:plugin-agent-federation","canonical":"@claude-flow/plugin-agent-federation","tier":"T0","t":["plugin-agent-federation","plugin","federation","cross-installation","cross","installation","zero-trust","zero","trust","security","pii-gated","pii","gated","flow","compliance-grade","compliance","grade","audit","trail","multi-agent","multi"],"s":["claude-flow","claude","flow","plugin-agent-federation","plugin","federation","audit","compliance","multi-agent","multi","pii","security","trust"]},
+{"id":"@claude-flow/plugin-agentic-qe","family":"plugin-agentic-qe","kind":"npm","store":"ruflo","version":"3.0.0-alpha.5","description":"Quality Engineering plugin for Ruflo V3 with 51 specialized agents across 12 DDD bounded contexts","keywords":["accessibility","ai-agents","chaos-engineering","claude-flow","coverage","plugin","quality-engineering","security","tdd","testing"],"source":"ruflo/v3/plugins/agentic-qe/package.json","sourceSha256":"8c1aa32ef36a98f56592319355c850848497c7ec2e0a4b9c7ab2f7c08e58f39b","variants":[],"readme":"AI-powered quality engineering that writes tests, finds bugs, and breaks things (safely) so your users don't have to. This plugin adds 58 AI agents to Claude Flow that handle all aspects of software quality: - Write tests for you - Unit tests, integration tests, E2E tests, even chaos tests - Find coverage gaps - Shows exactly which code paths aren't tested - Predict bugs before they happen - ML-based defect prediction from code patterns - Security scanning - Find vulnerabilities, secrets, and compliance issues - Break things on purpose - Chaos engineering to test resilience (safely!) Think of it as having a team of QA engineers who never sleep, never miss edge cases, and learn from every bug","readmeSource":"ruflo/v3/plugins/agentic-qe/README.md","product":"key:ruflo:plugin-agentic-qe","canonical":"@claude-flow/plugin-agentic-qe","tier":"T0","t":["plugin-agentic-qe","plugin","quality","engineer","ruflo","specializ","across","ddd","bound","context","accessibility","ai-agents","chaos-engineering","chao","claude-flow","claude","flow","coverage","quality-engineering","security","tdd","test"],"s":["claude-flow","claude","flow","plugin-agentic-qe","plugin","accessibility","ai-agents","chaos-engineering","chao","engineer","coverage","quality-engineering","quality","security","tdd","test"]},
+{"id":"@claude-flow/plugin-code-intelligence","family":"plugin-code-intelligence","kind":"npm","store":"ruflo","version":"3.0.0-alpha.3","description":"Advanced code intelligence plugin for Claude Flow V3 - semantic search, architecture analysis, refactoring impact, module splitting, and pattern learning","keywords":["architecture","claude-flow","code-analysis","code-intelligence","mcp","plugin","refactoring","semantic-search"],"source":"ruflo/v3/plugins/code-intelligence/package.json","sourceSha256":"7611f97b97eb5e9ad5dcd7d84b7e8dbe89d7845c93d9ad6ff5f9ecf0f2cb5c99","variants":[],"readme":"A comprehensive code intelligence plugin combining graph neural networks for code structure analysis with ultra-fast vector search for semantic code similarity. The plugin enables dead code detection, API surface analysis, refactoring impact prediction, and architectural drift monitoring while integrating seamlessly with existing IDE workflows. - Semantic Code Search: Find semantically similar code across the codebase using natural language or code snippets - Architecture Analysis: Analyze dependency graphs, detect layer violations, circular dependencies, and architectural drift - Refactoring Impact Prediction: Predict the impact of proposed code changes using GNN analysis - Module Splitting","readmeSource":"ruflo/v3/plugins/code-intelligence/README.md","product":"key:ruflo:plugin-code-intelligence","canonical":"@claude-flow/plugin-code-intelligence","tier":"T0","t":["plugin-code-intelligence","plugin","code","intelligence","advanc","claude","flow","semantic","search","architecture","analysis","refactor","impact","splitt","pattern","learn","claude-flow","code-analysis","code-intelligence","mcp","semantic-search"],"s":["claude-flow","claude","flow","plugin-code-intelligence","plugin","code","intelligence","architecture","code-analysis","analysis","code-intelligence","mcp","refactor","semantic-search","semantic","search"]},
+{"id":"@claude-flow/plugin-cognitive-kernel","family":"plugin-cognitive-kernel","kind":"npm","store":"ruflo","version":"3.0.0-alpha.1","description":"Cognitive kernel plugin for LLM augmentation with working memory, attention control, meta-cognition, and scaffolding","keywords":["attention","claude-flow","cognitive-kernel","cognitum","meta-cognition","scaffolding","sona","working-memory"],"source":"ruflo/v3/plugins/cognitive-kernel/package.json","sourceSha256":"131651b3b29f562e3a20b131fe544fb32156c4a1ae3abec6c4f74b2bd59293eb","variants":[],"readme":"A cutting-edge cognitive augmentation plugin combining the Cognitum Gate Kernel with SONA self-optimizing architecture to provide LLMs with enhanced cognitive capabilities. The plugin enables dynamic working memory, attention control mechanisms, meta-cognitive self-monitoring, and cognitive scaffolding while maintaining low latency through WASM acceleration. Manage dynamic working memory slots for complex reasoning tasks. Actions: allocate, update, retrieve, clear, consolidate Returns: Memory slot state with current contents and decay status. Control cognitive attention and information filtering. Modes: focus, diffuse, selective, divided, sustained Returns: Attention state with active target","readmeSource":"ruflo/v3/plugins/cognitive-kernel/README.md","product":"key:ruflo:plugin-cognitive-kernel","canonical":"@claude-flow/plugin-cognitive-kernel","tier":"T0","t":["plugin-cognitive-kernel","plugin","cognitive","kernel","augmentation","memory","attention","control","meta-cognition","meta","cognition","scaffold","claude-flow","claude","flow","cognitive-kernel","cognitum","sona","working-memory"],"s":["claude-flow","claude","flow","plugin-cognitive-kernel","plugin","cognitive","kernel","attention","cognitive-kernel","cognitum","meta-cognition","meta","cognition","scaffold","sona","working-memory","memory"]},
+{"id":"@claude-flow/plugin-financial-risk","family":"plugin-financial-risk","kind":"npm","store":"ruflo","version":"3.0.0-alpha.1","description":"High-performance financial risk analysis plugin with portfolio risk scoring, anomaly detection, market regime classification, and compliance reporting using RuVector WASM packages.","keywords":["aml","anomaly-detection","basel3","claude-flow","compliance","cvar","financial-risk","fraud-detection","mcp","mifid2","plugin","portfolio-risk","var"],"source":"ruflo/v3/plugins/financial-risk/package.json","sourceSha256":"7d3ae2b4423063a18d54bee42ec5fc9d62e62500dec283aa7927635268451792","variants":[],"readme":"A high-performance financial risk analysis plugin combining sparse inference for efficient market signal processing with graph neural networks for transaction network analysis. The plugin enables real-time anomaly detection, portfolio risk scoring, and automated compliance reporting while maintaining the explainability required by financial regulators (SEC, FINRA, Basel III). - Portfolio Risk Analysis: Calculate VaR, CVaR, Sharpe ratio, Sortino ratio, and max drawdown - Anomaly Detection: Detect fraud, AML violations, and market manipulation in transactions - Market Regime Classification: Identify current market regime through historical pattern matching - Compliance Checking: Automated veri","readmeSource":"ruflo/v3/plugins/financial-risk/README.md","product":"key:ruflo:plugin-financial-risk","canonical":"@claude-flow/plugin-financial-risk","tier":"T0","t":["plugin-financial-risk","plugin","financial","risk","high","analysis","portfolio","scor","anomaly","detection","market","regime","classification","compliance","report","aml","anomaly-detection","basel3","claude-flow","claude","flow","cvar","financial-risk","fraud-detection","fraud","mcp","mifid2","portfolio-risk","var"],"s":["claude-flow","claude","flow","plugin-financial-risk","plugin","financial","risk","aml","anomaly-detection","anomaly","detection","basel3","compliance","cvar","financial-risk","fraud-detection","fraud","mcp","mifid2","portfolio-risk","portfolio","var"]},
+{"id":"@claude-flow/plugin-gastown-bridge","family":"plugin-gastown-bridge","kind":"npm","store":"ruflo","version":"0.1.4","description":"Gas Town orchestrator integration for Claude Flow V3 with WASM-accelerated formula parsing and graph analysis","keywords":["beads","claude-flow","formulas","gastown","multi-agent","orchestration","plugin","wasm","workflows"],"source":"ruflo/v3/plugins/gastown-bridge/package.json","sourceSha256":"55b48ac8f8c35e48d2443ba0434c646816329c9e62550e0981055a53983691c9","variants":[],"readme":"The Gas Town Bridge Plugin brings Steve Yegge's powerful [Gas Town](https://github.com/steveyegge/gastown) multi-agent orchestrator to Claude Flow V3. Gas Town introduces battle-tested concepts for durable workflow execution that complement Claude Flow's swarm intelligence. Gas Town is a 75,000-line Go codebase that implements: - Beads - Git-backed issue tracking with graph semantics - Formulas - TOML-defined workflows (convoy, workflow, expansion, aspect) - Convoys - Work-order tracking for \"slung\" work between agents - GUPP - Gastown Universal Propulsion Principle for crash-resilient execution - Molecules/Wisps - Chained work units for durable, resumable workflows - Input Validation: Zod s","readmeSource":"ruflo/v3/plugins/gastown-bridge/README.md","product":"key:ruflo:plugin-gastown-bridge","canonical":"@claude-flow/plugin-gastown-bridge","tier":"T0","t":["plugin-gastown-bridge","plugin","gastown","bridge","gas","town","orchestrator","integration","claude","flow","wasm-accelerated","accelerat","formula","pars","graph","analysis","bead","claude-flow","multi-agent","multi","orchestration","workflow"],"s":["claude-flow","claude","flow","plugin-gastown-bridge","plugin","gastown","bridge","bead","formula","multi-agent","multi","orchestration","workflow"]},
+{"id":"@claude-flow/plugin-healthcare-clinical","family":"plugin-healthcare-clinical","kind":"npm","store":"ruflo","version":"3.0.0-alpha.2","description":"HIPAA-compliant clinical decision support plugin with patient similarity search, drug interaction detection, and clinical pathway recommendations using RuVector WASM packages.","keywords":["claude-flow","clinical-decision-support","drug-interactions","fhir","healthcare","hipaa","hl7","icd-10","mcp","patient-similarity","plugin","snomed-ct"],"source":"ruflo/v3/plugins/healthcare-clinical/package.json","sourceSha256":"cf5e79d3fc71858bcc5c2003294bf0073465f21d3547785dcd82fa82dc4679d4","variants":[],"readme":"A HIPAA-compliant clinical decision support plugin that combines ultra-fast vector search for medical literature retrieval with graph neural networks for patient pathway analysis. The plugin enables semantic search across medical records, drug interaction detection, and evidence-based treatment recommendations while maintaining strict data privacy through on-device WASM processing. - Patient Similarity Search: Find similar patient cases based on clinical features (diagnoses, labs, vitals, medications) - Drug Interaction Detection: Analyze drug-drug and drug-condition interactions using graph neural networks - Clinical Pathway Recommendations: Suggest evidence-based clinical pathways based on","readmeSource":"ruflo/v3/plugins/healthcare-clinical/README.md","product":"key:ruflo:plugin-healthcare-clinical","canonical":"@claude-flow/plugin-healthcare-clinical","tier":"T0","t":["plugin-healthcare-clinical","plugin","healthcare","clinical","hipaa-compliant","hipaa","compliant","decision","patient","similarity","search","drug","interaction","detection","pathway","recommendation","claude-flow","claude","flow","clinical-decision-support","drug-interactions","fhir","hl7","icd-10","icd","mcp","patient-similarity","snomed-ct","snom"],"s":["claude-flow","claude","flow","plugin-healthcare-clinical","plugin","healthcare","clinical","clinical-decision-support","decision","drug-interactions","drug","interaction","fhir","hipaa","hl7","icd-10","icd","mcp","patient-similarity","patient","similarity","snomed-ct","snom"]},
+{"id":"@claude-flow/plugin-hyperbolic-reasoning","family":"plugin-hyperbolic-reasoning","kind":"npm","store":"ruflo","version":"3.0.0-alpha.1","description":"Hyperbolic reasoning plugin providing Poincare ball embeddings, taxonomic reasoning, hierarchical search, and entailment graphs for superior hierarchical understanding.","keywords":["claude-flow","embeddings","entailment","gnn","hierarchy","hyperbolic","mcp","plugin","poincare","taxonomy"],"source":"ruflo/v3/plugins/hyperbolic-reasoning/package.json","sourceSha256":"2943637743c02b312722a36a245e10252caddf010cc39f7ff9ee2348087d1afc","variants":[],"readme":"An exotic reasoning plugin implementing hyperbolic neural networks for superior hierarchical understanding. The plugin enables efficient representation of tree structures, taxonomic reasoning, and hierarchical entailment using Poincare ball embeddings with Mobius operations. Applications include improved ontology navigation, hierarchical code understanding, and organizational relationship modeling. Embed hierarchical data in hyperbolic space (Poincare ball). Models: poincare_ball, lorentz, klein, half_plane Returns: Hyperbolic embeddings with learned curvature and embedding quality metrics. Perform taxonomic reasoning using hyperbolic entailment. Query Types: is_a, subsumption, lowest_common","readmeSource":"ruflo/v3/plugins/hyperbolic-reasoning/README.md","product":"key:ruflo:plugin-hyperbolic-reasoning","canonical":"@claude-flow/plugin-hyperbolic-reasoning","tier":"T0","t":["plugin-hyperbolic-reasoning","plugin","hyperbolic","reason","provid","poincare","ball","embedding","taxonomic","hierarchical","search","entailment","graph","superior","understand","claude-flow","claude","flow","gnn","hierarchy","mcp","taxonomy"],"s":["claude-flow","claude","flow","plugin-hyperbolic-reasoning","plugin","hyperbolic","reason","embedding","entailment","gnn","hierarchy","mcp","poincare","taxonomy"]},
+{"id":"@claude-flow/plugin-iot-cognitum","family":"plugin-iot-cognitum","kind":"npm","store":"ruflo","version":"1.0.0","description":"IoT Cognitum Seed device-agent bridge — treat every Seed as a Ruflo agent. Get a Seed at https://cognitum.one.","keywords":["agent","cognitum","device","fleet","iot","seed","swarm"],"source":"ruflo/v3/@claude-flow/plugin-iot-cognitum/package.json","sourceSha256":"911f8fbdd8a9fdc9eb51725c048e533e7d54ef75b08f7eab3042f3a03e66f56c","variants":[],"readme":"IoT Cognitum Seed device-agent bridge — treat every Seed as a Ruflo agent with hardware capabilities. This plugin requires a Cognitum Seed device. Get one at https://cognitum.one. The Seed is an edge appliance with on-device vector store, Ed25519 cryptographic identity, OTA firmware updates, mesh networking, and a witness chain for provenance. Default endpoints: - http://169.254.42.1 — link-local USB-C (no auth, read-only + pair window) - https://169.254.42.1:8443 — LAN/HTTPS with bearer token (full access including writes) Create a .env file in your project root (or any parent directory): The bin walks up from CWD looking for .env, loads it without overwriting existing process.env vars, the","readmeSource":"ruflo/v3/@claude-flow/plugin-iot-cognitum/README.md","product":"key:ruflo:plugin-iot-cognitum","canonical":"@claude-flow/plugin-iot-cognitum","tier":"T0","t":["plugin-iot-cognitum","plugin","iot","cognitum","seed","device-agent","device","bridge","treat","ruflo","http","cognitum.one","fleet","swarm"],"s":["claude-flow","claude","flow","plugin-iot-cognitum","plugin","iot","cognitum","device","fleet","seed","swarm"]},
+{"id":"@claude-flow/plugin-legal-contracts","family":"plugin-legal-contracts","kind":"npm","store":"ruflo","version":"3.0.0-alpha.2","description":"Legal contract analysis plugin for Claude Flow V3 - clause extraction, risk assessment, contract comparison, obligation tracking, and playbook matching","keywords":["claude-flow","clause-extraction","contract-comparison","contracts","legal","mcp","obligation-tracking","plugin","risk-assessment"],"source":"ruflo/v3/plugins/legal-contracts/package.json","sourceSha256":"4f5ddcfa47fdc8abbcad073e97a35cedbe9fee9722af2b20fc5c92e5bc0a5a22","variants":[],"readme":"A comprehensive legal contract analysis plugin combining hyperbolic embeddings for legal ontology navigation with fast vector search for clause similarity. The plugin enables automated clause extraction, risk scoring, obligation tracking, and regulatory compliance checking while maintaining attorney-client privilege through on-device processing. - Clause Extraction: Extract and classify clauses from contracts (indemnification, liability, termination, etc.) - Risk Assessment: Identify and score contractual risks by category and severity - Contract Comparison: Compare contracts with detailed diff and semantic alignment - Obligation Tracking: Extract obligations, deadlines, and dependencies usi","readmeSource":"ruflo/v3/plugins/legal-contracts/README.md","product":"key:ruflo:plugin-legal-contracts","canonical":"@claude-flow/plugin-legal-contracts","tier":"T0","t":["plugin-legal-contracts","plugin","legal","contract","analysis","claude","flow","clause","extraction","risk","assessment","comparison","obligation","track","playbook","match","claude-flow","clause-extraction","contract-comparison","mcp","obligation-tracking","risk-assessment"],"s":["claude-flow","claude","flow","plugin-legal-contracts","plugin","legal","contract","clause-extraction","clause","extraction","contract-comparison","comparison","mcp","obligation-tracking","obligation","track","risk-assessment","risk","assessment"]},
+{"id":"@claude-flow/plugin-neural-coordination","family":"plugin-neural-coordination","kind":"npm","store":"ruflo","version":"3.0.0-alpha.1","description":"Neural coordination plugin for multi-agent swarm intelligence using SONA, GNN, and attention mechanisms","keywords":["attention","claude-flow","consensus","gnn","multi-agent","neural-coordination","sona","swarm","topology"],"source":"ruflo/v3/plugins/neural-coordination/package.json","sourceSha256":"44d69482a27fd00183bf12b232c0f5c6fdb4a4f340ced2e1c07a22cee36a75a5","variants":[],"readme":"A cutting-edge multi-agent coordination plugin combining the SONA self-optimizing neural architecture with graph neural networks for agent communication topology optimization. The plugin enables emergent protocol development, neural consensus mechanisms, collective memory formation, and adaptive swarm behavior while maintaining interpretability of agent interactions. Achieve agent consensus using neural negotiation protocols. Returns: Consensus decision with voting breakdown, confidence scores, and round-by-round negotiation history. Optimize agent communication topology for efficiency using GNN analysis. Returns: Optimized communication graph with connection weights and routing recommendati","readmeSource":"ruflo/v3/plugins/neural-coordination/README.md","product":"key:ruflo:plugin-neural-coordination","canonical":"@claude-flow/plugin-neural-coordination","tier":"T0","t":["plugin-neural-coordination","plugin","neural","coordination","multi-agent","multi","swarm","intelligence","sona","gnn","attention","mechanism","claude-flow","claude","flow","consensus","neural-coordination","topology"],"s":["claude-flow","claude","flow","plugin-neural-coordination","plugin","neural","coordination","attention","consensus","gnn","multi-agent","multi","neural-coordination","sona","swarm","topology"]},
+{"id":"@claude-flow/plugin-perf-optimizer","family":"plugin-perf-optimizer","kind":"npm","store":"ruflo","version":"3.0.0-alpha.2","description":"AI-powered performance optimization plugin for bottleneck detection, memory analysis, and configuration tuning","keywords":["bottleneck-detection","claude-flow","memory-analysis","optimization","performance","plugin","query-optimization"],"source":"ruflo/v3/plugins/perf-optimizer/package.json","sourceSha256":"47adfd55da69c9d9665aecb8e461a20b04143239b332c582090d2194836dd630","variants":[],"readme":"A comprehensive performance optimization plugin combining sparse inference for efficient trace analysis with graph neural networks for dependency chain optimization. The plugin enables intelligent bottleneck detection, memory leak identification, N+1 query detection, and bundle size optimization while providing explainable recommendations based on historical performance patterns. - Bottleneck Detection: Identify performance bottlenecks using GNN-based dependency analysis - Memory Analysis: Detect memory leaks, retention chains, and GC pressure points - Query Optimization: Detect N+1 queries, missing indexes, and slow joins - Bundle Optimization: Analyze and optimize JavaScript bundle size wi","readmeSource":"ruflo/v3/plugins/perf-optimizer/README.md","product":"key:ruflo:plugin-perf-optimizer","canonical":"@claude-flow/plugin-perf-optimizer","tier":"T0","t":["plugin-perf-optimizer","plugin","perf","optimizer","ai-powered","power","bottleneck","detection","memory","analysis","configuration","tuning","bottleneck-detection","claude-flow","claude","flow","memory-analysis","query-optimization","query"],"s":["claude-flow","claude","flow","plugin-perf-optimizer","plugin","perf","optimizer","bottleneck-detection","bottleneck","detection","memory-analysis","memory","analysis","query-optimization","query"]},
+{"id":"@claude-flow/plugin-prime-radiant","family":"plugin-prime-radiant","kind":"npm","store":"ruflo","version":"0.1.5","description":"Mathematical AI interpretability plugin providing sheaf cohomology, spectral analysis, causal inference, and quantum topology for coherence validation, consensus verification, and hallucination prevention.","keywords":["ai-interpretability","causal-inference","claude-flow","coherence-engine","mcp","plugin","quantum-topology","rag","sheaf-cohomology"],"source":"ruflo/v3/plugins/prime-radiant/package.json","sourceSha256":"ab6b34b65b9756fa61094bf59eed1c2468a93b2c90bdee246856ef7e984f1e30","variants":[],"readme":"Mathematical AI that catches contradictions, verifies consensus, and prevents hallucinations before they cause problems. This plugin brings advanced mathematical techniques to Claude Flow for ensuring AI reliability: - Coherence Checking - Detect when information contradicts itself before storing it - Consensus Verification - Mathematically verify that multiple agents actually agree - Hallucination Prevention - Catch inconsistent RAG results before they reach users - Stability Analysis - Monitor swarm health using spectral graph theory - Causal Inference - Understand cause-and-effect, not just correlations Think of it as a mathematical \"sanity check\" layer that catches logical inconsistencie","readmeSource":"ruflo/v3/plugins/prime-radiant/README.md","product":"key:ruflo:plugin-prime-radiant","canonical":"@claude-flow/plugin-prime-radiant","tier":"T0","t":["plugin-prime-radiant","plugin","prime","radiant","mathematical","interpretability","provid","sheaf","cohomology","spectral","analysis","causal","inference","quantum","topology","coherence","validation","consensus","verification","hallucination","prevention","ai-interpretability","causal-inference","claude-flow","claude","flow","coherence-engine","mcp","quantum-topology","rag","sheaf-cohomology"],"s":["claude-flow","claude","flow","plugin-prime-radiant","plugin","prime","radiant","ai-interpretability","interpretability","causal-inference","causal","inference","coherence-engine","coherence","mcp","quantum-topology","quantum","topology","rag","sheaf-cohomology","sheaf","cohomology"]},
+{"id":"@claude-flow/plugin-quantum-optimizer","family":"plugin-quantum-optimizer","kind":"npm","store":"ruflo","version":"3.0.0-alpha.1","description":"Quantum-inspired optimization plugin providing simulated annealing, QAOA, Grover search, dependency resolution, and schedule optimization for combinatorial problems.","keywords":["annealing","claude-flow","dependency-resolution","grover","mcp","optimization","plugin","qaoa","quantum","scheduling"],"source":"ruflo/v3/plugins/quantum-optimizer/package.json","sourceSha256":"a312fd41591893f49de0de1e992f5d14c684690e0f55998a5c7317f17bd21ca6","variants":[],"readme":"An exotic optimization plugin implementing quantum-inspired algorithms including Quantum Annealing simulation, QAOA (Quantum Approximate Optimization Algorithm) emulation, and Grover-inspired search acceleration. The plugin provides dramatic speedups for dependency resolution, optimal scheduling, and constraint satisfaction while running entirely on classical WASM-accelerated hardware. Solve combinatorial optimization problems using simulated quantum annealing. Problem Types: qubo, ising, sat, max_cut, tsp, dependency Returns: Optimal or near-optimal solution with energy value and convergence statistics. Approximate optimization using Quantum Approximate Optimization Algorithm emulation. Pro","readmeSource":"ruflo/v3/plugins/quantum-optimizer/README.md","product":"key:ruflo:plugin-quantum-optimizer","canonical":"@claude-flow/plugin-quantum-optimizer","tier":"T0","t":["plugin-quantum-optimizer","plugin","quantum","optimizer","quantum-inspired","inspir","provid","simulat","anneal","qaoa","grover","search","dependency","resolution","schedule","combinatorial","problem","claude-flow","claude","flow","dependency-resolution","mcp","schedul"],"s":["claude-flow","claude","flow","plugin-quantum-optimizer","plugin","quantum","optimizer","anneal","dependency-resolution","dependency","resolution","grover","mcp","qaoa","schedul"]},
+{"id":"@claude-flow/plugin-test-intelligence","family":"plugin-test-intelligence","kind":"npm","store":"ruflo","version":"3.0.0-alpha.1","description":"AI-powered test intelligence plugin for predictive test selection, flaky detection, and coverage optimization","keywords":["claude-flow","coverage-analysis","flaky-detection","mutation-testing","plugin","predictive-testing","test-intelligence"],"source":"ruflo/v3/plugins/test-intelligence/package.json","sourceSha256":"7bce3ef5b382d70783eea82f96a923163a1048be68c15f80418ff6f9fd41c720","variants":[],"readme":"A comprehensive test intelligence plugin combining reinforcement learning for optimal test selection with graph neural networks for code-to-test mapping. The plugin enables predictive test selection (run only tests likely to fail), flaky test detection, mutation testing optimization, and test coverage gap identification while integrating seamlessly with popular testing frameworks. - Predictive Test Selection: Select tests most likely to fail based on code changes using RL - Flaky Test Detection: Identify and analyze flaky tests with root cause classification - Coverage Gap Analysis: Identify test coverage gaps using code-test graph analysis - Mutation Testing Optimization: Optimize mutation","readmeSource":"ruflo/v3/plugins/test-intelligence/README.md","product":"key:ruflo:plugin-test-intelligence","canonical":"@claude-flow/plugin-test-intelligence","tier":"T0","t":["plugin-test-intelligence","plugin","test","intelligence","ai-powered","power","predictive","selection","flaky","detection","coverage","claude-flow","claude","flow","coverage-analysis","analysis","flaky-detection","mutation-testing","mutation","predictive-testing","test-intelligence"],"s":["claude-flow","claude","flow","plugin-test-intelligence","plugin","test","intelligence","coverage-analysis","coverage","analysis","flaky-detection","flaky","detection","mutation-testing","mutation","predictive-testing","predictive","test-intelligence"]},
+{"id":"@claude-flow/plugins","family":"plugins","kind":"npm","store":"ruflo","version":"3.0.1","description":"Unified Plugin SDK for Claude Flow V3 - Worker, Hook, and Provider Integration","keywords":["agentdb","agentic-flow","ai-agents","claude-flow","hnsw","hooks","lora","neural-adaptation","plugins","ruvector","sdk","vector-search","workers"],"source":"ruflo/v3/@claude-flow/plugins/package.json","sourceSha256":"2930e653820c821d20c22e569420854a089697ab6170e7ab15de057860519d04","variants":[],"readme":"Unified Plugin SDK for Claude Flow V3 A comprehensive plugin development framework providing workers, hooks, providers, and security utilities for building Claude Flow extensions.","readmeSource":"ruflo/v3/@claude-flow/plugins/README.md","product":"key:ruflo:plugins","canonical":"@claude-flow/plugins","tier":"T0","t":["plugin","unify","claude","flow","worker","hook","provider","integration","agentdb","agentic-flow","ai-agents","claude-flow","hnsw","lora","neural-adaptation","neural","adaptation","vector-search","vector","search"],"s":["claude-flow","claude","flow","plugin","agentdb","agentic-flow","ai-agents","hnsw","hook","lora","neural-adaptation","neural","adaptation","vector-search","vector","search","worker"]},
+{"id":"@claude-flow/providers","family":"providers","kind":"npm","store":"ruflo","version":"3.0.0","description":"Multi-LLM Provider System for Claude Flow V3","keywords":["anthropic","claude-flow","cohere","failover","google","llm","load-balancing","multi-provider","ollama","openai","provider"],"source":"ruflo/v3/@claude-flow/providers/package.json","sourceSha256":"0395df8444950f6eaa727d2b4a72653c32857a8cc12faddaa3c87630e201f85f","variants":[],"readme":"- 6+ LLM Providers - Anthropic, OpenAI, Google, Cohere, Ollama, RuVector - Load Balancing - Round-robin, latency-based, least-loaded, cost-based strategies - Automatic Failover - Seamless provider switching on failures - Request Caching - LRU cache with configurable TTL - Cost Optimization - Up to 85%+ savings with intelligent routing - Streaming Support - Full streaming for all providers - Tool Calling - Unified tool/function calling interface - Health Monitoring - Real-time provider health checks - Cost Tracking - Per-request and aggregate cost metrics - claude-3-5-sonnet-20241022, claude-3-5-sonnet-latest - claude-3-opus-20240229 - claude-3-sonnet-20240229 - claude-3-haiku-20240307 - gpt-","readmeSource":"ruflo/v3/@claude-flow/providers/README.md","product":"key:ruflo:providers","canonical":"@claude-flow/providers","tier":"T0","t":["provider","multi-llm","multi","claude","flow","anthropic","claude-flow","cohere","failover","google","load-balancing","load","balanc","multi-provider","ollama","openai"],"s":["claude-flow","claude","flow","provider","anthropic","cohere","failover","google","load-balancing","load","balanc","multi-provider","multi","ollama","openai"]},
+{"id":"@claude-flow/ruvector-upstream","family":"upstream","kind":"npm","store":"ruflo","version":"3.0.0-alpha.1","description":"RuVector WASM package bridges for Claude Flow plugins","keywords":["claude-flow","graph","neural","quantum","ruvector","vector-search","wasm"],"source":"ruflo/v3/plugins/ruvector-upstream/package.json","sourceSha256":"00d01fef6f86651ff3ba3ad5b7bb6dc87a9265ad7bb6622308c2d61c022c8e03","variants":[],"readme":"This directory contains references and integration bridges for upstream RuVector WASM packages used by Claude Flow plugins. All packages are sourced from: https://github.com/ruvnet/ruvector","readmeSource":"ruflo/v3/plugins/ruvector-upstream/README.md","product":"key:ruflo:upstream","canonical":"@claude-flow/ruvector-upstream","tier":"T0","t":["ruvector-upstream","upstream","bridge","claude","flow","plugin","claude-flow","graph","neural","quantum","vector-search","vector","search"],"s":["claude-flow","claude","flow","ruvector-upstream","upstream","graph","neural","quantum","vector-search","vector","search"]},
+{"id":"@claude-flow/security","family":"security","kind":"npm","store":"ruflo","version":"3.0.1","description":"Security module - CVE fixes, input validation, path security","keywords":[],"source":"ruflo/v3/@claude-flow/security/package.json","sourceSha256":"d727fe69b7828c10039544d851f7cfd621e7c43ebb1de2b7eb18b55bf448a9d7","variants":[],"readme":"- CVE Remediation - Fixes for CVE-2 (Weak Password Hashing), CVE-3 (Hardcoded Credentials), HIGH-1 (Command Injection), HIGH-2 (Path Traversal) - Password Hashing - Secure bcrypt-based password hashing with configurable rounds (12+ recommended) - Credential Generation - Cryptographically secure credential and API key generation - Safe Command Execution - Allowlist-based command execution preventing injection attacks - Path Validation - Protection against path traversal and symlink attacks - Input Validation - Zod-based schema validation for all input types - Token Generation - Secure token creation with HMAC signing This is a plain library — every primitive (PasswordHasher, PathValidator, Sa","readmeSource":"ruflo/v3/@claude-flow/security/README.md","product":"key:ruflo:security","canonical":"@claude-flow/security","tier":"T0","t":["security","cve","fix","input","validation","path"],"s":["claude-flow","claude","flow","security"]},
+{"id":"@claude-flow/shared","family":"shared","kind":"npm","store":"ruflo","version":"3.0.0","description":"Shared module - common types, events, utilities, core interfaces","keywords":[],"source":"ruflo/v3/@claude-flow/shared/package.json","sourceSha256":"d87838cfddba1ce7718b61b81aa331e8ea23cc6826786a26d46d32a4fb907fdf","variants":[],"readme":"- Core Types - Agent, Task, Memory, MCP, and Swarm type definitions - Core Interfaces - Agent, Task, Memory, Event, and Coordinator interfaces - Configuration - Schema validation, loading, and default values - Event System - Event bus, coordinator, and handler utilities - Hooks System - Pre/post execution hooks for extensibility - MCP Infrastructure - Server, transport, connection pool, and tool registry - Health Monitoring - Health checks and monitoring utilities","readmeSource":"ruflo/v3/@claude-flow/shared/README.md","product":"key:ruflo:shared","canonical":"@claude-flow/shared","tier":"T0","t":["shar","common","type","event","utility","core","interface"],"s":["claude-flow","claude","flow","shar"]},
+{"id":"@claude-flow/swarm","family":"swarm","kind":"npm","store":"ruflo","version":"3.0.0","description":"Standalone swarm coordination - up to 100+ agents, 4 topologies, hive-mind, consensus","keywords":["ai-agents","consensus","coordination","distributed","hive-mind","multi-agent","swarm"],"source":"ruflo/v3/@claude-flow/swarm/package.json","sourceSha256":"597b44a09da52f17b26da506b57d5a2fedf43408194e05cad4c467e29ca387b8","variants":[],"readme":"This module provides a complete multi-agent coordination system with hive-mind capabilities: The default 15-agent architecture is a recommendation, not a hard limit:","readmeSource":"ruflo/v3/@claude-flow/swarm/README.md","product":"key:ruflo:swarm","canonical":"@claude-flow/swarm","tier":"T0","t":["swarm","standalone","coordination","100","topology","hive-mind","hive","mind","consensus","ai-agents","distribut","multi-agent","multi"],"s":["claude-flow","claude","flow","swarm","ai-agents","consensus","coordination","distribut","hive-mind","hive","mind","multi-agent","multi"]},
+{"id":"@claude-flow/teammate-plugin","family":"teammate-plugin","kind":"npm","store":"ruflo","version":"1.0.0-alpha.1","description":"Native TeammateTool integration plugin for Claude Flow - bridges Claude Code v2.1.19+ multi-agent capabilities","keywords":["claude","claude-code","claude-flow","mcp","multi-agent","orchestration","swarm","teammate"],"source":"ruflo/v3/plugins/teammate-plugin/package.json","sourceSha256":"3a48fc38b9788746bfe1bbe8e7be885be09b9c92e04d681a8a17f7489ae42e76","variants":[],"readme":"Native TeammateTool integration plugin for Claude Flow. Bridges Claude Code v2.1.19+ multi-agent orchestration capabilities with Claude Flow's swarm system. If your version is below 2.1.19, update Claude Code: Install directly using Claude Code's plugin system: Or with pnpm:","readmeSource":"ruflo/v3/plugins/teammate-plugin/README.md","product":"key:ruflo:teammate-plugin","canonical":"@claude-flow/teammate-plugin","tier":"T0","t":["teammate-plugin","teammate","plugin","teammatetool","integration","claude","flow","bridge","code","v2.1.19","multi-agent","multi","capability","claude-code","claude-flow","mcp","orchestration","swarm"],"s":["claude-flow","claude","flow","teammate-plugin","teammate","plugin","claude-code","code","mcp","multi-agent","multi","orchestration","swarm"]},
+{"id":"@claude-flow/testing","family":"testing","kind":"npm","store":"ruflo","version":"3.0.0","description":"Testing module - TDD London School framework, test utilities, fixtures, and mock services for V3 Claude-Flow","keywords":["claude-flow","fixtures","london-school","mocks","tdd","testing","vitest"],"source":"ruflo/v3/@claude-flow/testing/package.json","sourceSha256":"253f560e474e3eb40ef1f9f914794dd23ef537a50e6c39270c0a3f89a65b65e0","variants":[],"product":"key:ruflo:testing","canonical":"@claude-flow/testing","tier":"T0","t":["test","tdd","london","school","utility","fixture","mock","service","claude-flow","claude","flow","london-school","vitest"],"s":["claude-flow","claude","flow","test","fixture","london-school","london","school","mock","tdd","vitest"]},
+{"id":"@claude-flow/v3-monorepo","family":"v3-monorepo","kind":"npm","store":"ruflo","version":"3.0.0-alpha.1","description":"Claude Flow V3 - Modular AI Agent Coordination System","keywords":[],"source":"ruflo/v3/package.json","sourceSha256":"704ed677da42eced26e54f4768f537fe7b44521088a5734fc15ee21c76bbcd9d","variants":[],"readme":"Claude Flow V3 is a next-generation AI agent coordination system built on 10 Architecture Decision Records (ADRs). It provides a modular, security-first, high-performance platform for orchestrating multi-agent swarms with hierarchical mesh topology. V3 represents a complete architectural overhaul: - 10x faster testing with Vitest - 150x-12,500x faster search with HNSW indexing - 2.49x-7.47x Flash Attention speedup - 50-75% memory reduction - 15-Agent Hierarchical Mesh - Queen-led coordination with specialized worker agents - Domain-Driven Design - Clean bounded contexts with separation of concerns - Plugin Architecture - Microkernel pattern for extensibility - MCP-First API - Consistent inte","readmeSource":"ruflo/v3/README.md","product":"key:ruflo:v3-monorepo","canonical":"@claude-flow/v3-monorepo","tier":"T0","t":["v3-monorepo","monorepo","claude","flow","modular","coordination"],"s":["claude-flow","claude","flow","v3-monorepo","monorepo"]},
+{"id":"@claude-flow/watermark","family":"watermark","kind":"npm","store":"ruflo","version":"0.2.0","description":"SynthID-Text-style LLM text watermarking (generation + detection) as a WASM module — Tournament / non-distortionary / distortion-free schemes, indel-robust and short-text detectors. Node + browser. No watermark-removal tooling.","keywords":["eu-ai-act","llm","provenance","synthid","wasm","watermark"],"source":"ruflo/v3/@claude-flow/watermark/package.json","sourceSha256":"cca3fa67c7ab703759d04157c5e39a80d7a5537d283f92fb960887b42ed6b9be","variants":[],"readme":"SynthID-Text-style LLM text watermarking (generation + detection) as a WebAssembly module — no native addon, runs in Node. A watermark rides the tie-break randomness among *already-plausible* tokens: it never injects an out-of-distribution word, costs no extra tokens, and is detectable only by a holder of the secret key. Built for EU AI Act AI-content marking and for measuring watermark robustness. Backed by the Rust crate [ruflo-watermark](https://github.com/ruvnet/ruflo/tree/main/v3/crates/ruflo-watermark). The module is model-agnostic: at each step you pass the model's candidate token ids and their probabilities (typically a top-k slice); the watermarker picks which candidate to emit. - n","readmeSource":"ruflo/v3/@claude-flow/watermark/README.md","product":"key:ruflo:watermark","canonical":"@claude-flow/watermark","tier":"T0","t":["watermark","synthid-text-style","synthid","text","style","generation","detection","tournament","non-distortionary","non","distortionary","distortion-free","distortion","free","scheme","indel-robust","indel","robust","short-text","short","detector","browser","watermark-removal","removal","eu-ai-act","act","provenance"],"s":["claude-flow","claude","flow","watermark","eu-ai-act","act","provenance","synthid"]},
+{"id":"@cognitum/gate","family":"gate","kind":"npm","store":"ruvector","version":"0.1.0","description":"Browser and Node.js coherence gate for AI agent safety - real-time permit/defer/deny decisions in microseconds","keywords":["agent","ai","audit","claude","coherence","llm","permission","safety","wasm","webassembly"],"source":"ruvector/npm/packages/cognitum-gate-wasm/package.json","sourceSha256":"b496fe0cb44cc6ed1c1beb8bf36224779f92e107f546df3fb97ee7f336bb9a0a","variants":[],"readme":"Browser and Node.js coherence gate for AI agent safety The Cognitum Gate is a high-performance WASM-based coherence verification system designed to bring real-time safety guarantees to AI agent operations. Whether you're building autonomous agents in the browser or orchestrating complex workflows on Node.js, this package provides cryptographically-verifiable permit/defer/deny decisions in microseconds. Every action your agent considers passes through the gate, receiving an immediate verdict backed by witness receipts that create an immutable audit trail. Unlike traditional attention mechanisms that weight tokens by relevance, the coherence gate transforms attention into a permission system.","readmeSource":"ruvector/npm/packages/cognitum-gate-wasm/README.md","product":"key:ruvector:gate","canonical":"@cognitum/gate","tier":"T0","t":["gate","browser","coherence","safety","real","permit","defer","deny","decision","microsecond","audit","claude","permission"],"s":["cognitum","gate","audit","claude","coherence","permission","safety"]},
+{"id":"@daa/compute-wasm","family":"compute","kind":"npm","store":"daa","version":"0.1.0","description":"Decentralized AI training framework with WebAssembly support for browsers","keywords":["ai","blockchain","decentralized","federated-learning","machine-learning","p2p","rust","wasm","webassembly"],"source":"daa/daa-compute/package.json","sourceSha256":"b16a9eed6c223c53157bc0371929e6d75340b8b80ce40a7d20f42f0e3b3e5474","variants":[],"readme":"A high-performance peer-to-peer communication layer for decentralized AI training, built with Rust and libp2p. This implementation enables distributed gradient sharing across cloud servers, edge devices, and browser nodes. - libp2p Swarm: Full implementation with Kademlia DHT and Gossipsub protocols - Multi-Transport Support: TCP, WebSocket, and WebRTC (for browsers) - NAT Traversal: STUN/TURN support with automatic hole punching - Discovery Methods: Bootstrap nodes, mDNS, DHT, and capability-based filtering - All-Reduce Algorithms: Ring, Tree, Butterfly, and Hierarchical implementations - Compression: Multiple methods including Zstandard, LZ4, Snappy, and int8 quantization - Fault Tolerance","readmeSource":"daa/daa-compute/README.md","product":"key:daa:compute","canonical":"@daa/compute-wasm","tier":"T1","t":["compute-wasm","compute","decentraliz","train","browser","blockchain","federated-learning","federat","learn","machine-learning","machine","p2p"],"s":["daa","compute-wasm","compute","blockchain","decentraliz","federated-learning","federat","learn","machine-learning","machine","p2p"]},
+{"id":"@daa/orchestrator","family":"orchestrator","kind":"npm","store":"daa","version":"0.2.1","description":"Node.js bindings for DAA Orchestrator - autonomous agent orchestration with MRAP loop, workflow engine, rules engine, and token economy","keywords":["autonomous-agents","daa","napi-rs","orchestrator","rules-engine","rust","token-economy","workflow"],"source":"daa/daa-orchestrator/daa-napi/package.json","sourceSha256":"c1c3550ee4137b3f33ebc4f2618cf78048e573cc536b1bb20bdfba0934fde1e2","variants":[],"readme":"High-performance Node.js bindings for the DAA (Decentralized Autonomous Agents) Orchestrator, built with NAPI-rs. - MRAP Loop: Monitor, Reason, Act, Plan autonomous loop for continuous operation - Workflow Engine: Create and execute complex multi-step workflows - Rules Engine: Define and evaluate business rules and policies - Economy Manager: Token management, accounts, and trading operations - High Performance: Native Rust performance with zero-copy data transfer - Type Safe: Auto-generated TypeScript definitions - Cross-Platform: Support for Linux, macOS, and Windows The orchestrator provides an autonomous loop that continuously monitors system state, reasons about decisions, acts on them,","readmeSource":"daa/daa-orchestrator/daa-napi/README.md","product":"key:daa:orchestrator","canonical":"@daa/orchestrator","tier":"T1","t":["orchestrator","daa","autonomous","orchestration","mrap","loop","workflow","rule","token","economy","autonomous-agents","rules-engine","token-economy"],"s":["daa","orchestrator","autonomous-agents","autonomous","rules-engine","rule","token-economy","token","economy","workflow"]},
+{"id":"@daa/qudag-native","family":"qudag","kind":"npm","store":"daa","version":"0.1.0","description":"Native Node.js bindings for QuDAG quantum-resistant cryptography","keywords":["blake3","cryptography","ml-dsa","ml-kem","napi","napi-rs","native-addon","nodejs","post-quantum","quantum-resistant","qudag"],"source":"daa/qudag/qudag-napi/package.json","sourceSha256":"e55c838aff8f1afced43cbb7bab03c438229d1b9dd33f7a27fd613b548801f0e","variants":["qudag-napi"],"readme":"Native Node.js bindings for QuDAG quantum-resistant cryptography High-performance, quantum-resistant cryptographic operations for Node.js applications using NAPI-rs bindings to the Rust-based QuDAG library. - 🔐 Post-Quantum Cryptography: ML-DSA (Dilithium) signatures, ML-KEM (Kyber) key exchange - ⚡ Native Performance: Rust implementation with zero-overhead NAPI bindings - 🌍 Cross-Platform: Linux (x64, ARM64), macOS (Intel, Apple Silicon), Windows (x64) - 🛡️ Secure by Default: Quantum-resistant algorithms approved by NIST - 🔒 Static Linking: MUSL-based builds for maximum portability - 📦 Easy Integration: Drop-in replacement for traditional crypto operations - Node.js: 18.x, 20.x, or 22.","readmeSource":"daa/qudag/qudag-napi/README.md","product":"repo:qudag","canonical":"@daa/qudag-native","tier":"T1","t":["qudag-native","qudag","quantum-resistant","quantum","resistant","cryptography","blake3","ml-dsa","dsa","ml-kem","kem","native-addon","addon","post-quantum","post"],"s":["daa","qudag-native","qudag","blake3","cryptography","ml-dsa","dsa","ml-kem","kem","native-addon","addon","post-quantum","post","quantum","quantum-resistant","resistant"]},
+{"id":"@dream-machine/compile","family":"compile","kind":"npm","store":"dream-machine","version":"0.1.0","description":"Compile a dream.config into the full Dream Machine nightly routine prompt — one source, many target repos.","keywords":["config","dream-machine","prompt-compiler","routine"],"source":"dream-machine/packages/compile/package.json","sourceSha256":"8d555fe68fb43a64d1f3ed697553e39ac886d32bb8d44559f5e415cd44a74fda","variants":[],"product":"key:dream-machine:compile","canonical":"@dream-machine/compile","tier":null,"t":["compile","dream.config","full","dream","machine","nightly","routine","prompt","source","many","target","repo","config","dream-machine","prompt-compiler","compiler"],"s":["dream-machine","dream","machine","compile","config","prompt-compiler","prompt","compiler","routine"]},
+{"id":"@dream-machine/edge-contracts","family":"edge-contracts","kind":"npm","store":"dream-machine","version":"0.1.0","description":"Bounded edge-v1-prototype ticket codec and Ed25519 verification. No actuator authority.","keywords":[],"source":"dream-machine/packages/edge-contracts/package.json","sourceSha256":"cd24f1a09e832605dc5fa6a99ea8884b9809e86984b45effd9cb10b71a64818d","variants":[],"readme":"Status: edge-v1-prototype. No hardware, physical output, or complete Edge Contract v1 conformance is claimed. This private workspace package provides a small, inspectable boundary for simulated development without new dependencies. Input is an exact CueTicketV1 JSON object or at most 768 CBOR bytes. Output is canonical CBOR, the exact JSON projection, or signature validity. Malformed data fails closed with TicketFormatError; verification returns false. The codec follows [the existing numeric key table](../../docs/contracts/edge-v1/README.md#6-cueticketv1) and [JSON ticket schema](../../docs/contracts/edge-v1/safety-decision.schema.json). Keys 0 through 22 form the Ed25519 message. Key 23 con","readmeSource":"dream-machine/packages/edge-contracts/README.md","product":"key:dream-machine:edge-contracts","canonical":"@dream-machine/edge-contracts","tier":null,"t":["edge-contracts","edge","contract","bound","edge-v1-prototype","prototype","ticket","codec","ed25519","verification","actuator","authority"],"s":["dream-machine","dream","machine","edge-contracts","edge","contract"]},
+{"id":"@dream-machine/edge-sim","family":"edge-sim","kind":"npm","store":"dream-machine","version":"0.1.0","description":"Deterministic, nonactuating Edge v1 safety simulator. Not firmware or hardware evidence.","keywords":[],"source":"dream-machine/packages/edge-sim/package.json","sourceSha256":"b460200d662a653ab7de23a52bb728da8aeeb055ef1f8c3d1dd6227c683ceede","variants":[],"readme":"@dream-machine/edge-sim is an executable, deterministic, nonactuating simulator. It exercises the Edge Contract v1 ticket parser and authorization invariants without hardware, network, serial, filesystem, wall clock, or real signing keys. Its virtualGateOpen is a boolean in memory, never an output device. 1. A uint64 microsecond virtual clock with deterministic timer boundaries. 2. Explicit DISARMED → OBSERVE → ARMED simulated physical events, latched MUTED and FAULT, simulated maintenance, and terminal disposal. 3. Synthetic radar observations, packet dropout and multiple occupants. 4. Canonical CBOR decoding and Ed25519 verification through @dream-machine/edge-contracts, using publicly doc","readmeSource":"dream-machine/packages/edge-sim/README.md","product":"key:dream-machine:edge-sim","canonical":"@dream-machine/edge-sim","tier":null,"t":["edge-sim","edge","sim","deterministic","nonactuat","safety","simulator","firmware","hardware","evidence"],"s":["dream-machine","dream","machine","edge-sim","edge","sim"]},
+{"id":"@dream-machine/ledger","family":"ledger","kind":"npm","store":"dream-machine","version":"0.1.0","description":"The Dream Machine ledger — parse/append/verify the 10-column LEDGER.md and compute nightly learning signals.","keywords":["dream-machine","durable-memory","learning-signals","ledger"],"source":"dream-machine/packages/ledger/package.json","sourceSha256":"7801797cab9492e6ce96c84ebd4582dd7ed8cd577e929ca6ceaaaa6dd40b2bf8","variants":[],"product":"key:dream-machine:ledger","canonical":"@dream-machine/ledger","tier":null,"t":["ledger","dream","machine","parse","append","verify","10-column","column","ledger.md","compute","nightly","learn","signal","dream-machine","durable-memory","durable","memory","learning-signals"],"s":["dream-machine","dream","machine","ledger","durable-memory","durable","memory","learning-signals","learn","signal"]},
+{"id":"@dream-machine/memory","family":"memory","kind":"npm","store":"dream-machine","version":"0.1.0","description":"Bounded deterministic keyword memory with atomic flat-file persistence. Optional RuVector module probe; the RVF adapter is not implemented.","keywords":["dream-machine","memory","optional","ruvector","rvf","semantic-search","wasm"],"source":"dream-machine/packages/memory/package.json","sourceSha256":"4532c63427d1bfd26a0de02121b9be08cef4ac4aa9136925d4b9b86a3a78d530","variants":[],"readme":"This package is an implemented, bounded keyword store. auto may detect an optional RuVector module, but it still reports backend: 'flat-file'. An explicit ruvector-rvf request fails until a real adapter passes its persistence and query contract. Module availability is not evidence of vector retrieval. inMemory: true never accesses a file. Omitting a path defaults to in memory; explicit inMemory: false requires a path. Existing valid array files need no migration. New writes use compact JSON and do not retain formatting. all() and recall() return detached records. remember() captures a detached snapshot before waiting on earlier calls. The four required fields are strings; witness and detail","readmeSource":"dream-machine/packages/memory/README.md","product":"key:dream-machine:memory","canonical":"@dream-machine/memory","tier":null,"t":["memory","bound","deterministic","keyword","atomic","flat-file","flat","file","persistence","optional","probe","rvf","adapter","implement","dream-machine","dream","machine","semantic-search","semantic","search"],"s":["dream-machine","dream","machine","memory","optional","rvf","semantic-search","semantic","search"]},
+{"id":"@dream-machine/schedule","family":"schedule","kind":"npm","store":"dream-machine","version":"0.1.0","description":"Emit the cloud /schedule routine body for a Dream Machine target (compile → paste, or automate once an API is available).","keywords":["cloud-agent","cron","dream-machine","routine","schedule"],"source":"dream-machine/packages/schedule/package.json","sourceSha256":"a8b4ccfa6a66c656f2eb1e912165efe571fa9969b39252420a5a4eb7923e1e64","variants":[],"product":"key:dream-machine:schedule","canonical":"@dream-machine/schedule","tier":null,"t":["schedule","emit","cloud","routine","body","dream","machine","target","compile","paste","automate","once","available","cloud-agent","cron","dream-machine"],"s":["dream-machine","dream","machine","schedule","cloud-agent","cloud","cron","routine"]},
+{"id":"@dream-machine/witness","family":"witness","kind":"npm","store":"dream-machine","version":"0.1.0","description":"Witness provenance for the Dream Machine — sha256(sha256(report) + commit) stamp & 5-step verify.","keywords":["dream-machine","provenance","reproducibility","sha256","witness"],"source":"dream-machine/packages/witness/package.json","sourceSha256":"a45762c4623897b5251c36b8735717dba52be36bb13fc94772c6741fa57ed010","variants":[],"product":"key:dream-machine:witness","canonical":"@dream-machine/witness","tier":null,"t":["witness","provenance","dream","machine","sha256","report","commit","stamp","5-step","step","verify","dream-machine","reproducibility"],"s":["dream-machine","dream","machine","witness","provenance","reproducibility","sha256"]},
+{"id":"@latentmesh/evolve-metaharness","family":"evolve-metaharness","kind":"npm","store":"latentmesh","version":"0.1.0","description":"MetaHarness suite for the ADR-018 Darwin topology loop: runs the latentmesh-evolve binary and verifies its evidence-labelled receipt against ADR-006's acceptance bounds. Deterministic simulation only — no live-agent claims.","keywords":[],"source":"latentmesh/harness/evolve/package.json","sourceSha256":"a0f45307ecc1376cad883a764ddff8bb5e03f38288d6b09ba93214f102ab1b4a","variants":[],"product":"key:latentmesh:evolve-metaharness","canonical":"@latentmesh/evolve-metaharness","tier":null,"t":["evolve-metaharness","evolve","metaharness","suite","adr-018","adr","018","darwin","topology","loop","latentmesh-evolve","latentmesh","binary","verify","evidence-labelled","evidence","label","receipt","against","adr-006","006","acceptance","bound","deterministic","simulation","only","live-agent","live","claim"],"s":["latentmesh","evolve-metaharness","evolve","metaharness"]},
+{"id":"@latentmesh/integration-metaharness","family":"integration-metaharness","kind":"npm","store":"latentmesh","version":"0.1.0","description":"ADR-022 MetaHarness suite for the meshtastic/agentbbs/cognitum integration wave (ADRs 019-021): three e2e loopback suites, optional MetaHarness gates, and a Darwin-loop-style adapter-parameter optimizer, each emitting an evidence-labelled JSON receipt. Deterministic simulation only — no hardware, no live peers, no credentials.","keywords":[],"source":"latentmesh/harness/integration/package.json","sourceSha256":"39e5a180a2127e14257f5d2e56c89aea81318359219e178298cbfc2abe8bd6fc","variants":[],"product":"key:latentmesh:integration-metaharness","canonical":"@latentmesh/integration-metaharness","tier":null,"t":["integration-metaharness","integration","metaharness","adr-022","adr","022","suite","meshtastic","agentbb","cognitum","wave","adrs","019-021","019","021","three","e2e","loopback","optional","gate","darwin-loop-style","darwin","loop","style","adapter-parameter","adapter","parameter","optimizer","emitt","evidence-labelled","evidence","label","json","receipt","deterministic","simulation","only","hardware","live","peer","credential"],"s":["latentmesh","integration-metaharness","integration","metaharness"]},
+{"id":"@latentmesh/representation-metaharness","family":"representation-metaharness","kind":"npm","store":"latentmesh","version":"0.1.0","description":"ADR-046 MetaHarness suite for the eighth policy surface, representationPolicy: scores channel/admission/controls/alpha/costWeights genomes by measured causal utility density over receipts already committed to crates/latentmesh-runtime/receipts/. No simulator — fitness comes from real measurements, and an unmeasured channel is structurally ineligible rather than optimistically explored.","keywords":[],"source":"latentmesh/harness/representation/package.json","sourceSha256":"79a3d6daff9e1d11e1323f8db772226b27a0a30c9ff1b6e9314e8fd60e4a3c6c","variants":[],"readme":"This harness evolves representationPolicy — [ADR-046](../../docs/adr/046-representation-metaharness.md)'s eighth policy surface — while freezing the model, the receipts, the promotion rule, and the pre-registered significance level. Its fitness is not simulated. harness/air labels its results simulated because it runs a channel simulator, and that label is correct there. Here it would be a lie: every gain, cost and admission decision is read out of receipts already committed to crates/latentmesh-runtime/receipts/, so the evidence label is measured-committed-receipts, and lib/verify.mjs refuses a receipt that claims anything stronger — or anything weaker. npm run benchmark writes artifacts/re","readmeSource":"latentmesh/harness/representation/README.md","product":"key:latentmesh:representation-metaharness","canonical":"@latentmesh/representation-metaharness","tier":null,"t":["representation-metaharness","representation","metaharness","adr-046","adr","046","suite","eighth","policy","surface","representationpolicy","score","channel","admission","control","alpha","costweight","genome","measur","causal","utility","density","receipt","already","commit","crate","latentmesh-runtime","latentmesh","runtime","simulator","fitness","come","real","measurement","unmeasur","structurally","ineligible","rather","than","optimistically","explor"],"s":["latentmesh","representation-metaharness","representation","metaharness"]},
+{"id":"@marketing/ai-swarms","family":"ai-swarms","kind":"npm","store":"marketing","version":"1.0.0","description":"AI Marketing Swarms - 15-agent hierarchical mesh for autonomous media buying","keywords":["advertising","agents","ai","automation","marketing","media-buying","swarm"],"source":"marketing/package.json","sourceSha256":"8c92fc5a9d450beab5dc973ccf0ea4e7e0e26accf4781aa0f267a0f9a2bf9d5c","variants":[],"readme":"A 15-agent AI swarm system that handles media buying, creative optimization, and campaign management across multiple advertising platforms. Built with Claude-Flow V3 for enterprise-grade reliability. Imagine having a team of 15 AI specialists working around the clock on your marketing campaigns: - One agent optimizes your bids in real-time - Another predicts which creatives will fatigue before they do - Another detects fraud and wasted spend instantly - And 12 more specialists handling everything from attribution to cross-platform sync That's what AI Marketing Swarms does. It's like hiring an elite marketing team that never sleeps, never makes emotional decisions, and learns from every dolla","readmeSource":"marketing/README.md","product":"repo:marketing","canonical":"@marketing/ai-swarms","tier":"T2","t":["ai-swarms","swarm","market","15-agent","hierarchical","mesh","autonomous","media","buying","advertis","automation","media-buying"],"s":["market","ai-swarms","swarm","advertis","automation","media-buying","media","buying"]},
+{"id":"@medai/backend","family":"backend","kind":"npm","store":"agentic-flow","version":"1.0.0","description":"Medical AI Analysis Backend with Anti-Hallucination Features","keywords":["agentdb","ai","analysis","anti-hallucination","healthcare","medical"],"source":"agentic-flow/src/utils/package.json","sourceSha256":"7418fb05b27d05357fa1fbbba919cc1bf0466afe4544c4202843299e0189e3bb","variants":[],"product":"key:agentic-flow:backend","canonical":"@medai/backend","tier":"T1","t":["backend","medical","analysis","anti-hallucination","anti","hallucination","feature","agentdb","healthcare"],"s":["medai","backend","agentdb","analysis","anti-hallucination","anti","hallucination","healthcare","medical"]},
+{"id":"@metaharness/hackerone","family":"hackerone","kind":"npm","store":"hackerone","version":"0.1.0","description":"🛡️ HackerOne meta-harness for AI agents — defender-side triage (dedupe, CVSS scoring, scope-check) AND researcher-side recon (scope analysis, OWASP/CWE mapping, report formatting). Safety-contained: no live API by default, mock-mode for $0 CI.","keywords":["agent-harness","ai-agents","bug-bounty","cvss","cwe","defensive-security","hackerone","metaharness","owasp","security","triage","vdp","vulnerability-management"],"source":"hackerone/package.json","sourceSha256":"45b8ea1467c976ad6737344e23166a703e7d765f876aa71676858e460b36547c","variants":[],"readme":"@metaharness/hackerone is a deterministic, offline-first harness for the two sides of a HackerOne engagement: No ML, no embeddings, no external calls beyond the HackerOne API itself. Every classification is a transparent rule the SOC/researcher can audit and reproduce. The harness is read-only and capability-contained in v0.1. The defaults below are hard-frozen in src/safety.ts; nothing in this codebase relaxes them: Anyone reviewing src/safety.ts should be able to answer in one sentence: \"What's the worst this harness can do?\" → \"Issue read-only HackerOne API calls scoped to one program, with a hard rate limit, with the key fetched transiently and never logged.\" The assets subcommand emits","readmeSource":"hackerone/README.md","product":"repo:hackerone","canonical":"@metaharness/hackerone","tier":null,"t":["hackerone","meta-harness","meta","harness","defender-side","defender","side","triage","dedupe","cvss","scor","scope-check","scope","check","researcher-side","researcher","recon","analysis","owasp","cwe","mapp","report","formatt","safety-contained","safety","contain","live","default","mock-mode","mock","mode","agent-harness","ai-agents","bug-bounty","bug","bounty","defensive-security","defensive","security","metaharness","vdp","vulnerability-management","vulnerability","management"],"s":["metaharness","hackerone","agent-harness","harness","ai-agents","bug-bounty","bug","bounty","cvss","cwe","defensive-security","defensive","security","owasp","triage","vdp","vulnerability-management","vulnerability","management"]},
+{"id":"@midstream/lean-agentic","family":"lean-agentic","kind":"npm","store":"midstream","version":"1.0.0","description":"TypeScript/JavaScript client for Lean Agentic Learning System - Revolutionary stream learning with formal reasoning and autonomous agents","keywords":["agentic","agents","ai","formal-verification","knowledge-graph","lean","learning","llm","online-learning","real-time","streaming","theorem-proving"],"source":"midstream/lean-agentic-js/package.json","sourceSha256":"3a2314c8a7c7cdffbd6f9f139ad1917ec315e4d6f096e67f0f8c85c5802fe137","variants":["lean-agentic-wasm"],"readme":"A revolutionary learning framework combining formal reasoning, agentic AI, and stream learning for real-time adaptation. - 🎯 Formal Reasoning - Lean-style theorem proving for verified knowledge - 🤖 Agentic AI - Autonomous decision-making with Plan-Act-Observe-Learn loops - 📊 Stream Learning - Real-time online adaptation from data streams - 🧠 Knowledge Graph - Dynamic knowledge representation and evolution - ⚡ Real-Time Processing - Low-latency stream processing - 🔒 Type Safety - Full TypeScript support with comprehensive types","readmeSource":"midstream/lean-agentic-js/README.md","product":"key:midstream:lean-agentic","canonical":"@midstream/lean-agentic","tier":"T1","t":["lean-agentic","lean","client","learn","revolutionary","stream","formal","reason","autonomous","formal-verification","verification","knowledge-graph","knowledge","graph","online-learning","online","real","theorem-proving","theorem","prov"],"s":["midstream","lean-agentic","lean","formal-verification","formal","verification","knowledge-graph","knowledge","graph","learn","online-learning","online","real","stream","theorem-proving","theorem","prov"]},
+{"id":"@prime/ml-napi","family":"ml","kind":"npm","store":"daa","version":"0.2.1","description":"Node.js bindings for Prime ML distributed federated learning framework","keywords":["distributed-ml","federated-learning","machine-learning","napi","pytorch","rust","tensorflow","zero-copy"],"source":"daa/prime-rust/prime-napi/package.json","sourceSha256":"02d923c2bf9c7c08162c7ab6bcc8d9ab5c203118b2a34c66cf919e5b8eb9d14f","variants":[],"readme":"High-performance Node.js bindings for the Prime ML distributed federated learning framework, built with Rust and NAPI-rs. - Zero-Copy Operations: Direct memory access for tensor data using napi::Buffer - High Performance: Rust-powered computation with minimal overhead - Federated Learning: Built-in support for distributed training coordination - Parallel Gradient Aggregation: Multiple aggregation strategies (FedAvg, Secure Aggregation, Trimmed Mean, Krum) - Type Safety: Full TypeScript type definitions - Cross-Platform: Supports Linux, macOS, Windows (x64 and ARM64)","readmeSource":"daa/prime-rust/prime-napi/README.md","product":"key:daa:ml","canonical":"@prime/ml-napi","tier":"T1","t":["ml-napi","prime","distribut","federat","learn","distributed-ml","federated-learning","machine-learning","machine","pytorch","tensorflow","zero-copy","zero","copy"],"s":["prime","ml-napi","distributed-ml","distribut","federated-learning","federat","learn","machine-learning","machine","pytorch","tensorflow","zero-copy","zero","copy"]},
+{"id":"@qudag/ruvector-index","family":"index","kind":"npm","store":"qudag","version":"1.0.0","description":"Native RuVector disposable retrieval index","keywords":[],"source":"qudag/integrations/ruvector-index/package.json","sourceSha256":"6a6b975ff3d25952e9df3e121890622b826a3d3f58e860c27fcc05a1b7a399bc","variants":[],"readme":"Disposable native retrieval index for the federation gateway. The gateway's SQLite ledger remains the source of truth: rebuild this index from its persisted vectors after restart. It is not a consensus or authorization mechanism. search returns {id, score, metadata}; score is cosine distance, lower is closer: 0 identical direction, 1 orthogonal, 2 opposite. Input vectors are normalized before conversion to native Float32. Finite nonzero vectors of exactly the configured dimension are required. IDs are immutable and duplicates reject, including concurrent insertions. Capacity and k are bounded. Metadata is cloned on both insertion and retrieval. textEmbedding is deterministic normalized lexic","readmeSource":"qudag/integrations/ruvector-index/README.md","product":"key:qudag:index","canonical":"@qudag/ruvector-index","tier":"T1","t":["ruvector-index","index","disposable","retrieval"],"s":["qudag","ruvector-index","index"]},
+{"id":"@ruflo/ai-team-service","family":"ai-team-service","kind":"npm","store":"ruflo","version":"0.1.6","description":"Multi-tenant MCP control plane and Claude plugin for coordinated AI teams.","keywords":[],"source":"ruflo/plugins/ruflo-ai-team/package.json","sourceSha256":"22cff6c3c290f8cdb0c1f4c7fcd5de5254fc981fac06935f1169b9bbb0180cb2","variants":[],"readme":"RuFlo AI Team is a separate, multi-tenant MCP service and Claude plugin. It turns a reviewed goal into explicit team, run, task, memory, and evidence records without exposing raw operator tools or credentials. The public v0.1 surface coordinates work; it does not silently send messages, deploy software, execute shell commands, make purchases, or approve consequential actions. Claude Code and Cowork agents use the service as a shared control plane while the user remains the authority for external effects. - OAuth 2.1 resource server with RFC 9728 discovery and issuer/audience/scope verification. - Tenant identity derived only from verified token claims; no tool accepts a tenant ID. - Firestor","readmeSource":"ruflo/plugins/ruflo-ai-team/README.md","product":"key:ruflo:ai-team-service","canonical":"@ruflo/ai-team-service","tier":"T0","t":["ai-team-service","team","service","multi-tenant","multi","tenant","mcp","control","plane","claude","plugin","coordinat"],"s":["ruflo","ai-team-service","team","service"]},
+{"id":"@ruflo/chatgpt-federation","family":"chatgpt-federation","kind":"npm","store":"ruflo","version":"0.1.0","description":"ChatGPT Federation publisher — signs ruflo swarm events with its own Nostr key and publishes them over its own NIP-42 authenticated relay connection. The key lives in this service and never leaves it.","keywords":[],"source":"ruflo/plugins/ruflo-chatgpt-federation/package.json","sourceSha256":"ac39775f32dd019960590a44214eb0503e577d21c17852a36c622989386d1192","variants":[],"readme":"The ChatGPT Federation connector's publisher. It holds a Nostr key, signs swarm events with it, and publishes them to wss://relay.ruv.io over a connection it authenticated itself. buzz-relay refuses any EVENT whose pubkey differs from the NIP-42 identity that authenticated the connection: So \"sign it here, let the gateway relay it for you\" cannot work — not as a policy choice, as a protocol one. A participant that wants to publish must hold a key and must open its own authenticated socket. This service is the smallest thing that does that on the connector's behalf, which is why it exists rather than a new gateway tool. It is also why the x.ruv.io gateway is uninvolved here: it does not sign","readmeSource":"ruflo/plugins/ruflo-chatgpt-federation/README.md","product":"key:ruflo:chatgpt-federation","canonical":"@ruflo/chatgpt-federation","tier":"T0","t":["chatgpt-federation","chatgpt","federation","publisher","sign","ruflo","swarm","event","own","nostr","key","publish","them","nip-42","nip","authenticat","relay","connection","live","service","never","leave"],"s":["ruflo","chatgpt-federation","chatgpt","federation"]},
+{"id":"@ruflo/research","family":"research","kind":"npm","store":"ruflo","version":"0.1.0","description":"RuFlo Research — Goal-Oriented Action Planning UI for autonomous AI research workflows","keywords":[],"source":"ruflo/v3/goal_ui/package.json","sourceSha256":"f7e7a83b7efa3b5a1ec5238bc62090a046794c45985ca440c590562ca54fb550","variants":[],"readme":"Turn plain-English research goals into executable agent plans. RuFlo Research applies classic Goal-Oriented Action Planning (GOAP) — A* search through a state space of actions with preconditions and effects — to autonomous AI research, then dispatches the work to live agents you can inspect in real time. For the embeddable widget: The widget exposes a global window.RufloResearchWidget with init(containerId) and version for programmatic control. See [docs/WIDGET-INTEGRATION.md](docs/WIDGET-INTEGRATION.md) for the full integration guide. React 18 · TypeScript 5 · Vite 5 · Tailwind 3 · shadcn/ui · Radix UI · React Query · React Router · Supabase Edge Functions · GOAP A* planner Hosted on Netlif","readmeSource":"ruflo/v3/goal_ui/README.md","product":"key:ruflo:research","canonical":"@ruflo/research","tier":"T0","t":["research","ruflo","goal-oriented","goal","orient","action","plann","autonomous","workflow"],"s":["ruflo","research"]},
+{"id":"@ruflo/x-gateway","family":"x-gateway","kind":"npm","store":"ruflo","version":"0.8.1","description":"x.ruv.io gateway — MCP server for ruflo swarm federation + claims over a signed, membership-gated Nostr relay. Exposes /mcp tools and ruv:// resources.","keywords":[],"source":"ruflo/plugins/ruflo-x-gateway/package.json","sourceSha256":"063d3681254a78a1350b336e6d672157b2f2d743a47ea7cdf196a59fed20d38d","variants":[],"readme":"MCP gateway for the open ruflo swarm federation. Coordination rides an open, membership-gated, signed Nostr relay — every message is a signed Nostr event (verifiable authorship), and the relay admits members + NIP-42 auth (security). - GET / — service info - GET /health — health probe - POST /mcp — full compatibility MCP surface (Streamable HTTP, stateless) - POST /chatgpt/mcp — directory-safe ChatGPT surface - POST /claude/mcp — directory-safe Claude connector surface - GET /.well-known/oauth-protected-resource/{mcp|chatgpt/mcp|claude/mcp} — RFC 9728 OAuth discovery - GET /privacy, /terms, /support — public legal and support pages The two directory-safe endpoints expose twelve tools with ex","readmeSource":"ruflo/plugins/ruflo-x-gateway/README.md","product":"key:ruflo:x-gateway","canonical":"@ruflo/x-gateway","tier":"T0","t":["x-gateway","gateway","x.ruv.io","mcp","server","ruflo","swarm","federation","claim","sign","membership-gated","membership","gated","nostr","relay","expos","resource"],"s":["ruflo","x-gateway","gateway"]},
+{"id":"@ruv/cognitum","family":"cognitum","kind":"npm","store":"cognitum-cogs","version":"0.1.0","description":"Node.js native bindings for Cognitum ASIC simulator","keywords":["asic","cognitum","hardware","napi","napi-rs","native","rust","simulator","wasm","webassembly"],"source":"cognitum-cogs/cognitum-sim/cognitum-napi/package.json","sourceSha256":"e383c6dcdec219832e43da76416fee2c2740cfe8527a651ac66cb9386e968fc3","variants":["@ruv/cognitum-wasm","cognitum","cognitum-cli","cognitum-core","cognitum-napi","cognitum-wasm"],"readme":"High-performance Node.js native bindings for the Cognitum ASIC simulator using NAPI-RS. - Native performance - Direct Rust-to-Node.js integration with zero overhead - Async/await support - Non-blocking simulation using Tokio runtime - TypeScript support - Auto-generated type definitions - Cross-platform - Pre-built binaries for all major platforms - Thread-safe - Safe concurrent access using RwLock Pre-built binaries are available for: - Linux: x64, ARM64, ARM7 (glibc and musl) - macOS: x64, ARM64 (Apple Silicon) - Windows: x64, ARM64 - Execution speed: 10-20M cycles/second (native) - Me","readmeSource":"cognitum-cogs/cognitum-sim/cognitum-napi/README.md","product":"key:cognitum-cogs:cognitum","canonical":"@ruv/cognitum","tier":null,"t":["cognitum","asic","simulator","hardware"],"s":["cognitum","asic","hardware","simulator"]},
+{"id":"@ruv/ruv-swarm-wasm","family":"ruv-swarm","kind":"npm","store":"synaptic-mesh","version":"1.0.5","description":"WebAssembly bindings for ruv-swarm neural network orchestration","keywords":[],"source":"synaptic-mesh/src/js/ruv-swarm/wasm-unified/package.json","sourceSha256":"1bfdf89a11af4f316ce2762ec6289d5fbf309cb9b6464887335ab2f7a77d8329","variants":[],"readme":"High-performance WebAssembly neural network orchestration with SIMD optimization for browser and Node.js environments. ruv-swarm-wasm is a cutting-edge WebAssembly implementation of the ruv-swarm neural network orchestration engine, specifically designed for maximum performance in web browsers and Node.js environments. By leveraging SIMD (Single Instruction, Multiple Data) optimizations and WebAssembly's near-native performance, this crate delivers unprecedented speed for neural network operations in JavaScript environments. - SIMD-accelerated operations: 2-4x performance improvement over scalar implementations - Near-native performance: WebAssembly execution with optimized memory management","readmeSource":"synaptic-mesh/src/js/ruv-swarm/wasm-unified/README.md","product":"key:synaptic-mesh:ruv-swarm","canonical":"@ruv/ruv-swarm-wasm","tier":"T1","t":["ruv-swarm-wasm","swarm","ruv-swarm","neural","network","orchestration"],"s":["ruv-swarm-wasm","swarm","ruv-swarm"]},
+{"id":"@ruv/rvcsi","family":"rvcsi","kind":"npm","store":"rvcsi","version":"0.3.1","description":"rvCSI — edge RF sensing runtime: Node.js bindings (napi-rs) over the Rust CSI pipeline (ADR-095, ADR-096)","keywords":["cli","csi","napi","napi-rs","presence","rf-sensing","rvcsi","wifi"],"source":"rvcsi/crates/rvcsi-node/package.json","sourceSha256":"6101b57ffb5470763fb352def32ae4a4fb05024381948b7627c5fcee225563e2","variants":["rvcsi-cli","rvcsi-core","rvcsi-node"],"readme":"Node.js bindings (napi-rs) for rvCSI — the edge RF sensing runtime: ingest WiFi CSI from files / Nexmon dumps, validate and normalize it, run reusable DSP, emit typed presence / motion / quality / anomaly events, and export temporal embeddings to an RF-memory store. See [ADR-095](../../docs/adr/ADR-095-rvcsi-edge-rf-sensing-platform.md) and [ADR-096](../../docs/adr/ADR-096-rvcsi-ffi-crate-layout.md). The native addon is produced from the rvcsi-node Rust crate: (cargo build -p rvcsi-node also compiles the addon as a cdylib; napi build additionally emits the platform loader and .d.ts.) TypeScript types ship in index.d.ts (CsiFrame, CsiWindow, CsiEvent, SourceHealth, CaptureSummary, ValidationS","readmeSource":"rvcsi/crates/rvcsi-node/README.md","product":"repo:rvcsi","canonical":"@ruv/rvcsi","tier":"T1","t":["rvcsi","edge","sens","runtime","csi","pipeline","adr-095","adr","095","adr-096","096","presence","rf-sensing","wifi"],"s":["rvcsi","csi","presence","rf-sensing","sens","wifi"]},
+{"id":"@ruv/sparc-ui","family":"sparc-ui","kind":"npm","store":"sparc","version":"0.1.4","description":"SPARC (Specification, Pseudocode, Architecture, Refinement, and Completion) Framework UI Components","keywords":[],"source":"sparc/ui/package.json","sourceSha256":"212b8eba2af07945b23fb64f0034c97cc3bb06124e83190374dcd63d49e57a7a","variants":[],"readme":"A comprehensive React component library for the SPARC (Specification, Pseudocode, Architecture, Refinement, and Completion) Framework. This library provides a robust set of UI components designed to streamline the development of AI-powered applications. The SPARC UI Framework is built on modern web technologies and best practices, offering: - AI-First Architecture: Deep integration with multiple LLM providers - Component-Driven Design: Modular, reusable components following atomic design principles - Type Safety: Full TypeScript support with comprehensive type definitions - Modern Styling: Built with Tailwind CSS for responsive and customizable designs - Accessibility: ARIA-compliant compone","readmeSource":"sparc/ui/README.md","product":"key:sparc:sparc-ui","canonical":"@ruv/sparc-ui","tier":"T1","t":["sparc-ui","sparc","specification","pseudocode","architecture","refinement","completion","component"],"s":["sparc-ui","sparc"]},
+{"id":"@ruvector/agentic-integration","family":"agentic-integration","kind":"npm","store":"ruvector","version":"1.0.2","description":"Distributed agent coordination for ruvector with claude-flow integration","keywords":["agent-coordination","claude-flow","distributed-systems","mesh-coordination","ruvector","swarm","vector-search"],"source":"ruvector/npm/packages/agentic-integration/package.json","sourceSha256":"c630ffe56510093ef7f09ea9d6d7e262258b76ca9f565e2f7f3ce0cbaef5d198","variants":[],"product":"key:ruvector:agentic-integration","canonical":"@ruvector/agentic-integration","tier":"T0","t":["agentic-integration","integration","distribut","coordination","claude-flow","claude","flow","agent-coordination","distributed-systems","mesh-coordination","mesh","swarm","vector-search","vector","search"],"s":["agentic-integration","integration","agent-coordination","coordination","claude-flow","claude","flow","distributed-systems","distribut","mesh-coordination","mesh","swarm","vector-search","vector","search"]},
+{"id":"@ruvector/agentic-synth","family":"agentic-synth","kind":"npm","store":"ruvector","version":"0.1.7","description":"High-performance synthetic data generator for AI/ML training, RAG systems, and agentic workflows with DSPy.ts, Gemini, OpenRouter, and vector databases","keywords":["agentic-ai","agenticdb","ai-training","claude","cli-tool","context-caching","data-augmentation","data-generation","data-synthesis","dataset-generator","dspy","dspy-ts","edge-cases","event-generation","gemini","gpt","langchain","llm","machine-learning","ml-training","mock-data","natural-language-processing","nlp","nodejs","openrouter","prompt-engineering","rag","retrieval-augmented-generation","ruvector","streaming","structured-data","synthetic-data","synthetic-dataset","test-data","time-series","training-data","training-datasets","typescript","vector-embeddings"],"source":"ruvector/npm/packages/agentic-synth/package.json","sourceSha256":"16003a15d3cc76c6abbfc8fd0984814dcbe5bc28d0b9380ace8aab791616e139","variants":[],"readme":"Powered by Gemini, OpenRouter, and DSPy.ts | 98% Test Coverage | 50+ Production Examples [🎯 Get Started](#-quick-start-5-minutes) • [📚 Examples](#-examples-as-npx-packages) • [📖 Documentation](#-api-reference) • [💬 Community](#-community--support) Training AI models and testing agentic systems requires massive amounts of diverse, high-quality data. Real data is: - 💰 Expensive to collect and curate - 🔒 Privacy-sensitive with compliance risks - 🐌 Slow to generate at scale - ⚠️ Insufficient for edge cases and stress tests - 🔄 Hard to reproduce across environments Agentic-Synth generates unlimited synthetic data tailored to your exact needs with: - ⚡ 10-100x faster than manual creation -","readmeSource":"ruvector/npm/packages/agentic-synth/README.md","product":"key:ruvector:agentic-synth","canonical":"@ruvector/agentic-synth","tier":"T0","t":["agentic-synth","synth","high","synthetic","generator","train","rag","workflow","dspy.t","gemini","openrouter","vector","databas","agentic-ai","agenticdb","ai-training","claude","cli-tool","context-caching","context","cach","data-augmentation","augmentation","data-generation","generation","data-synthesis","synthesis","dataset-generator","dataset","dspy","dspy-ts","edge-cases","edge","cas","event-generation","event","gpt","langchain","machine-learning","machine","learn","ml-training","mock-data","mock","natural-language-processing","natural","language","process","nlp","prompt-engineering","prompt","engineer","retrieval-augmented-generation","retrieval","augment","stream","structured-data","structur","synthetic-data","synthetic-dataset","test-data","test","time-series","sery","training-data","training-datasets","vector-embeddings","embedding"],"s":["agentic-synth","synth","agentic-ai","agenticdb","ai-training","train","claude","cli-tool","context-caching","context","cach","data-augmentation","augmentation","data-generation","generation","data-synthesis","synthesis","dataset-generator","dataset","generator","dspy","dspy-ts","edge-cases","edge","cas","event-generation","event","gemini","gpt","langchain","machine-learning","machine","learn","ml-training","mock-data","mock","natural-language-processing","natural","language","process","nlp","openrouter","prompt-engineering","prompt","engineer","rag","retrieval-augmented-generation","retrieval","augment","stream","structured-data","structur","synthetic-data","synthetic","synthetic-dataset","test-data","test","time-series","sery","training-data","training-datasets","vector-embeddings","vector","embedding"]},
+{"id":"@ruvector/attention","family":"attention","kind":"npm","store":"ruvector","version":"0.1.4","description":"High-performance attention mechanisms for Node.js","keywords":["ai","attention","cgt","coherence","contrastive-learning","deep-learning","flash-attention","gpu","graph-attention","hyperbolic","infonce","llm","machine-learning","mixture-of-experts","moe","multi-head-attention","napi-rs","neural-network","neural-networks","rust","sheaf-attention","simd","transformer","vector-search","wasm","webassembly"],"source":"ruvector/crates/ruvector-attention-node/package.json","sourceSha256":"52f9c89404feafde284e0866e72fc9f91d0b1ea85479d064d9f5f3fb6e7c4cfd","variants":["@ruvector/attention-wasm","ruvector-attention","ruvector-attention-cli","ruvector-attention-node","ruvector-attention-wasm"],"readme":"High-performance attention mechanisms for Node.js, powered by Rust. - Scaled Dot-Product Attention: Classic attention mechanism with optional scaling - Multi-Head Attention: Parallel attention heads for richer representations - Flash Attention: Memory-efficient attention with block-wise computation - Linear Attention: O(N) complexity attention using kernel approximations - Hyperbolic Attention: Attention in hyperbolic space for hierarchical data - Mixture-of-Experts (MoE) Attention: Dynamic expert routing for specialized attention - constructor(dim: number, scale?: number) - compute(query: Float32Array, keys: Float32Array[], values: Float32Array[]): Float32Array - constructor(dim: number, nu","readmeSource":"ruvector/crates/ruvector-attention-node/README.md","product":"key:ruvector:attention","canonical":"@ruvector/attention","tier":"T0","t":["attention","high","mechanism","cgt","coherence","contrastive-learning","contrastive","learn","deep-learning","deep","flash-attention","flash","gpu","graph-attention","graph","hyperbolic","infonce","machine-learning","machine","mixture-of-experts","mixture","expert","moe","multi-head-attention","multi","head","neural-network","neural","network","neural-networks","sheaf-attention","sheaf","transformer","vector-search","vector","search"],"s":["attention","cgt","coherence","contrastive-learning","contrastive","learn","deep-learning","deep","flash-attention","flash","gpu","graph-attention","graph","hyperbolic","infonce","machine-learning","machine","mixture-of-experts","mixture","expert","moe","multi-head-attention","multi","head","neural-network","neural","network","neural-networks","sheaf-attention","sheaf","transformer","vector-search","vector","search"]},
+{"id":"@ruvector/burst-scaling","family":"burst-scaling","kind":"npm","store":"ruvector","version":"1.0.2","description":"Adaptive burst scaling system for ruvector - handles 10-50x traffic spikes","keywords":["auto-scaling","burst","capacity","cloud-run","gcp","predictive-scaling","ruvector","scaling"],"source":"ruvector/npm/packages/burst-scaling/package.json","sourceSha256":"1f4d4cd7a422f93b5c78a5506dcf2b76083652bd7e6a34ebb9ba1f7ff4f35535","variants":[],"readme":"This burst scaling system enables Ruvector to handle massive traffic spikes (e.g., World Cup events with 25 billion concurrent streams) while maintaining strict latency SLAs and cost controls. - Predictive Scaling: ML-based forecasting pre-warms capacity before known events - Reactive Scaling: Real-time auto-scaling based on CPU, memory, connections, and latency - Global Orchestration: Cross-region capacity allocation with budget controls - Cost Management: Sophisticated budget tracking with graceful degradation - Infrastructure as Code: Complete Terraform configuration for GCP Cloud Run - Comprehensive Monitoring: Cloud Monitoring dashboard with 15+ key metrics - Node.js 18+ - Terraform 1.0","readmeSource":"ruvector/npm/packages/burst-scaling/README.md","product":"key:ruvector:burst-scaling","canonical":"@ruvector/burst-scaling","tier":"T0","t":["burst-scaling","burst","scal","adaptive","handle","10-50x","50x","traffic","spike","auto-scaling","auto","capacity","cloud-run","cloud","gcp","predictive-scaling","predictive"],"s":["burst-scaling","burst","scal","auto-scaling","auto","capacity","cloud-run","cloud","gcp","predictive-scaling","predictive"]},
+{"id":"@ruvector/cli","family":"cli","kind":"npm","store":"ruvector","version":"0.1.29","description":"Command-line interface for RuVector vector database with self-learning hooks","keywords":["claude-code","cli","command-line","database","hooks","intelligence","vector"],"source":"ruvector/npm/packages/cli/package.json","sourceSha256":"c895706e18e18de4d1c44c6d7a5f56a7609ab50a7aeaefcc6b38e4243983c339","variants":["ruvector-cli"],"product":"repo:ruvector","canonical":"ruvector","tier":"T0","t":["command-line","command","line","interface","vector","database","self-learning","self","learn","hook","claude-code","claude","code","intelligence"],"s":["claude-code","claude","code","command-line","command","line","database","hook","intelligence","vector"]},
+{"id":"@ruvector/cluster","family":"cluster","kind":"npm","store":"ruvector","version":"0.1.0","description":"Distributed clustering and coordination for RuVector - auto-sharding, Raft consensus integration","keywords":["cluster","consensus","distributed","raft","sharding","vector-database"],"source":"ruvector/crates/ruvector-cluster/package.json","sourceSha256":"0f0b002e71538f22e6a0822ed959782b57d36289cae2ce92690884ad76573e8d","variants":["ruvector-cluster"],"readme":"Distributed clustering and sharding for Ruvector vector databases. ruvector-cluster provides horizontal scaling capabilities with consistent hashing, shard management, and cluster coordination. Enables Ruvector to scale to billions of vectors across multiple nodes. Part of the [Ruvector](https://github.com/ruvnet/ruvector) ecosystem. - Horizontal Scaling: Distribute data across multiple nodes - Consistent Hashing: Minimal rebalancing on cluster changes - Auto-Sharding: Automatic shard distribution and balancing - Fault Tolerant: Handle node failures gracefully - Async-First: Built on Tokio for high-performance networking - Cluster Membership: Node discovery and health monitoring - Consistent","readmeSource":"ruvector/crates/ruvector-cluster/README.md","product":"key:ruvector:cluster","canonical":"@ruvector/cluster","tier":"T0","t":["cluster","distribut","coordination","auto-sharding","auto","shard","raft","consensus","integration","vector-database","vector","database"],"s":["cluster","consensus","distribut","raft","shard","vector-database","vector","database"]},
+{"id":"@ruvector/cnn","family":"cnn","kind":"npm","store":"ruvector","version":"0.1.1","description":"CNN feature extraction for image embeddings - SIMD-optimized, pure Rust/WASM","keywords":["cnn","contrastive-learning","embeddings","image","machine-learning","mobilenet","ruvector","simd","wasm"],"source":"ruvector/npm/packages/ruvector-cnn/package.json","sourceSha256":"a082d1d7815568c17ae11e6ebd0ef249e5e6d206b829e1e75e665888aef4c1e0","variants":["ruvector-cnn","ruvector-cnn-wasm"],"readme":"Turn images into searchable vectors — runs in browsers, no backend needed. This package converts images into numbers (called \"embeddings\") that describe what's in the picture. Similar images produce similar numbers. Use it to: - Build \"find similar images\" features - Group photos by what they show - Create visual search for products - Compare images without AI APIs","readmeSource":"ruvector/npm/packages/ruvector-cnn/README.md","product":"key:ruvector:cnn","canonical":"@ruvector/cnn","tier":"T0","t":["cnn","feature","extraction","image","embedding","simd-optimized","pure","contrastive-learning","contrastive","learn","machine-learning","machine","mobilenet"],"s":["cnn","contrastive-learning","contrastive","learn","embedding","image","machine-learning","machine","mobilenet"]},
+{"id":"@ruvector/core","family":"core","kind":"npm","store":"ruvector","version":"0.1.17","description":"High-performance Rust vector database for Node.js with HNSW indexing and SIMD optimizations","keywords":["database","embeddings","hnsw","machine-learning","napi","napi-rs","performance","rag","rust","semantic-search","simd","similarity-search","vector"],"source":"ruvector/npm/core/package.json","sourceSha256":"4c91b2c2703967d764b1f0d33173883ea2c8cc704a36a2291ebbe3cbc21e8424","variants":["@metabiohacker/core","ruvector-core"],"readme":"High-performance Rust vector database for Node.js with HNSW indexing and SIMD optimizations. - 🚀 Blazing Fast: Rust + SIMD optimizations for maximum performance - 🎯 HNSW Indexing: State-of-the-art approximate nearest neighbor search - 📦 Zero-Copy: Efficient buffer sharing between Rust and Node.js - 🔍 Multiple Distance Metrics: Euclidean, Cosine, Dot Product, Manhattan - 💾 Persistent Storage: Optional disk-based storage with memory mapping - 🔧 Quantization: Scalar, Product, and Binary quantization support - 📊 TypeScript: Full type definitions included - 🌍 Cross-Platform: Linux, macOS, and Windows support The package will automatically install the correct native binding for your platfo","readmeSource":"ruvector/npm/core/README.md","product":"repo:ruvector","canonical":"ruvector","tier":"T0","t":["core","high","vector","database","hnsw","index","embedding","machine-learning","machine","learn","rag","semantic-search","semantic","search","similarity-search","similarity"],"s":["core","database","embedding","hnsw","machine-learning","machine","learn","rag","semantic-search","semantic","search","similarity-search","similarity","vector"]},
+{"id":"@ruvector/diskann","family":"diskann","kind":"npm","store":"ruvector","version":"0.1.1","description":"DiskANN/Vamana — SSD-friendly billion-scale approximate nearest neighbor search with product quantization","keywords":["ann","approximate-nearest-neighbor","billion-scale","diskann","napi","product-quantization","rust","ruvector","ssd","vamana","vector-search"],"source":"ruvector/npm/packages/diskann/package.json","sourceSha256":"b545bc0e2ed442b8fcdb23f0471450f4506c267e116ac663393ad9b4788c6a5b","variants":["ruvector-diskann","ruvector-diskann-node"],"readme":"DiskANN / Vamana approximate-nearest-neighbor (ANN) search for Node.js — a Rust core compiled to native .node addons via [NAPI-RS](https://napi.rs/) for Linux x64/arm64, macOS x64/arm64, and Windows x64. DiskANN is the SSD-friendly graph index from Microsoft Research that powers billion-scale vector search on a single machine. This package implements the Vamana graph construction with α-robust pruning ([NeurIPS 2019](https://proceedings.neurips.cc/paper/2019/hash/09853c7fb1d3f8ee67a61b6bf4a7f8e6-Abstract.html)) plus optional Product Quantization (PQ) and mmap persistence so working set ≪ dataset size. - Vamana graph with two-pass construction (α=1.0 then α=1.2) and α-robust pruning — the pub","readmeSource":"ruvector/npm/packages/diskann/README.md","product":"key:ruvector:diskann","canonical":"@ruvector/diskann","tier":"T0","t":["diskann","vamana","ssd-friendly","ssd","friendly","billion-scale","billion","scale","approximate","nearest","neighbor","search","product","quantization","ann","approximate-nearest-neighbor","product-quantization","vector-search","vector"],"s":["diskann","ann","approximate-nearest-neighbor","approximate","nearest","neighbor","billion-scale","billion","scale","product-quantization","product","quantization","ssd","vamana","vector-search","vector","search"]},
+{"id":"@ruvector/emergent-time","family":"emergent-time","kind":"npm","store":"ruvector","version":"0.1.0","description":"Agentic Time: measure an AI agent's internal time and 7-state health from per-channel state deltas, in the browser, on the edge, or in Node — via a tiny WASM build of the emergent-time Rust crate.","keywords":["agentic-time","agents","ai-agents","change-point","emergent-time","observability","quantum-gravity","relational","thermodynamics","time","wasm","webassembly"],"source":"ruvector/npm/packages/emergent-time/package.json","sourceSha256":"ebd86e823517bb25cf7f9db646ec8a9a8cf379f3a0ae15eb40ae677b65402767","variants":["emergent-time","emergent-time-wasm"],"readme":"Agentic Time for the browser, the edge, and Node — a tiny WASM build of the agentic-time layer of the [emergent-time](https://crates.io/crates/emergent-time) Rust crate. Agentic time measures how much an AI agent has *changed internally*, not how many seconds, steps, or tokens have elapsed. You feed it the six channel deltas of a transition — belief, memory, retrieval, goal-graph, contradiction, plan — and it returns: - an explainable tick (a post-floor internal-time increment, its class, a human-readable reason, and the per-channel contributions), - a cumulative agentic time reading, - the Agentic Time Index (ATI = progress per unit of structural change), and - a 7-state health classificati","readmeSource":"ruvector/npm/packages/emergent-time/README.md","product":"key:ruvector:emergent-time","canonical":"@ruvector/emergent-time","tier":"T0","t":["emergent-time","emergent","measure","internal","7-state","state","health","per-channel","per","channel","delta","browser","edge","via","tiny","crate","agentic-time","ai-agents","change-point","change","point","observability","quantum-gravity","quantum","gravity","relational","thermodynamic"],"s":["emergent-time","emergent","agentic-time","ai-agents","change-point","change","point","observability","quantum-gravity","quantum","gravity","relational","thermodynamic"]},
+{"id":"@ruvector/gnn","family":"gnn","kind":"npm","store":"ruvector","version":"0.1.25","description":"Graph Neural Network capabilities for Ruvector - Node.js bindings","keywords":["browser","gnn","graph-neural-network","hnsw","machine-learning","napi-rs","neural-networks","ruvector","vector-database","wasm","webassembly"],"source":"ruvector/crates/ruvector-gnn-node/package.json","sourceSha256":"6d861f968047b20671a8feb141d1b788f13cd56d8f802cdd1587d11cb1d1afea","variants":["@ruvector/gnn-wasm","ruvector-gnn","ruvector-gnn-node","ruvector-gnn-wasm"],"readme":"High-performance Graph Neural Network (GNN) capabilities for Ruvector, powered by Rust and NAPI-RS. - GNN Layers: Multi-head attention, layer normalization, GRU cells - Tensor Compression: Adaptive compression with 5 levels (None, Half, PQ8, PQ4, Binary) - Differentiable Search: Soft attention-based search with temperature scaling - Hierarchical Processing: Multi-layer GNN forward pass - Zero-copy: Efficient data transfer between JavaScript and Rust - TypeScript Support: Full type definitions included - forward(nodeEmbedding: number[], neighborEmbeddings: number[][], edgeWeights: number[]): number[] - toJson(): string - Serialize layer to JSON - fromJson(json: string): RuvectorLayer - Deseri","readmeSource":"ruvector/crates/ruvector-gnn-node/README.md","product":"key:ruvector:gnn","canonical":"@ruvector/gnn","tier":"T0","t":["gnn","graph","neural","network","capability","browser","graph-neural-network","hnsw","machine-learning","machine","learn","neural-networks","vector-database","vector","database"],"s":["gnn","browser","graph-neural-network","graph","neural","network","hnsw","machine-learning","machine","learn","neural-networks","vector-database","vector","database"]},
+{"id":"@ruvector/graph-data-generator","family":"graph-data-generator","kind":"npm","store":"ruvector","version":"0.1.0","description":"AI-powered synthetic graph data generation with OpenRouter/Kimi K2 integration for Neo4j knowledge graphs, social networks, and temporal events","keywords":["ai-training","cypher","entity-relationships","graph-data","graph-generation","kimi-k2","knowledge-graph","llm","neo4j","openrouter","ruvector","social-network","synthetic-data","temporal-graphs","typescript","vector-embeddings"],"source":"ruvector/npm/packages/graph-data-generator/package.json","sourceSha256":"d25e475b6b07b7c92cd2ad6fb4dcf1395b1de8fdd3e9c3a62c104147aa11fb73","variants":[],"readme":"AI-powered synthetic graph data generation with OpenRouter/Kimi K2 integration for Neo4j knowledge graphs, social networks, and temporal events. - Knowledge Graph Generation: Create realistic knowledge graphs with entities and relationships - Social Network Generation: Generate social networks with various topology patterns - Temporal Events: Create time-series graph data with events and entities - Entity Relationships: Generate domain-specific entity-relationship graphs - Cypher Generation: Automatic Neo4j Cypher statement generation - Vector Embeddings: Enrich graphs with semantic embeddings - OpenRouter Integration: Powered by Kimi K2 and other OpenRouter models - Type-Safe: Full TypeScri","readmeSource":"ruvector/npm/packages/graph-data-generator/README.md","product":"key:ruvector:graph-data-generator","canonical":"@ruvector/graph-data-generator","tier":"T0","t":["graph-data-generator","graph","generator","ai-powered","power","synthetic","generation","openrouter","kimi","integration","neo4j","knowledge","social","network","temporal","event","ai-training","train","cypher","entity-relationships","entity","relationship","graph-data","graph-generation","kimi-k2","knowledge-graph","social-network","synthetic-data","temporal-graphs","vector-embeddings","vector","embedding"],"s":["graph-data-generator","graph","generator","ai-training","train","cypher","entity-relationships","entity","relationship","graph-data","graph-generation","generation","kimi-k2","kimi","knowledge-graph","knowledge","neo4j","openrouter","social-network","social","network","synthetic-data","synthetic","temporal-graphs","temporal","vector-embeddings","vector","embedding"]},
+{"id":"@ruvector/graph-node","family":"graph","kind":"npm","store":"ruvector","version":"2.1.1","description":"Native Node.js bindings for RuVector Graph Database with hypergraph support, Cypher queries, and persistence - 10x faster than WASM","keywords":["ai","browser","cypher","database","fast","graph","graph-database","graph-query","hypergraph","knowledge-graph","machine-learning","napi","native","neo4j","performance","property-graph","rag","rust","ruv","ruvector","vector-database","wasm","webassembly","zero-copy"],"source":"ruvector/npm/packages/graph-node/package.json","sourceSha256":"c5e8263d10c583da27513c26e4c1d29428d00e171736e01bdbd67d71356d3cf2","variants":["@ruvector/graph-wasm","ruvector-graph","ruvector-graph-node","ruvector-graph-wasm"],"readme":"Native Node.js bindings for RuVector Graph Database with hypergraph support, Cypher queries, and persistence. 10x faster than WASM. - Native Performance: Direct NAPI-RS bindings - no WASM overhead - Hypergraph Support: Multi-node relationships with vector embeddings - Cypher Queries: a practical subset of Cypher — see [Cypher support](#cypher-support) - Persistence: ACID-compliant storage with redb backend - Vector Similarity Search: Fast k-NN search on embeddings - Graph Traversal: k-hop neighbor discovery - Transactions: Full ACID support with begin/commit/rollback - Batch Operations: High-throughput bulk inserts (131K+ ops/sec) - Zero-Copy: Efficient Float32Array handling - TypeScript: Fu","readmeSource":"ruvector/npm/packages/graph-node/README.md","product":"key:ruvector:graph","canonical":"@ruvector/graph-node","tier":"T0","t":["graph-node","graph","database","hypergraph","cypher","query","persistence","10x","than","browser","graph-database","graph-query","knowledge-graph","knowledge","machine-learning","machine","learn","neo4j","property-graph","property","rag","vector-database","vector","zero-copy","zero","copy"],"s":["graph-node","graph","browser","cypher","database","graph-database","graph-query","query","hypergraph","knowledge-graph","knowledge","machine-learning","machine","learn","neo4j","property-graph","property","rag","vector-database","vector","zero-copy","zero","copy"]},
+{"id":"@ruvector/graph-transformer","family":"graph-transformer","kind":"npm","store":"ruvector","version":"2.0.4","description":"Proof-gated graph transformer with 8 verified modules — physics, biological, manifold, temporal, economic graph intelligence via NAPI-RS","keywords":["attention","gnn","graph-neural-network","graph-transformer","napi-rs","neural-network","proof-gated","ruvector","verified","verified-training","wasm","webassembly"],"source":"ruvector/crates/ruvector-graph-transformer-node/package.json","sourceSha256":"5127f1d79294a1e7f9a79f006460253961c83dc14d8764bbebdebf6fe29797d2","variants":["ruvector-graph-transformer","ruvector-graph-transformer-node","ruvector-graph-transformer-wasm"],"readme":"Node.js bindings for RuVector Graph Transformer — proof-gated graph attention, verified training, and 8 specialized graph layers via NAPI-RS. Use graph transformers from JavaScript and TypeScript with native Rust performance. Every graph operation — adding nodes, computing attention, training weights — produces a formal proof receipt proving it was done correctly. The heavy computation runs in compiled Rust via NAPI-RS, so you get sub-millisecond proof verification without leaving the Node.js ecosystem. Prebuilt binaries are provided for:","readmeSource":"ruvector/crates/ruvector-graph-transformer-node/README.md","product":"key:ruvector:graph-transformer","canonical":"@ruvector/graph-transformer","tier":"T0","t":["graph-transformer","graph","transformer","proof-gated","proof","gated","verify","physic","biological","manifold","temporal","economic","intelligence","via","attention","gnn","graph-neural-network","neural","network","neural-network","verified-training","train"],"s":["graph-transformer","graph","transformer","attention","gnn","graph-neural-network","neural","network","neural-network","proof-gated","proof","gated","verify","verified-training","train"]},
+{"id":"@ruvector/kge","family":"kge","kind":"npm","store":"ruvector","version":"0.2.0","description":"Holographic knowledge-graph embeddings (HolE / RotatE) for ruvector: link prediction, relation composition and compositional queries, with ANN candidate retrieval over HNSW. Native (napi-rs) with a WASM fallback; no network.","keywords":["ann","complex","embeddings","graph","hnsw","hole","holographic-embeddings","kge","knowledge-graph","link-prediction","local-first","relation-composition","rotate","ruvector"],"source":"ruvector/npm/packages/kge/package.json","sourceSha256":"fdb81a147a501e495b9746e490c00530ad260a42a86a5aafd948933e4d2a34b4","variants":["ruvector-kge","ruvector-kge-ffi","ruvector-kge-wasm"],"readme":"Holographic knowledge-graph embeddings for [ruvector](https://github.com/ruvnet/ruvector): score a triple (subject, relation, object) for plausibility, then answer link prediction, relation similarity, and 2-hop composition — locally, with no network and no per-query cost. Native (napi-rs) with a WASM fallback. A trained pair of embedding tables (entities E, relations R) plus a scorer. The default scorer is HolE (Holographic Embeddings, Nickel, Rosasco & Poggio, AAAI 2016): it scores r · (e_s ⋆ e_o), a relation vector dotted with the *circular correlation* of the head and tail entity vectors. HolE ≡ ComplEx. Hayashi & Shimbo (ACL 2017, arXiv:1702.05563) proved HolE is algebraically the same","readmeSource":"ruvector/npm/packages/kge/README.md","product":"key:ruvector:kge","canonical":"@ruvector/kge","tier":"T0","t":["kge","holographic","knowledge-graph","knowledge","graph","embedding","hole","rotate","link","prediction","relation","composition","compositional","query","ann","candidate","retrieval","hnsw","fallback","network","complex","holographic-embeddings","link-prediction","local-first","local","relation-composition"],"s":["kge","ann","complex","embedding","graph","hnsw","hole","holographic-embeddings","holographic","knowledge-graph","knowledge","link-prediction","link","prediction","local-first","local","relation-composition","relation","composition","rotate"]},
+{"id":"@ruvector/mincut-wasm","family":"mincut","kind":"npm","store":"ruvector","version":"2.3.0","description":"Weighted dynamic minimum cut with validated typed-array batches","keywords":["algorithm","dynamic","dynamic-graph","graph","mincut","minimum-cut","napi","network-analysis","nodejs","self-healing","wasm"],"source":"ruvector/npm/packages/mincut-wasm/package.json","sourceSha256":"2251590628422fc2c6c179d1a74e07dc77e890676cbdc5e830525dc98f881cd8","variants":["ruvector-mincut","ruvector-mincut-node","ruvector-mincut-wasm"],"product":"key:ruvector:mincut","canonical":"@ruvector/mincut-wasm","tier":"T0","t":["mincut-wasm","mincut","weight","dynamic","minimum","cut","validat","typed-array","typed","array","batch","algorithm","dynamic-graph","graph","minimum-cut","network-analysis","network","analysis","self-healing","self","heal"],"s":["mincut-wasm","mincut","algorithm","dynamic","dynamic-graph","graph","minimum-cut","minimum","cut","network-analysis","network","analysis","self-healing","self","heal"]},
+{"id":"@ruvector/node","family":"node","kind":"npm","store":"ruvector","version":"0.1.19","description":"High-performance Rust vector database for Node.js with HNSW indexing and SIMD optimizations","keywords":["database","embeddings","hnsw","machine-learning","napi","performance","rag","rust","semantic-search","simd","similarity-search","vector"],"source":"ruvector/crates/ruvector-node/package.json","sourceSha256":"084b629a08b53f179019cbaff1ca82114ef1069c1d1d949d87355fbbc40297c3","variants":["ruvector-node"],"readme":"Native Rust performance for Node.js vector databases via NAPI-RS Bring the power of Ruvector's blazing-fast vector search to your Node.js and TypeScript applications. Built with NAPI-RS for zero-overhead native bindings, async/await support, and complete type safety. In the age of AI, Node.js applications need fast, efficient vector search for RAG systems, semantic search, and recommendation engines. But existing JavaScript solutions are slow, memory-intensive, or lack critical features. Ruvector Node.js eliminates these limitations. - ⚡ Native Performance: <0.5ms search latency with Rust-powered HNSW indexing - 🚀 10-100x Faster: Outperforms pure JavaScript vector databases by orders of mag","readmeSource":"ruvector/crates/ruvector-node/README.md","product":"key:ruvector:node","canonical":"@ruvector/node","tier":"T0","t":["high","vector","database","hnsw","index","embedding","machine-learning","machine","learn","rag","semantic-search","semantic","search","similarity-search","similarity"],"s":["database","embedding","hnsw","machine-learning","machine","learn","rag","semantic-search","semantic","search","similarity-search","similarity","vector"]},
+{"id":"@ruvector/ospipe","family":"ospipe","kind":"npm","store":"ruvector","version":"0.1.3","description":"OSpipe SDK - RuVector-enhanced personal AI memory system for Screenpipe pipes","keywords":["ai-memory","ospipe","pipes","ruvector","screenpipe","semantic-search","vector-search","wasm"],"source":"ruvector/npm/packages/ospipe/package.json","sourceSha256":"6c64ff997e717f67e3b1f779bfca451e3878149b3a01e9e072bdc89f88f9399a","variants":["@ruvector/ospipe-wasm"],"readme":"RuVector-enhanced personal AI memory SDK for Screenpipe [Screenpipe](https://github.com/screenpipe/screenpipe) is an open-source desktop application that continuously records your screen, audio, and UI interactions locally. It builds a searchable timeline of everything you see, hear, and do on your computer. Out of the box, Screenpipe stores its data in SQLite with FTS5 full-text indexing -- effective for keyword lookups, but limited to literal string matching. If you search for \"auth discussion,\" you will not find a frame that says \"we talked about login security.\" OSpipe replaces Screenpipe's storage and search backend with the [RuVector](https://github.com/ruvnet/ruvector) ecosystem -- a","readmeSource":"ruvector/npm/packages/ospipe/README.md","product":"key:ruvector:ospipe","canonical":"@ruvector/ospipe","tier":"T0","t":["ospipe","ruvector-enhanced","enhanc","personal","memory","screenpipe","pipe","ai-memory","semantic-search","semantic","search","vector-search","vector"],"s":["ospipe","ai-memory","memory","pipe","screenpipe","semantic-search","semantic","search","vector-search","vector"]},
+{"id":"@ruvector/pi-brain","family":"pi-brain","kind":"npm","store":"ruvector","version":"0.1.2","description":"CLI and SDK for π — the RuVector shared brain. Share, search, and transfer learning across AI sessions.","keywords":["ai","brain","claude","mcp","pi","ruvector","shared-learning"],"source":"ruvector/npm/packages/pi-brain/package.json","sourceSha256":"24dd82d94b1c60a4ae85e7735a5bbbf79ac884e55e66b2b4d1f8cac289fd551c","variants":[],"readme":"CLI and SDK for π.ruv.io — the RuVector shared AI brain. Search, share, and transfer knowledge across AI sessions with cryptographic verification and federated learning. Or run directly: Set your API key via environment variable: Your identity is pseudonymous — the server derives a contributor ID from your key via SHAKE-256. No PII is stored. Register with Claude Code: Available MCP tools: brain_search, brain_share, brain_vote, brain_list, brain_get, brain_status, brain_drift, brain_partition, brain_transfer, brain_delete.","readmeSource":"ruvector/npm/packages/pi-brain/README.md","product":"key:ruvector:pi-brain","canonical":"@ruvector/pi-brain","tier":"T0","t":["pi-brain","brain","shar","share","search","transfer","learn","across","session","claude","mcp","shared-learning"],"s":["pi-brain","brain","claude","mcp","shared-learning","shar","learn"]},
+{"id":"@ruvector/postgres-cli","family":"postgres","kind":"npm","store":"ruvector","version":"0.2.9","description":"Advanced AI vector database CLI for PostgreSQL - pgvector drop-in replacement with 53+ SQL functions, 39 attention mechanisms, GNN layers, hyperbolic embeddings, and self-learning capabilities","keywords":["agent-routing","ai","ann","approximate-nearest-neighbor","artificial-intelligence","attention","bm25","cli","command-line","cypher","deep-learning","embeddings","flash-attention","gat","gcn","gnn","graph-database","graph-neural-network","graphsage","hierarchical-embeddings","hnsw","hybrid-search","hyperbolic","ivfflat","knowledge-graph","llm","lorentz","machine-learning","multi-head-attention","neural-network","pgvector","poincare","postgres","postgresql","quantization","rag","reasoning","retrieval-augmented-generation","ruvector","self-learning","semantic-search","similarity-search","sparse-vectors","splade","tf-idf","transformer","vector","vector-database","vector-quantization"],"source":"ruvector/npm/packages/postgres-cli/package.json","sourceSha256":"ad16a342a4ea368ff884f4a73dcb02efe92009a1c5ea7024e84c6a882711b060","variants":["ruvector-postgres"],"readme":"The most advanced AI vector database CLI for PostgreSQL. A drop-in pgvector replacement with 53+ SQL functions, 39 attention mechanisms, GNN layers, hyperbolic embeddings, and self-learning capabilities. That's it! The installer will: 1. Detect your system (Linux/macOS) 2. Install PostgreSQL if not present 3. Install Rust if not present (for native installs) 4. Build and install the RuVector extension 5. Create a ready-to-use database PostgreSQL Versions: 14, 15, 16, 17 (native), Docker supports all Direct Docker Hub Usage (without CLI):","readmeSource":"ruvector/npm/packages/postgres-cli/README.md","product":"key:ruvector:postgres","canonical":"@ruvector/postgres-cli","tier":"T0","t":["postgres-cli","postgre","advanc","vector","database","postgresql","pgvector","drop-in","drop","replacement","sql","function","attention","mechanism","gnn","layer","hyperbolic","embedding","self-learning","self","learn","capability","agent-routing","rout","ann","approximate-nearest-neighbor","approximate","nearest","neighbor","artificial-intelligence","artificial","intelligence","bm25","command-line","command","line","cypher","deep-learning","deep","flash-attention","flash","gat","gcn","graph-database","graph","graph-neural-network","neural","network","graphsage","hierarchical-embeddings","hierarchical","hnsw","hybrid-search","hybrid","search","ivfflat","knowledge-graph","knowledge","lorentz","machine-learning","machine","multi-head-attention","multi","head","neural-network","poincare","quantization","rag","reason","retrieval-augmented-generation","retrieval","augment","generation","semantic-search","semantic","similarity-search","similarity","sparse-vectors","sparse","splade","tf-idf","idf","transformer","vector-database","vector-quantization"],"s":["postgres-cli","postgre","agent-routing","rout","ann","approximate-nearest-neighbor","approximate","nearest","neighbor","artificial-intelligence","artificial","intelligence","attention","bm25","command-line","command","line","cypher","deep-learning","deep","learn","embedding","flash-attention","flash","gat","gcn","gnn","graph-database","graph","database","graph-neural-network","neural","network","graphsage","hierarchical-embeddings","hierarchical","hnsw","hybrid-search","hybrid","search","hyperbolic","ivfflat","knowledge-graph","knowledge","lorentz","machine-learning","machine","multi-head-attention","multi","head","neural-network","pgvector","poincare","postgresql","quantization","rag","reason","retrieval-augmented-generation","retrieval","augment","generation","self-learning","self","semantic-search","semantic","similarity-search","similarity","sparse-vectors","sparse","vector","splade","tf-idf","idf","transformer","vector-database","vector-quantization"]},
+{"id":"@ruvector/raft","family":"raft","kind":"npm","store":"ruvector","version":"0.1.1","description":"Raft consensus implementation for distributed systems - leader election, log replication, and fault tolerance","keywords":["cluster","consensus","distributed-consensus","distributed-systems","fault-tolerance","leader-election","log-replication","raft","ruvector"],"source":"ruvector/npm/packages/raft/package.json","sourceSha256":"d8a7701b336adc3c68b7036ecee36626df22ad4b331eeb5018ccc2d01442f9c7","variants":["ruvector-raft"],"product":"key:ruvector:raft","canonical":"@ruvector/raft","tier":"T0","t":["raft","consensus","implementation","distribut","leader","election","log","replication","fault","tolerance","cluster","distributed-consensus","distributed-systems","fault-tolerance","leader-election","log-replication"],"s":["raft","cluster","consensus","distributed-consensus","distribut","distributed-systems","fault-tolerance","fault","tolerance","leader-election","leader","election","log-replication","log","replication"]},
+{"id":"@ruvector/replication","family":"replication","kind":"npm","store":"ruvector","version":"0.1.1","description":"Data replication and synchronization - multi-node replicas, conflict resolution with vector clocks, change data capture, and automatic failover","keywords":["change-data-capture","conflict-resolution","crdt","distributed-systems","failover","high-availability","replication","ruvector","synchronization","vector-clock"],"source":"ruvector/npm/packages/replication/package.json","sourceSha256":"31e0a15671c2b45028834c345893ea18c77bd4992221ae6384e9b6ea488996f3","variants":["ruvector-replication"],"product":"key:ruvector:replication","canonical":"@ruvector/replication","tier":"T0","t":["replication","synchronization","multi-node","multi","replica","conflict","resolution","vector","clock","change","capture","automatic","failover","change-data-capture","conflict-resolution","crdt","distributed-systems","distribut","high-availability","high","availability","vector-clock"],"s":["replication","change-data-capture","change","capture","conflict-resolution","conflict","resolution","crdt","distributed-systems","distribut","failover","high-availability","high","availability","synchronization","vector-clock","vector","clock"]},
+{"id":"@ruvector/router","family":"router","kind":"npm","store":"ruvector","version":"0.1.32","description":"Semantic router for AI agents - vector-based intent matching with HNSW indexing and SIMD acceleration","keywords":["agent-routing","ai-routing","browser","embeddings","fast","hnsw","intent-matching","llm","napi","native","performance","rust","ruv","ruvector","semantic-router","simd","similarity-search","vector-search","wasm","webassembly"],"source":"ruvector/npm/packages/router/package.json","sourceSha256":"933ac57fc5e8edd8948985b7397d97f66632c17ef9957f8ef0d2b5b071604228","variants":["@ruvector/router-wasm","ruvector-router-cli","ruvector-router-core","ruvector-router-ffi","ruvector-router-wasm"],"readme":"Semantic router for AI agents - vector-based intent matching with HNSW indexing and SIMD acceleration. - Semantic Intent Matching: Route queries to intents based on meaning, not keywords - HNSW Indexing: Fast approximate nearest neighbor search - SIMD Optimized: Native Rust performance with vectorized operations - Quantization: Memory-efficient storage for large intent sets - Multi-Platform: Works on Linux, macOS, and Windows The package automatically installs the correct native binary for your platform. Main class for semantic routing. RouterConfig: Add an intent to the router. Route a query to matching intents. Route with a pre-computed embedding (synchronous). Remove an intent from the ro","readmeSource":"ruvector/npm/packages/router/README.md","product":"key:ruvector:router","canonical":"@ruvector/router","tier":"T0","t":["router","semantic","vector-based","vector","intent","match","hnsw","index","acceleration","agent-routing","rout","ai-routing","browser","embedding","intent-matching","semantic-router","similarity-search","similarity","search","vector-search"],"s":["router","agent-routing","rout","ai-routing","browser","embedding","hnsw","intent-matching","intent","match","semantic-router","semantic","similarity-search","similarity","search","vector-search","vector"]},
+{"id":"@ruvector/rudag","family":"rudag","kind":"npm","store":"ruvector","version":"0.1.1","description":"Fast DAG (Directed Acyclic Graph) library with Rust/WASM. Topological sort, critical path, task scheduling, dependency resolution, workflow optimization. Self-learning ML attention. Browser & Node.js with auto-persistence.","keywords":["attention-mechanism","bottleneck","browser","build-system","critical-path","dag","data-pipeline","dependency-graph","dependency-resolution","directed-acyclic-graph","etl","graph","indexeddb","job-scheduler","machine-learning","nodejs","performance","persistence","pipeline","query-optimizer","rust","self-learning","sql-optimizer","task-scheduler","task-scheduling","topo-sort","topological-sort","typescript","wasm","webassembly","workflow","workflow-engine"],"source":"ruvector/npm/packages/rudag/package.json","sourceSha256":"4e379b9fc291ade0f1bec8bc3668ba490bbbfddf6e7155c095b83fc575e646da","variants":[],"readme":"Smart task scheduling with self-learning optimization — powered by Rust/WASM rudag answers these questions instantly. It's a Directed Acyclic Graph (DAG) library that helps you manage dependencies, find bottlenecks, and optimize execution — all with self-learning intelligence that gets smarter over time. You have tasks with dependencies. Task C needs A and B to finish first: You need answers:","readmeSource":"ruvector/npm/packages/rudag/README.md","product":"key:ruvector:rudag","canonical":"@ruvector/rudag","tier":"T0","t":["rudag","dag","direct","acyclic","graph","topological","sort","critical","path","task","schedul","dependency","resolution","workflow","self-learning","self","learn","attention","browser","auto-persistence","auto","persistence","attention-mechanism","mechanism","bottleneck","build-system","critical-path","data-pipeline","pipeline","dependency-graph","dependency-resolution","directed-acyclic-graph","etl","indexeddb","job-scheduler","job","scheduler","machine-learning","machine","query-optimizer","query","optimizer","sql-optimizer","sql","task-scheduler","task-scheduling","topo-sort","topo","topological-sort","workflow-engine"],"s":["rudag","attention-mechanism","attention","mechanism","bottleneck","browser","build-system","critical-path","critical","path","dag","data-pipeline","pipeline","dependency-graph","dependency","graph","dependency-resolution","resolution","directed-acyclic-graph","direct","acyclic","etl","indexeddb","job-scheduler","job","scheduler","machine-learning","machine","learn","persistence","query-optimizer","query","optimizer","self-learning","self","sql-optimizer","sql","task-scheduler","task","task-scheduling","schedul","topo-sort","topo","sort","topological-sort","topological","workflow","workflow-engine"]},
+{"id":"@ruvector/ruqu","family":"ruqu","kind":"npm","store":"ruqu","version":"0.2.0","description":"Quantum circuit simulator in WebAssembly + agent harness CLI — run state-vector simulation, Grover search, and QAOA MaxCut from your terminal. npx @ruvector/ruqu","keywords":["agent","agent-harness","ai-agent","browser","claude-code","cli","coherence","error-correction","gate","grover","harness","metaharness","min-cut","npx","qaoa","quantum","quantum-circuit","quantum-computing","quantum-simulation","quantum-simulator","qubits","ruqu","rust","simulation","state-vector","surface-code","vqe","wasm","webassembly"],"source":"ruqu/cli/package.json","sourceSha256":"d0251ede0442657c3705f3e0e532b8c27e120a053278368af3e55e78f35590b1","variants":["@ruvector/ruqu-wasm","ruqu","ruqu-core","ruqu-wasm"],"readme":"Quantum computing from your terminal — a state-vector quantum circuit simulator compiled to WebAssembly (the pure-Rust [ruqu](https://github.com/ruvnet/ruqu) crates), wrapped in a [metaharness](https://github.com/ruvnet/agent-harness-generator) agent CLI. Gates: h x y z s t rx ry rz cnot cz swap rzz measure reset barrier. Up to 25 qubits. The CLI bundles a --target nodejs WebAssembly build of the ruqu-wasm crate (real state-vector simulation in Rust → WASM) and loads it directly in Node — no native addon, no Python. It also boots the metaharness kernel + Claude Code host adapter for the agent-harness commands (init/doctor). The observer laboratory is available from this source checkout with","readmeSource":"ruqu/cli/README.md","product":"repo:ruqu","canonical":"@ruvector/ruqu","tier":"T2","t":["ruqu","quantum","circuit","simulator","harness","state-vector","state","vector","simulation","grover","search","qaoa","maxcut","terminal","npx","agent-harness","ai-agent","browser","claude-code","claude","code","coherence","error-correction","error","correction","gate","metaharness","min-cut","min","cut","quantum-circuit","quantum-computing","comput","quantum-simulation","quantum-simulator","qubit","surface-code","surface","vqe"],"s":["ruqu","agent-harness","harness","ai-agent","browser","claude-code","claude","code","coherence","error-correction","error","correction","gate","grover","metaharness","min-cut","min","cut","npx","qaoa","quantum","quantum-circuit","circuit","quantum-computing","comput","quantum-simulation","simulation","quantum-simulator","simulator","qubit","state-vector","state","vector","surface-code","surface","vqe"]},
+{"id":"@ruvector/ruvllm","family":"ruvllm","kind":"npm","store":"ruvector","version":"2.7.0","description":"Self-learning LLM runtime — TurboQuant KV-cache (6-8x compression), SONA adaptive learning, FlashAttention, speculative decoding, GGUF inference","keywords":["adaptive-learning","ai","benchmarking","browser","cli","continual-learning","cuda","deep-learning","edge","edge-ai","ewc","fastgrnn","federated-learning","flash-attention","gguf","hnsw","inference","kv-cache","llm","local-llm","lora","machine-learning","mamba","metal","model-compression","model-serving","napi","offline","paged-attention","quantization","router","rust","ruvector","ruvllm","self-learning","simd","sona","speculative-decoding","transformer","transformers","turboquant","vector-database","wasm","webassembly","webgpu"],"source":"ruvector/npm/packages/ruvllm/package.json","sourceSha256":"d3fb2008bfd09fe84b3dc0442d24a95c67189daed03ec375b66afcd463b95b22","variants":["@ruvector/ruvllm-cli","@ruvector/ruvllm-wasm","ruvllm","ruvllm-cli","ruvllm-wasm"],"readme":"Self-learning LLM runtime for Node.js — GGUF inference, TurboQuant KV-cache compression (6-8x memory savings), SONA adaptive learning, FlashAttention, speculative decoding, and SIMD-optimized kernels. Built in Rust, runs everywhere. Reduce inference memory by 6-8x with <0.5% quality loss:","readmeSource":"ruvector/npm/packages/ruvllm/README.md","product":"key:ruvector:ruvllm","canonical":"@ruvector/ruvllm","tier":"T0","t":["ruvllm","self-learning","self","learn","runtime","turboquant","kv-cache","cache","6-8x","compression","sona","adaptive","flashattention","speculative","decod","gguf","inference","adaptive-learning","benchmark","browser","continual-learning","continual","cuda","deep-learning","deep","edge","edge-ai","ewc","fastgrnn","federated-learning","federat","flash-attention","flash","attention","hnsw","local-llm","local","lora","machine-learning","machine","mamba","metal","model-compression","model","model-serving","serv","offline","paged-attention","paged","quantization","router","speculative-decoding","transformer","vector-database","vector","database","webgpu"],"s":["ruvllm","adaptive-learning","adaptive","learn","benchmark","browser","continual-learning","continual","cuda","deep-learning","deep","edge","edge-ai","ewc","fastgrnn","federated-learning","federat","flash-attention","flash","attention","gguf","hnsw","inference","kv-cache","cache","local-llm","local","lora","machine-learning","machine","mamba","metal","model-compression","model","compression","model-serving","serv","offline","paged-attention","paged","quantization","router","self-learning","self","sona","speculative-decoding","speculative","decod","transformer","turboquant","vector-database","vector","database","webgpu"]},
+{"id":"@ruvector/rvdna","family":"rvdna","kind":"npm","store":"rvdna","version":"0.3.0","description":"rvDNA — AI-native genomic analysis. 20-SNP biomarker risk scoring, streaming anomaly detection, 64-dim profile vectors, 23andMe genotyping, CYP2D6/CYP2C19 pharmacogenomics, variant calling, protein prediction, and HNSW vector search.","keywords":["23andme","ai","anomaly-detection","bioinformatics","biomarker","dna","genomics","health","hnsw","napi","pharmacogenomics","protein","risk-score","rust","rvdna","streaming","variant-calling","vector-search","wasm"],"source":"rvdna/npm/packages/rvdna/package.json","sourceSha256":"d8e44cdcfe2db5d0f9b452d999af9b44c4f40bf38ad9d9b2fdb6eebc96e36db0","variants":["rvdna"],"readme":"DNA analysis in JavaScript. Encode sequences, translate proteins, search genomes by similarity, and read the .rvdna AI-native file format — all from Node.js or the browser. Built on Rust via NAPI-RS for native speed. Falls back to pure JavaScript when native bindings aren't available. Packs a DNA string into 2-bit bytes. Each byte holds 4 bases: A=00, C=01, G=10, T=11. Ambiguous bases (N) map to A. Decodes 2-bit packed bytes back to a DNA string. You must pass the original sequence length since the last byte may have padding. Translates a DNA string to a protein amino acid string using the standard genetic code. Stops at the first stop codon (TAA, TAG, TGA). Returns cosine similarity between","readmeSource":"rvdna/npm/packages/rvdna/README.md","product":"repo:rvdna","canonical":"@ruvector/rvdna","tier":"T2","t":["rvdna","ai-native","genomic","analysis","20-snp","snp","biomarker","risk","scor","stream","anomaly","detection","64-dim","dim","profile","vector","23andme","genotyp","cyp2d6","cyp2c19","pharmacogenomic","variant","call","protein","prediction","hnsw","search","anomaly-detection","bioinformatic","dna","health","risk-score","score","variant-calling","vector-search"],"s":["rvdna","23andme","anomaly-detection","anomaly","detection","bioinformatic","biomarker","dna","genomic","health","hnsw","pharmacogenomic","protein","risk-score","risk","score","stream","variant-calling","variant","call","vector-search","vector","search"]},
+{"id":"@ruvector/rvf","family":"rvf","kind":"npm","store":"ruvector","version":"0.3.4","description":"RuVector Format — unified TypeScript SDK for vector intelligence","keywords":["binary-format","binding","cli","cognitive-container","database","embedding","hnsw","napi","nodejs","rvf","simd","vector","wasm","webassembly"],"source":"ruvector/npm/packages/rvf/package.json","sourceSha256":"982165801dc48d885c6de852194208056b87d0d243692b4af161e1cdc8cac76b","variants":["@ruvector/rvf-node","@ruvector/rvf-wasm","rvf-cli","rvf-node","rvf-types","rvf-wasm"],"readme":"Unified TypeScript/JavaScript SDK for the RuVector Format (RVF) — a cognitive container that stores vectors, carries models, boots compute kernels, and proves everything in a single .rvf file. Deno: The WASM build targets wasm32-unknown-unknown, which runs natively in Deno. Import via npm: specifier or load the .wasm bundle directly. Browser: The @ruvector/rvf-wasm package provides a ~46 KB control-plane WASM module plus a ~5.5 KB tile-compute module. Works in any browser with WebAssembly support. RVF (RuVector Format) is a universal binary substrate that merges database, model, graph engine, kernel, and attestation into a single deployable file. A .rvf file is segmented — each segment carri","readmeSource":"ruvector/npm/packages/rvf/README.md","product":"key:ruvector:rvf","canonical":"@ruvector/rvf","tier":"T0","t":["rvf","format","unify","vector","intelligence","binary-format","binary","cognitive-container","cognitive","container","database","embedd","hnsw"],"s":["rvf","binary-format","binary","format","cognitive-container","cognitive","container","database","embedd","hnsw","vector"]},
+{"id":"@ruvector/rvf-mcp-server","family":"rvf-mcp-server","kind":"npm","store":"ruvector","version":"0.1.4","description":"MCP server for RuVector Format (RVF) vector database — stdio and SSE transports","keywords":["mcp","model-context-protocol","ruvector","rvf","vector-database"],"source":"ruvector/npm/packages/rvf-mcp-server/package.json","sourceSha256":"6d653d1cbf1204ec72b733e1872102558d5041513369ff0e71701d4b9afc5c78","variants":[],"readme":"MCP (Model Context Protocol) server for RuVector Format (RVF) vector stores. Exposes RVF capabilities to AI agents like Claude Code, Cursor, and other MCP-compatible tools. Add to your MCP config: - Node.js >= 18.0.0 MIT","readmeSource":"ruvector/npm/packages/rvf-mcp-server/README.md","product":"key:ruvector:rvf-mcp-server","canonical":"@ruvector/rvf-mcp-server","tier":"T0","t":["rvf-mcp-server","rvf","mcp","server","format","vector","database","stdio","sse","transport","model-context-protocol","model","context","protocol","vector-database"],"s":["rvf-mcp-server","rvf","mcp","server","model-context-protocol","model","context","protocol","vector-database","vector","database"]},
+{"id":"@ruvector/rvf-solver","family":"rvf-solver","kind":"npm","store":"ruvector","version":"0.1.8","description":"RVF self-learning temporal solver — Thompson Sampling, PolicyKernel, ReasoningBank","keywords":["agi","reasoning-bank","rvf","self-learning","solver","temporal","thompson-sampling","wasm"],"source":"ruvector/npm/packages/rvf-solver/package.json","sourceSha256":"9099da756e020dab3645f2f9a1a46fb4717e9a174b57c0605ce164fc4fc99f9b","variants":["rvf-solver-wasm"],"readme":"Self-learning temporal solver with Thompson Sampling, PolicyKernel, ReasoningBank, and SHAKE-256 tamper-evident witness chains. Runs in the browser, Node.js, and edge runtimes via WebAssembly. Or via the unified SDK: - Thompson Sampling two-signal model — safety Beta distribution + cost EMA for adaptive policy selection - 18 context-bucketed bandits — 3 range x 3 distractor x 2 noise levels for fine-grained context awareness - KnowledgeCompiler with signature-based pattern cache — distills learned patterns into reusable compiled configurations - Speculative dual-path execution — runs two candidate arms in parallel, picks the winner - Three-loop adaptive solver — fast: constraint propagation","readmeSource":"ruvector/npm/packages/rvf-solver/README.md","product":"key:ruvector:rvf-solver","canonical":"@ruvector/rvf-solver","tier":"T0","t":["rvf-solver","rvf","solver","self-learning","self","learn","temporal","thompson","sampl","policykernel","reasoningbank","agi","reasoning-bank","reason","bank","thompson-sampling"],"s":["rvf-solver","rvf","solver","agi","reasoning-bank","reason","bank","self-learning","self","learn","temporal","thompson-sampling","thompson","sampl"]},
+{"id":"@ruvector/rvforge","family":"rvforge","kind":"npm","store":"ruvector","version":"0.2.0","description":"RVForge — one canonical RVF to signed platform installers","keywords":["installer","packaging","provenance","rvf","rvforge","signing","tauri"],"source":"ruvector/npm/packages/rvforge/package.json","sourceSha256":"938e2df2c5b8226ed6b1c8711a9961c1e36cec209d206accfa9da2235d886859","variants":[],"readme":"RVForge — one canonical RVF to signed platform installers. forge turns a single .rvf agent into installable packages for Windows, macOS, Linux, and RVM, without the agent inside each package ever differing. The RVF identity, contents, policies, and signatures are the same in the .dmg as in the .msi. [The Sandbox Is Not the Boundary →](https://ruvnet.github.io/RuVector/rvforge/) An illustrated walkthrough at why process isolation does not bound an agent, then goes command by command from authoring a signed artifact to running it under the capability gate — ending by flipping one byte to show what the identity check is actually for. See [ADR-283](../../../docs/adr/ADR-283-rvf-forge-canonical-i","readmeSource":"ruvector/npm/packages/rvforge/README.md","product":"key:ruvector:rvforge","canonical":"@ruvector/rvforge","tier":"T0","t":["rvforge","canonical","rvf","sign","installer","packag","provenance","tauri"],"s":["rvforge","installer","packag","provenance","rvf","sign","tauri"]},
+{"id":"@ruvector/scipix","family":"scipix","kind":"npm","store":"ruvector","version":"0.1.1","description":"OCR client for scientific documents - extract LaTeX, MathML from equations, research papers, and technical diagrams","keywords":["document-processing","equation-recognition","image-recognition","latex","math-ocr","mathml","ocr","pdf-extraction","ruvector","scientific-computing"],"source":"ruvector/npm/packages/scipix/package.json","sourceSha256":"b93ce4b741e548acbd1c0e5c8fb3f251b3ec7ddb81c78d8df33029e02a622193","variants":[],"product":"key:ruvector:scipix","canonical":"@ruvector/scipix","tier":"T0","t":["scipix","ocr","client","scientific","document","extract","latex","mathml","equation","research","paper","technical","diagram","document-processing","process","equation-recognition","recognition","image-recognition","image","math-ocr","math","pdf-extraction","pdf","extraction","scientific-computing","comput"],"s":["scipix","document-processing","document","process","equation-recognition","equation","recognition","image-recognition","image","latex","math-ocr","math","ocr","mathml","pdf-extraction","pdf","extraction","scientific-computing","scientific","comput"]},
+{"id":"@ruvector/server","family":"server","kind":"npm","store":"ruvector","version":"0.1.0","description":"HTTP/gRPC server for RuVector - REST API with streaming support","keywords":["grpc","http","rest-api","server","streaming","vector-database"],"source":"ruvector/crates/ruvector-server/package.json","sourceSha256":"9f8d65ced094fea51704bbef203ce4ea89045ce07590d406d69ba52dba41b128","variants":["ruvector-server"],"readme":"High-performance REST API server for Ruvector vector databases. ruvector-server provides a production-ready HTTP API built on Axum with CORS support, compression, and OpenAPI documentation. Exposes full Ruvector functionality via RESTful endpoints. Part of the [Ruvector](https://github.com/ruvnet/ruvector) ecosystem. - Fast: Built on Axum and Tokio for high throughput - Production Ready: CORS, compression, tracing built-in - RESTful API: Standard HTTP endpoints for all operations - OpenAPI: Auto-generated API documentation - Multi-Collection: Support multiple vector collections - Vector CRUD: Insert, get, update, delete vectors - Search API: k-NN search with filtering - Batch Operations: Bul","readmeSource":"ruvector/crates/ruvector-server/README.md","product":"key:ruvector:server","canonical":"@ruvector/server","tier":"T0","t":["server","http","grpc","rest","stream","rest-api","vector-database","vector","database"],"s":["server","grpc","http","rest-api","rest","stream","vector-database","vector","database"]},
+{"id":"@ruvector/solver","family":"solver","kind":"npm","store":"ruvector","version":"0.1.0","description":"Sublinear-time sparse linear solver for Node.js via NAPI-RS","keywords":["linear-algebra","napi-rs","neumann-series","pagerank","rust","ruvector","solver","sparse-matrix","sublinear"],"source":"ruvector/crates/ruvector-solver-node/package.json","sourceSha256":"99c99bd04b0f2a3ead3df2799b6e87bfe3762adb2425332d2a0111e715cd7089","variants":["ruvector-solver","ruvector-solver-node","ruvector-solver-wasm"],"product":"key:ruvector:solver","canonical":"@ruvector/solver","tier":"T0","t":["solver","sublinear-time","sublinear","sparse","linear","via","linear-algebra","algebra","neumann-series","neumann","sery","pagerank","sparse-matrix","matrix"],"s":["solver","linear-algebra","linear","algebra","neumann-series","neumann","sery","pagerank","sparse-matrix","sparse","matrix","sublinear"]},
+{"id":"@ruvector/sona","family":"sona","kind":"npm","store":"ruvector","version":"0.1.8","description":"Self-Optimizing Neural Architecture (SONA) - Runtime-adaptive learning with LoRA, EWC++, and ReasoningBank for LLM routers and AI systems. Sub-millisecond learning overhead, WASM and Node.js support.","keywords":["adaptive","adaptive-learning","ai","continual-learning","deep-learning","elastic-weight-consolidation","ewc","learning","llm","llm-router","lora","low-rank-adaptation","machine-learning","napi","neural","neural-network","reasoningbank","rust","ruvector","sona"],"source":"ruvector/npm/packages/sona/package.json","sourceSha256":"ad2bca914422fd76d3ac6d3580b1812d97d1d2eaf41c8c11268a241d3eb0936b","variants":["ruvector-sona"],"readme":"Self-Optimizing Neural Architecture (SONA) - Node.js bindings for adaptive learning with ReasoningBank. SONA is a cutting-edge adaptive learning system that combines: - Micro-LoRA (rank 1-2): Ultra-fast inference-time adaptation - Base LoRA (rank 8+): Deeper background learning - EWC++: Catastrophic forgetting prevention - ReasoningBank: Pattern extraction and storage - Dual Learning Loops: Instant (<1ms) and background (periodic) learning - 🚀 Instant Adaptation: Sub-millisecond learning updates during inference - 🧠 Pattern Recognition: Automatic extraction and clustering of learned patterns - 🔄 Dual Learning Loops: Balance speed and depth with instant and background learning - 💾 Memory","readmeSource":"ruvector/npm/packages/sona/README.md","product":"key:ruvector:sona","canonical":"@ruvector/sona","tier":"T0","t":["sona","self-optimizing","self","optimiz","neural","architecture","runtime-adaptive","runtime","adaptive","learn","lora","ewc","reasoningbank","router","sub-millisecond","sub","millisecond","overhead","adaptive-learning","continual-learning","continual","deep-learning","deep","elastic-weight-consolidation","elastic","weight","consolidation","llm-router","low-rank-adaptation","low","rank","adaptation","machine-learning","machine","neural-network","network"],"s":["sona","adaptive","adaptive-learning","learn","continual-learning","continual","deep-learning","deep","elastic-weight-consolidation","elastic","weight","consolidation","ewc","llm-router","router","lora","low-rank-adaptation","low","rank","adaptation","machine-learning","machine","neural","neural-network","network","reasoningbank"]},
+{"id":"@ruvector/spiking-neural","family":"spiking-neural","kind":"npm","store":"ruvector","version":"1.0.3","description":"High-performance Spiking Neural Network (SNN) with SIMD optimization - CLI & SDK","keywords":["ai","biologically-inspired","deep-learning","lif-neuron","machine-learning","neural-network","neuromorphic","pattern-recognition","simd","snn","spiking-neural-network","stdp"],"source":"ruvector/npm/packages/spiking-neural/package.json","sourceSha256":"8a7e8605e3089d3a79f12164390f1db641fc91e6bddaae6d26057298c517cbad","variants":[],"readme":"A high-performance Spiking Neural Network (SNN) library for Node.js with biologically-inspired neuron models and unsupervised learning. SNNs process information through discrete spike events rather than continuous values, making them more energy-efficient and better suited for temporal pattern recognition. Unlike traditional neural networks that use continuous activation values, SNNs communicate through discrete spike events timed in milliseconds. This mirrors how biological neurons work, enabling: - Temporal pattern recognition - naturally process time-series data - Energy efficiency - event-driven computation (10-100x lower power) - Online learning - adapt in real-time with STDP (no batche","readmeSource":"ruvector/npm/packages/spiking-neural/README.md","product":"key:ruvector:spiking-neural","canonical":"@ruvector/spiking-neural","tier":"T0","t":["spiking-neural","spik","neural","high","network","snn","biologically-inspired","biologically","inspir","deep-learning","deep","learn","lif-neuron","lif","neuron","machine-learning","machine","neural-network","neuromorphic","pattern-recognition","pattern","recognition","spiking-neural-network","stdp"],"s":["spiking-neural","spik","neural","biologically-inspired","biologically","inspir","deep-learning","deep","learn","lif-neuron","lif","neuron","machine-learning","machine","neural-network","network","neuromorphic","pattern-recognition","pattern","recognition","snn","spiking-neural-network","stdp"]},
+{"id":"@ruvector/tiny-dancer","family":"tiny-dancer","kind":"npm","store":"ruvector","version":"0.1.22","description":"Neural router for AI agent orchestration - FastGRNN-based intelligent routing with circuit breaker, uncertainty estimation, and hot-reload","keywords":["agent-orchestration","ai-routing","browser","circuit-breaker","fast","fastgrnn","hot-reload","inference","llm-routing","model-routing","napi","native","neural-network","neural-router","performance","rust","ruv","ruvector","simd","uncertainty-estimation","wasm","webassembly"],"source":"ruvector/npm/packages/tiny-dancer/package.json","sourceSha256":"2fb2ad37ddc8f02907c0214569adcf41eb2fb26ae8ea4c85b997af3ed7cbb841","variants":["@ruvector/tiny-dancer-wasm","ruvector-tiny-dancer-core","ruvector-tiny-dancer-node","ruvector-tiny-dancer-wasm"],"readme":"Neural router for AI agent orchestration - FastGRNN-based intelligent routing with circuit breaker, uncertainty estimation, and hot-reload. - FastGRNN Neural Routing: Efficient gated recurrent network for fast inference - Uncertainty Estimation: Know when the router is confident vs. uncertain - Circuit Breaker: Automatic fallback when routing fails repeatedly - Hot-Reload: Update models without restarting the application - SIMD Optimized: Native Rust performance with SIMD acceleration - Multi-Platform: Works on Linux, macOS, and Windows The package automatically installs the correct native binary for your platform. Main class for neural routing. RouterConfig: Route a query to the best candid","readmeSource":"ruvector/npm/packages/tiny-dancer/README.md","product":"key:ruvector:tiny-dancer","canonical":"@ruvector/tiny-dancer","tier":"T0","t":["tiny-dancer","tiny","dancer","neural","router","orchestration","fastgrnn-based","fastgrnn","intelligent","rout","circuit","breaker","uncertainty","estimation","hot-reload","hot","reload","agent-orchestration","ai-routing","browser","circuit-breaker","inference","llm-routing","model-routing","model","neural-network","network","neural-router","uncertainty-estimation"],"s":["tiny-dancer","tiny","dancer","agent-orchestration","orchestration","ai-routing","rout","browser","circuit-breaker","circuit","breaker","fastgrnn","hot-reload","hot","reload","inference","llm-routing","model-routing","model","neural-network","neural","network","neural-router","router","uncertainty-estimation","uncertainty","estimation"]},
+{"id":"@ruvector/typesafe","family":"typesafe","kind":"npm","store":"ruvector","version":"0.2.0","description":"Local typed decisions (choice / score / noul) over sentence embeddings — a ruvector-powered, self-optimizing alternative to typesafe.ai's System One API. Native (napi-rs) with a WASM fallback; no network, no per-token cost.","keywords":["calibration","choice","classification","confidence","decisions","embeddings","jev","local-first","noul","ruvector","system-one","typesafe","zero-dependency"],"source":"ruvector/npm/packages/typesafe/package.json","sourceSha256":"87510561f99fc26e6da326b512e04242c6e1644a7f0728d77bf3bd9fb38e6815","variants":["ruvector-typesafe-core","ruvector-typesafe-ffi","ruvector-typesafe-wasm"],"readme":"Local typed decisions over sentence embeddings, with the wire contract of Jev ([typesafe.ai](https://typesafe.ai)) \"System One\". You send one text state plus a batch of questions — choice, score, noul — and get back per-question answers with confidence, an abstain mass, and a receipt naming the head, model and temperature that produced each answer. It is a bounded classifier, not an LLM host: no network by default, no per-token cost, no subprocess in the decision path. A native (napi-rs) core does the deciding; a WASM build is the portable fallback. Latency targets are release gates (p95 ≤ 50 ms native, ≤ 150 ms WASM — ADR-006); the engine's own numbers are measured by typesafe bench and rec","readmeSource":"ruvector/npm/packages/typesafe/README.md","product":"key:ruvector:typesafe","canonical":"@ruvector/typesafe","tier":"T0","t":["typesafe","local","typed","decision","choice","score","noul","sentence","embedding","ruvector-powered","power","self-optimizing","self","optimiz","alternative","typesafe.ai","fallback","network","per-token","per","token","cost","calibration","classification","confidence","jev","local-first","system-one","zero-dependency","zero","dependency"],"s":["typesafe","calibration","choice","classification","confidence","decision","embedding","jev","local-first","local","noul","system-one","zero-dependency","zero","dependency"]},
+{"id":"@ruvector/wasm","family":"wasm","kind":"npm","store":"ruvector","version":"0.2.0","description":"High-performance Rust vector database for browsers via WASM","keywords":["browser","database","embeddings","indexeddb","rust","simd","vector","wasm","web-workers"],"source":"ruvector/crates/ruvector-wasm/package.json","sourceSha256":"638e41de97af2639a02eacbb572c9e4e7afd1fa943543c4cf0b254b3d2e59343","variants":["ruvector-wasm"],"readme":"High-performance vector database running entirely in your browser via WebAssembly. In the age of privacy-first, offline-capable web applications, running AI workloads entirely in the browser is no longer optional—it's essential. Ruvector WASM brings enterprise-grade vector search to the browser: - ⚡ Blazing Fast: <1ms query latency with HNSW indexing and SIMD acceleration - 🔒 Privacy First: All data stays in the browser—zero server round-trips - 📴 Offline Capable: Full functionality without internet via IndexedDB persistence - 🌐 Edge Computing: Deploy to CDNs for ultra-low latency globally - 💾 Persistent Storage: IndexedDB integration with automatic synchronization - 🧵 Multi-threaded: W","readmeSource":"ruvector/crates/ruvector-wasm/README.md","product":"key:ruvector:wasm","canonical":"@ruvector/wasm","tier":"T0","t":["high","vector","database","browser","via","embedding","indexeddb","web-workers","web","worker"],"s":["browser","database","embedding","indexeddb","vector","web-workers","web","worker"]},
+{"id":"@ruvector/wasm-unified","family":"wasm-unified","kind":"npm","store":"ruvector","version":"1.0.0","description":"Unified TypeScript API surface for RuVector WASM - attention, learning, economy, and exotic features","keywords":["ai","attention","learning","machine-learning","neural","ruvector","wasm","webassembly"],"source":"ruvector/npm/packages/ruvector-wasm-unified/package.json","sourceSha256":"5b1b4d7fab686f172d1caa9c7870a439b6aefc08e4a8b53168e45faf17a3c36d","variants":[],"readme":"Unified TypeScript API surface for RuVector WASM - exposing attention, learning, economy, and exotic computation features through a clean, type-safe interface. - 14+ Attention Mechanisms: Neural (scaled-dot, multi-head, hyperbolic, linear, flash, local-global, MoE, Mamba) and DAG (topological, mincut-gated, hierarchical, spectral, flow, causal, sparse) - Adaptive Learning: Micro-LoRA adaptation, SONA pre-query, BTSP one-shot learning, RL algorithms, meta-learning - Nervous System Simulation: Spiking neural networks, synaptic plasticity, multiple neuron models - Compute Credit Economy: Balance management, staking, rewards, contribution multipliers - Exotic Computation: Quantum-inspired, hyper","readmeSource":"ruvector/npm/packages/ruvector-wasm-unified/README.md","product":"key:ruvector:wasm-unified","canonical":"@ruvector/wasm-unified","tier":"T0","t":["wasm-unified","unify","surface","attention","learn","economy","exotic","feature","machine-learning","machine","neural"],"s":["wasm-unified","unify","attention","learn","machine-learning","machine","neural"]},
+{"id":"@ruvnet/entrypoint","family":"entrypoint","kind":"npm","store":"ruvnet","version":"0.10.0","description":"Read-only RuV Stack discovery, integration plans, CLI and MCP","keywords":[],"source":"ruvnet/plugins/ruvnet/package.json","sourceSha256":"42b6e34c20bd5cfd268792c50ef5054cbbe464e22f26e91afba42004b52148a1","variants":[],"product":"key:ruvnet:entrypoint","canonical":"@ruvnet/entrypoint","tier":"T2","t":["entrypoint","read-only","read","only","stack","discovery","integration","plan","mcp"],"s":["entrypoint"]},
+{"id":"@ruvnet/nvsim-dashboard","family":"nvsim-dashboard","kind":"npm","store":"ruview","version":"0.1.0","description":"Vite + Lit dashboard for the nvsim NV-diamond magnetometer pipeline simulator (ADR-092).","keywords":[],"source":"ruview/dashboard/package.json","sourceSha256":"f91383c25d6a22175ae4516aa14f61b47d77af2932021069c80488a8be5a4e82","variants":[],"product":"key:ruview:nvsim-dashboard","canonical":"@ruvnet/nvsim-dashboard","tier":"T0","t":["nvsim-dashboard","nvsim","dashboard","vite","lit","nv-diamond","diamond","magnetometer","pipeline","simulator","adr-092","adr","092"],"s":["nvsim-dashboard","nvsim","dashboard"]},
+{"id":"@ruvnet/rulake-console","family":"rulake-console","kind":"npm","store":"rulake","version":"0.1.0","description":"ruLake Console — verifiable vector memory dashboard for AI agents (Vite + React)","keywords":[],"source":"rulake/ui/package.json","sourceSha256":"2f0093b9e2813b4ba03b3a1b5c1704f44e1719351ae0f5dd1f052a6b088c6a23","variants":[],"readme":"Verifiable vector memory dashboard for AI agents. Per [ADR-006](../docs/adrs/ADR-006-rulake-console-vite-github-pages.md). Deployed to https://ruvnet.github.io/RuLake/ on every push to main that touches ui/ or sdk/node-wasm/ (CI: .github/workflows/release-ui.yml). npm run build produces ui/dist/; npm run preview serves it at http://127.0.0.1:4173/RuLake/. Per ADR-006 §10, the console runs in three modes from the same build: 1. Demo — fixture data only, witness verification still cryptographically real via rulake-wasm. Default for the GitHub Pages URL. 2. WASM-local — full ruLake substrate in the browser via IndexedDB + Web Workers + WebGL + WASM. Optional cloud + IPFS for storage. 3. Live (M","readmeSource":"rulake/ui/README.md","product":"key:rulake:rulake-console","canonical":"@ruvnet/rulake-console","tier":"T1","t":["rulake-console","rulake","console","verifiable","vector","memory","dashboard","vite","react"],"s":["rulake-console","rulake","console"]},
+{"id":"@ruvnet/ruview","family":"ruview","kind":"npm","store":"ruview","version":"0.5.1","description":"RuView WiFi-sensing operator harness — onboard, calibrate, verify, enforce evidence guardrails, and read Cognitum Spaces through explicitly granted OAuth.","keywords":["agent-harness","ambient-intelligence","channel-state-information","claude-code","cli","cognitum","cognitum-spaces","csi","esp32","mcp","mcp-server","metaharness","oauth","pose-estimation","presence-detection","ruview","wifi","wifi-densepose","wifi-sensing"],"source":"ruview/harness/ruview/package.json","sourceSha256":"ec11f4244bc77be216dc54da0cb6585a646e2b9c3a94efbf72bd8b7206d1dbcf","variants":["@ruv/ruview-cli"],"readme":"An AI agent harness that knows how to operate RuView (WiFi-DensePose): onboard a newcomer, provision an ESP32 CSI node, calibrate a room, train pose models, and — crucially — refuse to overstate accuracy. Minted from the RuView monorepo via [metaharness](https://www.npmjs.com/package/metaharness) and hardened per ADR-182. WiFi sensing infers *coarse* pose/presence/breathing from Channel State Information. It is not a camera. Every accuracy number this harness emits must be MEASURED against a baseline — that rule is enforced in code (ruview_claim_check). The operator tools are pure Node and the published package has no runtime dependencies (ADR-263 O3). MetaHarness, Darwin and Flywheel are ex","readmeSource":"ruview/harness/ruview/README.md","product":"repo:ruview","canonical":"@ruvnet/ruview","tier":"T0","t":["ruview","wifi-sensing","wifi","sens","operator","harness","onboard","calibrate","verify","enforce","evidence","guardrail","read","cognitum","space","through","explicitly","grant","oauth","agent-harness","ambient-intelligence","ambient","intelligence","channel-state-information","channel","state","information","claude-code","claude","code","cognitum-spaces","csi","esp32","mcp","mcp-server","server","metaharness","pose-estimation","pose","estimation","presence-detection","presence","detection","wifi-densepose","densepose"],"s":["ruview","agent-harness","harness","ambient-intelligence","ambient","intelligence","channel-state-information","channel","state","information","claude-code","claude","code","cognitum","cognitum-spaces","space","csi","esp32","mcp","mcp-server","server","metaharness","oauth","pose-estimation","pose","estimation","presence-detection","presence","detection","wifi","wifi-densepose","densepose","wifi-sensing","sens"]},
+{"id":"@ruvnet/ruvn","family":"ruvn","kind":"npm","store":"ruvn","version":"0.1.1","description":"Gamma-entrainment protocol research agent — compare modalities/dosing, grade evidence, verify signed session bundles (ruv-neural)","keywords":["agent-harness","dossier","fact-checking","research"],"source":"ruvn/package.json","sourceSha256":"dcb53e04078f2c1659bbcc01c1f61f7ffe6e96aaa621667c606673823676f92c","variants":[],"readme":"A research agent that turns a question into a graded, cited evidence dossier — built for gamma-entrainment (40 Hz) protocol research. ruvn is a small AI-agent \"harness\": a set of specialist agents plus the config to drop them into your AI coding host (Claude Code, Codex, Copilot, and more). You ask a research question; it runs a disciplined pipeline that searches, grades every source, synthesizes only from the good ones, adversarially fact-checks the result, and hands back a dossier where every claim is cited. It was built for the [ruv-neural](https://github.com/ruvnet/ruv-neural) project — comparing 40 Hz stimulation modalities, dosing, responder profiles, and safety evidence — but it works","readmeSource":"ruvn/README.md","product":"repo:ruvn","canonical":"@ruvnet/ruvn","tier":"T2","t":["ruvn","gamma-entrainment","gamma","entrainment","protocol","research","compare","modality","dosing","grade","evidence","verify","sign","session","bundle","ruv-neural","neural","agent-harness","harness","dossier","fact-checking","fact","check"],"s":["ruvn","agent-harness","harness","dossier","fact-checking","fact","check","research"]},
+{"id":"@ruvnet/rvagent","family":"rvagent","kind":"npm","store":"ruview","version":"0.2.0","description":"SENSE-BRIDGE: dual-transport MCP server (stdio default; Streamable HTTP opt-in via RVAGENT_HTTP_PORT) exposing RuView WiFi-DensePose sensing primitives to AI agents","keywords":["cognitum","csi","mcp","pose-estimation","ruview","ruvnet","rvagent","sense-bridge","wifi"],"source":"ruview/tools/ruview-mcp/package.json","sourceSha256":"4ea53234ef8471038baa7da0a5c3049f7a1817617dce0d82e905791cd0b1c722","variants":[],"readme":"SENSE-BRIDGE is a dual-transport [Model Context Protocol](https://modelcontextprotocol.io/) (MCP) server that bridges the RuView WiFi-DensePose sensing stack to AI agents (Claude Code, Cursor, ruflo swarms, and any MCP-compatible client). Install once; AI agents can then call ruview_presence_now, ruview_vitals_get_heart_rate, ruview_bfld_last_scan, and more — without writing HTTP or WebSocket client code. Requirements: Node.js >= 20. The wifi-densepose-sensing-server Rust binary must be reachable at RUVIEW_SENSING_SERVER_URL (default http://localhost:3000). Canonical tool names are underscore-form (ADR-264 — host tool-name validators commonly enforce ^[a-zA-Z0-9_-]{1,64}$). The pre-0.1.1 dot","readmeSource":"ruview/tools/ruview-mcp/README.md","product":"key:ruview:rvagent","canonical":"@ruvnet/rvagent","tier":"T0","t":["rvagent","sense-bridge","sense","bridge","dual-transport","dual","transport","mcp","server","stdio","default","streamable","http","opt-in","opt","via","port","expos","ruview","wifi-densepose","wifi","densepose","sens","primitive","cognitum","csi","pose-estimation","pose","estimation"],"s":["rvagent","cognitum","csi","mcp","pose-estimation","pose","estimation","ruview","sense-bridge","sense","bridge","wifi"]},
+{"id":"@ruvnet/sparc","family":"sparc","kind":"npm","store":"sparc","version":"1.0.0","description":"Deterministic SPARC metaharness with MCP transports and shared Claude and ChatGPT skills","keywords":["chatgpt","claude","codex","mcp","metaharness","sparc"],"source":"sparc/package.json","sourceSha256":"85ce7b714f4db9b8d1a74a914d2de16bffbde86fde36bbc6e147ce5a4d3c5e3d","variants":[],"readme":"SPARC is an evidence-gated engineering metaharness for Specification, Pseudocode, Architecture, Refinement, and Completion. Version 1.0 adds an executable TypeScript state machine, an npm CLI, local and remote MCP transports, shared Claude and OpenAI skills, deterministic phase gates, principal isolation, compare and set revisions, exact idempotency, atomic persistence, and tamper-evident receipts. The established Python CLI remains available as a compatibility surface. The five phases are enforced by code, not suggested by a prompt. SPARC records development evidence. It never executes an acceptance-test command stored in a run and never treats repository or model text as authority. The npm","readmeSource":"sparc/README.md","product":"repo:sparc","canonical":"@ruvnet/sparc","tier":"T1","t":["sparc","deterministic","metaharness","mcp","transport","shar","claude","chatgpt","skill","codex"],"s":["sparc","chatgpt","claude","codex","mcp","metaharness"]},
+{"id":"@synaptic-mesh/mcp-integration","family":"mcp-integration","kind":"npm","store":"synaptic-mesh","version":"1.0.0","description":"Model Context Protocol integration for Synaptic Neural Mesh","keywords":["dag-consensus","distributed-ai","mcp","model-context-protocol","neural-mesh","synaptic-mesh"],"source":"synaptic-mesh/src/mcp/package.json","sourceSha256":"2a0b7eea6a78ac374afd33c646080d8adf045aea80cfb511a4466a9710f15785","variants":[],"readme":"The Synaptic Neural Mesh MCP Server now includes comprehensive integration with Kimi-K2 AI models through multiple providers. - Multi-Provider Support: Moonshot AI, OpenRouter, and local LLM support - Tool Calling: Full function calling capabilities - 128k Context Window: Extended context management for long conversations - Streaming Responses: Real-time response streaming - Error Handling: Robust error handling and fallback mechanisms - Context Management: Intelligent context window management - Native Kimi-K2 models - 128k context window - Optimized for Chinese and English - Best performance for neural mesh tasks - Access to multiple AI providers - Claude, GPT-4, Llama, and more - Up to 20","readmeSource":"synaptic-mesh/src/mcp/README.md","product":"key:synaptic-mesh:mcp-integration","canonical":"@synaptic-mesh/mcp-integration","tier":"T1","t":["mcp-integration","mcp","integration","model","context","protocol","synaptic","neural","mesh","dag-consensus","dag","consensus","distributed-ai","distribut","model-context-protocol","neural-mesh","synaptic-mesh"],"s":["synaptic-mesh","synaptic","mesh","mcp-integration","mcp","integration","dag-consensus","dag","consensus","distributed-ai","distribut","model-context-protocol","model","context","protocol","neural-mesh","neural"]},
+{"id":"@synaptic-neural-mesh/claude-container","family":"claude-container","kind":"npm","store":"synaptic-mesh","version":"1.0.0","description":"Minimal, secure Claude container for task execution","keywords":["ai","claude","container","sandbox","security","synaptic-neural-mesh"],"source":"synaptic-mesh/docker/claude-container/package.json","sourceSha256":"08b031cd5c735e071d00775624cf12b480e7cd990b83a60ef166a59469039640","variants":[],"readme":"A minimal, secure Docker container for running Claude AI tasks in a sandboxed environment. - ✅ Local API Access Only: Each node runs Claude with their own API credentials - ✅ No Shared Secrets: No API keys or proxy access shared between nodes - ✅ Secure Sandbox: Read-only filesystem with tmpfs workspace - ✅ User Control: User maintains full control with their own API keys - ✅ Network Isolation: API access only, no lateral network access - ✅ No Persistent State: Tmpfs workspace cleared on container restart - Alpine Linux Base: Minimal attack surface (< 50MB) - Non-root User: Runs as uid:1000, no privileged access - Read-only Filesystem: Container filesystem is immutable - Resource Limits: 512","readmeSource":"synaptic-mesh/docker/claude-container/README.md","product":"key:synaptic-mesh:claude-container","canonical":"@synaptic-neural-mesh/claude-container","tier":"T1","t":["claude-container","claude","container","minimal","secure","task","execution","sandbox","security","synaptic-neural-mesh","synaptic","neural","mesh"],"s":["synaptic-neural-mesh","synaptic","neural","mesh","claude-container","claude","container","sandbox","security"]},
+{"id":"@synaptic-neural-mesh/claude-max-market","family":"claude-max-market","kind":"npm","store":"synaptic-mesh","version":"1.0.0","description":"NPX wrapper for Claude Max Docker orchestration and market integration with compliance","keywords":["claude","compliance","docker","market","npx","orchestration","synaptic-neural-mesh"],"source":"synaptic-mesh/packages/claude-max-market/package.json","sourceSha256":"754e5b875ac2487137de8c3c97261c5442c3d6206c3db2e6451bcd9bf909a934","variants":[],"readme":"NPX wrapper for Docker orchestration and market integration with comprehensive compliance features for Claude Max capacity sharing. This implementation is designed to be fully compliant with Anthropic's Terms of Service by ensuring: - ✅ No shared API keys - Each user maintains their own Claude credentials - ✅ Peer orchestrated model - Tasks route to willing participants, not centralized accounts - ✅ Voluntary participation - Users opt-in explicitly with granular controls - ✅ User control & transparency - Full approval and audit mechanisms - ✅ Token incentives - RUV tokens reward contribution, not access purchase The system uses a comprehensive configuration system with these main sections:","readmeSource":"synaptic-mesh/packages/claude-max-market/README.md","product":"key:synaptic-mesh:claude-max-market","canonical":"@synaptic-neural-mesh/claude-max-market","tier":"T1","t":["claude-max-market","claude","max","market","npx","wrapper","docker","orchestration","integration","compliance","synaptic-neural-mesh","synaptic","neural","mesh"],"s":["synaptic-neural-mesh","synaptic","neural","mesh","claude-max-market","claude","max","market","compliance","docker","npx","orchestration"]},
+{"id":"@worldgraph/supersplat-bridge","family":"supersplat-bridge","kind":"npm","store":"worldgraph","version":"0.3.1","description":"TypeScript / PlayCanvas integration layer that renders the WorldGraph digital twin over a SuperSplat Gaussian splat via WASM (ADR-200/201/202)","keywords":["digital-twin","gaussian-splat","playcanvas","privacy","supersplat","wasm","worldgraph"],"source":"worldgraph/supersplat-bridge/package.json","sourceSha256":"e3725fce178ac45c8ffadfe91e05c55184ed219b92e1e21b0a942c3e89c10099","variants":[],"readme":"TypeScript / PlayCanvas integration that renders the [WorldGraph](https://github.com/ruvnet/worldgraph) privacy-aware digital twin over a [SuperSplat](https://github.com/playcanvas/supersplat) Gaussian splat. It supports local WebAssembly twins and authenticated live snapshot/delta streams. See the repo's [INTEGRATION.md](../INTEGRATION.md) for the end-to-end build and [docs/adr/ADR-200..203](../docs/adr/) for the design. The Rust→WASM module (src/worldgraph-wasm/) is produced separately by ../worldgraph-wasm/build-wasm.sh and is git-ignored. - SemanticVisualizer — typed wrapper over the WASM WorldgraphBridge; the module is injected so it unit-tests without a browser. - WorldgraphScene + Pla","readmeSource":"worldgraph/supersplat-bridge/README.md","product":"key:worldgraph:supersplat-bridge","canonical":"@worldgraph/supersplat-bridge","tier":"T2","t":["supersplat-bridge","supersplat","bridge","playcanva","integration","layer","render","worldgraph","digital","twin","gaussian","splat","via","adr-200","adr","200","201","202","digital-twin","gaussian-splat","privacy"],"s":["worldgraph","supersplat-bridge","supersplat","bridge","digital-twin","digital","twin","gaussian-splat","gaussian","splat","playcanva","privacy"]},
+{"id":"agent-booster","family":"agent-booster","kind":"npm","store":"agentic-flow","version":"0.2.2","description":"Ultra-fast code editing engine - 52x faster than Morph LLM at $0 cost","keywords":["ast","code-editing","morph","performance","refactoring","rust","wasm"],"source":"agentic-flow/packages/agent-booster/package.json","sourceSha256":"6cae7b7cf95b71a45cd5daf5abccbdf8d18fa13a01eada4465bac86c45aebd39","variants":["agent-booster-native"],"readme":"Agent Booster is a high-performance code transformation engine designed to eliminate the latency and cost bottleneck in AI coding agents, autonomous systems, and developer tools. Built in Rust with WebAssembly, it applies code edits 350x faster than LLM-based alternatives while maintaining 100% accuracy. When building AI coding agents, LLM-based code application APIs create severe bottlenecks: - 🐌 Slow: 200-500ms latency per edit blocks agent execution - 💸 Expensive: $0.01+ per edit = $100+ monthly costs for active agents - 🔒 Privacy Concerns: Code must be sent to external APIs - ⚠️ Unreliable: Non-deterministic results, rate limits, network issues Agent Booster solves all of these: - ⚡ I","readmeSource":"agentic-flow/packages/agent-booster/README.md","product":"key:agentic-flow:agent-booster","canonical":"agent-booster","tier":"T1","t":["agent-booster","booster","ultra-fast","ultra","code","edit","52x","than","morph","cost","ast","code-editing","refactor"],"s":["agent-booster","booster","ast","code-editing","code","edit","morph","refactor"]},
+{"id":"agent-booster-wasm","family":"agent-booster","kind":"npm","store":"agentic-qe","version":"0.1.0","description":"WebAssembly bindings for agent-booster","keywords":[],"source":"agentic-qe/src/integrations/agent-booster-wasm/package.json","sourceSha256":"13d54d8928f72ca0857ff26e7698f3af4a04d9307ea60454613ab9ede476969e","variants":[],"readme":"WebAssembly bindings for the Agent Booster library, providing fast code editing capabilities using tree-sitter and similarity matching in Node.js and browser environments. Phase: Initial implementation complete, waiting for core library API finalization. The WASM bindings infrastructure is ready with: - ✅ Type definitions and conversions - ✅ WASM-bindgen setup - ✅ JavaScript interop layer - ⏳ Core apply_edit function (pending core library implementation) Once published, you can install via npm: 1. Install Rust (1.70+): 2. Add WASM target: 3. Install wasm-pack: For dynamic use cases, you can use the JSON API: Main interface for applying code edits. - new() - Create with default configuration","readmeSource":"agentic-qe/src/integrations/agent-booster-wasm/README.md","product":"key:agentic-qe:agent-booster","canonical":"agent-booster-wasm","tier":"T1","t":["agent-booster-wasm","booster","agent-booster"],"s":["agent-booster-wasm","booster","agent-booster"]},
+{"id":"agent-name-service-agent","family":"agent-name-service-agent","kind":"npm","store":"agent-name-service","version":"0.1.0","description":"Maintain and validate agent-name-service","keywords":[],"source":"agent-name-service/.harness/generated/package.json","sourceSha256":"faed73d6fa5e4ffc7ac67423acbcd33c20cb92a145f781e2314fd83fbd7066ef","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"agent-name-service/.harness/generated/README.md","product":"key:agent-name-service:agent-name-service-agent","canonical":"agent-name-service-agent","tier":"T3","t":["agent-name-service-agent","name","service","maintain","validate","agent-name-service"],"s":["agent-name-service-agent","name","service"]},
+{"id":"agentdb","family":"agentdb","kind":"npm","store":"agentdb","version":"3.0.0-alpha.20","description":"Self-learning vector memory for AI agents — single-file .rvf cognitive container with HNSW search, episodic Reflexion memory, causal graph + Cypher, 9 RL algorithms, Thompson Sampling bandit, 41 MCP tools, hybrid (BM25 + dense) retrieval, GNN attention. 150x faster than SQLite. Up to +36% search quality from feedback. Runs in Node, browsers (WASM), edge, offline. Backs the agentic-flow / ruflo / Claude Code memory layer.","keywords":["agent-database","agent-memory","agent-orchestration","agentdb","agentic","agentic-ai","agentic-flow","ai-agents","ann","anthropic","attention","audit-log","bandit","bm25","browser","causal-graph","causal-reasoning","claude-code","claude-code-plugin","cypher","edge","embedding-search","embeddings","episodic-memory","explainable-ai","explainable-recall","gnn","graph-database","hnsw","hybrid-search","knn","knowledge-graph","lifelong-learning","llm-memory","llm-tools","mcp","memory","mmr","model-context-protocol","napi","offline-first","provenance","rag","reflexion","reinforcement-learning","rerank","retrieval-augmented-generation","rl","rrf","ruflo","rust","ruvector","self-improving","self-learning","semantic-search","skill-library","sql.js","sqlite","thompson-sampling","typescript","vector-database","vector-search","vector-store","wasm"],"source":"agentdb/package.json","sourceSha256":"3f93359be65f3304f6f066b51f47f1cc3ce7145c4d52b69be13a03057bdb7f66","variants":[],"readme":"Vector memory that gets smarter every time your agent uses it. A single-file cognitive container — vectors, indexes, learning state, and a cryptographic audit trail in one .rvf. Self-learning search improves up to 36% from feedback alone, with no manual tuning. Runs in Node, the browser, edge runtimes, and offline. There are three ways to use AgentDB depending on what you're building. Pick whichever matches your stack:","readmeSource":"agentdb/README.md","product":"repo:agentdb","canonical":"agentdb","tier":"T1","t":["agentdb","self-learning","self","learn","vector","memory","single-file","single","file","rvf","cognitive","container","hnsw","search","episodic","reflexion","causal","graph","cypher","algorithm","thompson","sampl","bandit","mcp","hybrid","bm25","dense","retrieval","gnn","attention","150x","than","sqlite","quality","feedback","browser","edge","offline","back","agentic-flow","flow","ruflo","claude","code","layer","agent-database","database","agent-memory","agent-orchestration","orchestration","agentic-ai","ai-agents","ann","anthropic","audit-log","audit","log","causal-graph","causal-reasoning","reason","claude-code","claude-code-plugin","plugin","embedding-search","embedd","embedding","episodic-memory","explainable-ai","explainable","explainable-recall","recall","graph-database","hybrid-search","knn","knowledge-graph","knowledge","lifelong-learning","lifelong","llm-memory","llm-tools","mmr","model-context-protocol","model","context","protocol","offline-first","provenance","rag","reinforcement-learning","reinforcement","rerank","retrieval-augmented-generation","augment","generation","rrf","self-improving","improv","semantic-search","semantic","skill-library","skill","sql.j","thompson-sampling","vector-database","vector-search","vector-store","store"],"s":["agentdb","agent-database","database","agent-memory","memory","agent-orchestration","orchestration","agentic-ai","agentic-flow","flow","ai-agents","ann","anthropic","attention","audit-log","audit","log","bandit","bm25","browser","causal-graph","causal","graph","causal-reasoning","reason","claude-code","claude","code","claude-code-plugin","plugin","cypher","edge","embedding-search","embedd","search","embedding","episodic-memory","episodic","explainable-ai","explainable","explainable-recall","recall","gnn","graph-database","hnsw","hybrid-search","hybrid","knn","knowledge-graph","knowledge","lifelong-learning","lifelong","learn","llm-memory","llm-tools","mcp","mmr","model-context-protocol","model","context","protocol","offline-first","offline","provenance","rag","reflexion","reinforcement-learning","reinforcement","rerank","retrieval-augmented-generation","retrieval","augment","generation","rrf","ruflo","self-improving","self","improv","self-learning","semantic-search","semantic","skill-library","skill","sql.j","sqlite","thompson-sampling","thompson","sampl","vector-database","vector","vector-search","vector-store","store"]},
+{"id":"agentdb-onnx","family":"agentdb-onnx","kind":"npm","store":"agentic-flow","version":"1.0.0","description":"AgentDB with optimized ONNX embeddings - 100% local, GPU-accelerated AI agent memory","keywords":["agentdb","ai-agents","embeddings","gpu-acceleration","local-inference","onnx","vector-database"],"source":"agentic-flow/packages/agentdb-onnx/package.json","sourceSha256":"861c528432abef08d86dba961b477979abd6b9cd8f8a83c5148623ec3fdf9823","variants":[],"readme":"100% local, GPU-accelerated embeddings with AgentDB vector storage for intelligent AI agents. - No API calls, no cloud dependencies - Complete data privacy - Zero latency overhead from network requests - Free unlimited embeddings - ONNX Runtime with CUDA support (Linux/Windows) - DirectML support (Windows) - CoreML support (macOS) - Automatic fallback to CPU - ⚡ Batch processing: 3-4x faster than sequential - 💾 LRU caching: 80%+ hit rate for common queries - 🔥 Model warmup: Pre-JIT compilation for consistent latency - 📊 Smart batching: Automatic chunking for large datasets - ReasoningBank: Store and retrieve successful patterns - Reflexion Memory: Self-improving episodic memory - Comprehe","readmeSource":"agentic-flow/packages/agentdb-onnx/README.md","product":"key:agentic-flow:agentdb-onnx","canonical":"agentdb-onnx","tier":"T1","t":["agentdb-onnx","agentdb","onnx","embedding","100","local","gpu-accelerated","gpu","accelerat","memory","ai-agents","gpu-acceleration","acceleration","local-inference","inference","vector-database","vector","database"],"s":["agentdb-onnx","agentdb","onnx","ai-agents","embedding","gpu-acceleration","gpu","acceleration","local-inference","local","inference","vector-database","vector","database"]},
+{"id":"agentic-devops-agent","family":"agentic-devops-agent","kind":"npm","store":"agentic-devops","version":"0.1.0","description":"Maintain and validate agentic-devops","keywords":[],"source":"agentic-devops/.harness/generated/package.json","sourceSha256":"6644a1d006493aca445097a3846f83d07e36dfa2e3ddca65afd2293756b71d89","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"agentic-devops/.harness/generated/README.md","product":"key:agentic-devops:agentic-devops-agent","canonical":"agentic-devops-agent","tier":"T3","t":["agentic-devops-agent","devop","maintain","validate","agentic-devops"],"s":["agentic-devops-agent","devop"]},
+{"id":"agentic-flow","family":"agentic-flow","kind":"npm","store":"agentic-flow","version":"2.0.2-alpha","description":"Production-ready AI agent orchestration platform with 66 specialized agents, 213 MCP tools, ReasoningBank learning memory, quantum-resistant Jujutsu VCS, and autonomous multi-agent swarms. Built by @ruvnet with Claude Agent SDK, neural networks, memory persistence, GitHub integration, and distributed consensus protocols.","keywords":[],"source":"agentic-flow/package.json","sourceSha256":"d275a1c924db5a3ab98c87dcf12e5004ec2297a965441dce4d39c9f9f5a171ee","variants":[],"readme":"That's it! Your project now has: - 🧠 Self-learning hooks that improve agent routing over time - 🤖 80+ specialized agents (coder, tester, reviewer, architect, etc.) - ⚡ Background workers triggered by keywords (ultralearn, optimize, audit) - 📊 213 MCP tools for swarm coordination Agentic-Flow v2 now includes SONA (@ruvector/sona) for sub-millisecond adaptive learning: - 🎓 +55% Quality Improvement: Research profile with LoRA fine-tuning - ⚡ <1ms Learning Overhead: Sub-millisecond pattern learning and retrieval - 🔄 Continual Learning: EWC++ prevents catastrophic forgetting - 💡 Pattern Discovery: 300x faster pattern retrieval (150ms → 0.5ms) - 💰 60% Cost Savings: LLM router with intellige","readmeSource":"agentic-flow/README.md","product":"repo:agentic-flow","canonical":"agentic-flow","tier":"T1","t":["agentic-flow","flow","production-ready","orchestration","specializ","213","mcp","reasoningbank","learn","memory","quantum-resistant","quantum","resistant","jujutsu","vcs","autonomous","multi-agent","multi","swarm","built","claude","neural","network","persistence","github","integration","distribut","consensus","protocol"],"s":["agentic-flow","flow"]},
+{"id":"agentic-flow-quic","family":"agentic-flow-quic","kind":"npm","store":"agentic-flow","version":"0.1.0","description":"High-performance QUIC transport layer for agentic-flow with WASM support","keywords":[],"source":"agentic-flow/agentic-flow/wasm/quic/package.json","sourceSha256":"9261f0652ec5ed9b5d858b70ab572f5e34d6a01deab0631e0af885f1b652885d","variants":[],"readme":"High-performance QUIC transport layer for agentic-flow with WASM support. - QUIC Client with connection pooling - QUIC Server with stream multiplexing - 0-RTT connection establishment - BBR congestion control - WASM bindings for browser usage - Bidirectional streams - Connection migration handling MIT","readmeSource":"agentic-flow/agentic-flow/wasm/quic/README.md","product":"key:agentic-flow:agentic-flow-quic","canonical":"agentic-flow-quic","tier":"T1","t":["agentic-flow-quic","flow","quic","high","transport","layer","agentic-flow"],"s":["agentic-flow-quic","flow","quic"]},
+{"id":"agentic-jujutsu","family":"agentic-jujutsu","kind":"crate","store":"agentic-flow","version":null,"description":"AI-powered Jujutsu VCS wrapper for multi-agent collaboration - 10-100x faster than Git with MCP protocol support","keywords":["ai-agents","jujutsu","mcp","vcs","wasm"],"source":"agentic-flow/packages/agentic-jujutsu/Cargo.toml","sourceSha256":"448caca99c6814b6ebd8d4c795c3e5893c9f0dd13fd83ab072ec36c1861fd20a","variants":[],"readme":"- [⚡ Quick Start](#-quick-start---try-it-now) - Get started in 30 seconds - [🔧 Installation](#-installation) - Complete installation guide - [🏗️ Architecture](#-architecture) - How it works - [🚀 CLI Commands](#-npx-cli-commands---complete-reference) - All npx commands - [🤖 MCP Tools](#-mcp-tools-for-ai-agents---quick-reference) - AI agent integration - [🎯 Use Cases](#-ai-coding-agent-use-cases) - Real-world examples - [🔗 Rust/Cargo](#-rustcargo-advanced-use) - Advanced Rust usage - [📖 Full Documentation](#-links--resources) - More resources Quantum-ready, self-learning version control for AI agents - like Git, but designed for multiple AI agents working on code simultaneously without","readmeSource":"agentic-flow/packages/agentic-jujutsu/README.md","product":"key:agentic-flow:agentic-jujutsu","canonical":"agentic-jujutsu","tier":"T1","t":["agentic-jujutsu","jujutsu","ai-powered","power","vcs","wrapper","multi-agent","multi","collaboration","10-100x","100x","than","git","mcp","protocol","ai-agents"],"s":["agentic-jujutsu","jujutsu","ai-agents","mcp","vcs"]},
+{"id":"agentic-qe","family":"agentic-qe","kind":"npm","store":"agentic-qe","version":"3.14.6","description":"Agentic Quality Engineering V3 - Domain-Driven Design Architecture with 13 Bounded Contexts, O(log n) coverage analysis, ReasoningBank learning, 60 specialized QE agents, mathematical Coherence verification, deep Claude Flow integration","keywords":[],"source":"agentic-qe/package.json","sourceSha256":"848d01654d106e93986d352036b4d06f75fee95ba93ed4056a058665100588f9","variants":[],"readme":"[Release Notes](docs/releases/README.md) | [Changelog](CHANGELOG.md) | [Issues](https://github.com/proffesor-for-testing/agentic-qe/issues) | [Discussions](https://github.com/proffesor-for-testing/agentic-qe/discussions) AI-powered quality engineering agents that generate tests, find coverage gaps, detect flaky tests, and learn your codebase patterns — across 11 coding agent platforms. - Generates comprehensive tests automatically — unit, integration, property-based, and BDD scenarios for your codebase with framework-specific output (Jest, Vitest, Playwright, Cypress, pytest, JUnit, Go, Rust, Swift, Flutter, and more) - Finds coverage gaps and prioritizes what to test — risk-weighted analysi","readmeSource":"agentic-qe/README.md","product":"repo:agentic-qe","canonical":"agentic-qe","tier":"T1","t":["agentic-qe","quality","engineer","domain-driven","domain","driven","design","architecture","bound","context","log","coverage","analysis","reasoningbank","learn","specializ","mathematical","coherence","verification","deep","claude","flow","integration"],"s":["agentic-qe"]},
+{"id":"agentic-reports-agent","family":"agentic-reports-agent","kind":"npm","store":"agentic-reports","version":"0.1.0","description":"Maintain and validate agentic-reports","keywords":[],"source":"agentic-reports/.harness/generated/package.json","sourceSha256":"0d21a582ef140a51145b542ffd5b1f742f803d76e827b6250e84ee2c74790f8d","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"agentic-reports/.harness/generated/README.md","product":"key:agentic-reports:agentic-reports-agent","canonical":"agentic-reports-agent","tier":"T3","t":["agentic-reports-agent","report","maintain","validate","agentic-reports"],"s":["agentic-reports-agent","report"]},
+{"id":"agentic-robotics","family":"agentic-robotics","kind":"npm","store":"agentic-robotics","version":"0.2.5","description":"High-performance agentic robotics framework with ROS2 compatibility - Complete toolkit","keywords":["agents","cli","mcp","middleware","robotics","ros","ros2"],"source":"agentic-robotics/npm/agentic-robotics/package.json","sourceSha256":"d78776ada09d50a620d0b49e740b2b822360bce13f0d7c588d7c41e6017dfdba","variants":[],"readme":"📚 [Documentation](https://ruv.io/agentic-robotics/docs) | 🚀 [Getting Started](#-quick-start) | 💬 [Community](https://github.com/ruvnet/agentic-robotics/discussions) | 🐛 [Issues](https://github.com/ruvnet/agentic-robotics/issues) Agentic Robotics is a revolutionary framework that bridges the gap between traditional robotics and modern AI. It empowers developers to build intelligent, self-learning robots that can perceive, decide, and act autonomously in complex environments. Traditional robotics frameworks require extensive programming for every scenario. Agentic Robotics changes that: - 🧠 AI-First Design: Built-in integration with Large Language Models (LLMs) like Claude, GPT-4, and mor","readmeSource":"agentic-robotics/npm/agentic-robotics/README.md","product":"repo:agentic-robotics","canonical":"agentic-robotics","tier":"T2","t":["agentic-robotics","robotic","high","ros2","compatibility","complete","toolkit","mcp","middleware","ros"],"s":["agentic-robotics","robotic","mcp","middleware","ros","ros2"]},
+{"id":"agentic-search-agent","family":"agentic-search-agent","kind":"npm","store":"agentic-search","version":"0.1.0","description":"Maintain and validate agentic-search","keywords":[],"source":"agentic-search/.harness/generated/package.json","sourceSha256":"04a612b6eafd60452266fd101bdc9f0e1ae978326a056a3947ec62fffffae858","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"agentic-search/.harness/generated/README.md","product":"key:agentic-search:agentic-search-agent","canonical":"agentic-search-agent","tier":"T3","t":["agentic-search-agent","search","maintain","validate","agentic-search"],"s":["agentic-search-agent","search"]},
+{"id":"agentic-voice-agent","family":"agentic-voice-agent","kind":"npm","store":"agentic-voice","version":"0.1.0","description":"Maintain and validate agentic-voice","keywords":[],"source":"agentic-voice/.harness/generated/package.json","sourceSha256":"98382e11c7ad3194025693b2502871009d35129d50c5bb33b45d67d852196900","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"agentic-voice/.harness/generated/README.md","product":"key:agentic-voice:agentic-voice-agent","canonical":"agentic-voice-agent","tier":"T3","t":["agentic-voice-agent","voice","maintain","validate","agentic-voice"],"s":["agentic-voice-agent","voice"]},
+{"id":"agenticow","family":"agenticow","kind":"npm","store":"agenticow","version":"0.2.4","description":"Git for Agent Memory: Copy-On-Write vector branching for embedded multi-agent memory. Branch a base memory in ~0.5ms / 162 bytes regardless of base size — 83x faster, 3000x smaller than full-copy snapshots. Exact read-through queries (parent ∪ edits, child wins).","keywords":["agent-memory","ai-agents","checkpoint","copy-on-write","cow","cow-vector-store","embedded-vector-db","git-for-vectors","llm-memory","memory-checkpointing","multi-agent-memory","rollback","ruvector","rvf","vector-branching","vector-database","vector-database-branching","vector-snapshot"],"source":"agenticow/package.json","sourceSha256":"bbfa5802d647ef540f2d511df2f111348d150bc519067f35bbfa1be5e9837c47","variants":[],"readme":"Branch a base vector memory in ~0.5 ms / 162 bytes — independent of base size. Exact read-through queries (parent ∪ edits, child wins). Built for embedded multi-agent memory. [Website / Demo →](https://ruvnet.github.io/agenticow/) · [npm](https://www.npmjs.com/package/agenticow) · [Benchmarks](#benchmarks) · [Acceptance proof](#acceptance-the-1000-branch-proof) Every other vector store makes you full-copy the index to snapshot, fork, or checkpoint it. agenticow branches it — copy-on-write, like Git. Creating a branch costs ~0.5 ms and 162 bytes whether the base holds 10,000 or 1,000,000 vectors. Query a branch and you transparently see parent ∪ your edits, with the child winning on id collis","readmeSource":"agenticow/README.md","product":"repo:agenticow","canonical":"agenticow","tier":"T1","t":["agenticow","git","memory","copy-on-write","copy","write","vector","branch","embed","multi-agent","multi","base","0.5ms","162","byte","regardless","size","83x","3000x","smaller","than","full-copy","full","snapshot","exact","read-through","read","through","query","parent","edit","child","wins","agent-memory","ai-agents","checkpoint","cow","cow-vector-store","store","embedded-vector-db","git-for-vectors","llm-memory","memory-checkpointing","multi-agent-memory","rollback","rvf","vector-branching","vector-database","database","vector-database-branching","vector-snapshot"],"s":["agenticow","agent-memory","memory","ai-agents","checkpoint","copy-on-write","copy","write","cow","cow-vector-store","vector","store","embedded-vector-db","embed","git-for-vectors","git","llm-memory","memory-checkpointing","multi-agent-memory","multi","rollback","rvf","vector-branching","branch","vector-database","database","vector-database-branching","vector-snapshot","snapshot"]},
+{"id":"agenticsjs","family":"agenticsjs","kind":"npm","store":"agenticsjs","version":"1.0.5","description":"AgenticJS is a powerful and flexible JavaScript library designed to provide an intelligent and interactive search experience with real-time results and advanced visualization.","keywords":["tailwindcss","vite"],"source":"agenticsjs/package.json","sourceSha256":"00caaffb47e087ac24b10c5a0d6350479db41d8973807f2e26027b55e428df72","variants":[],"readme":"AgenticsJS is a powerful and flexible JavaScript library designed to provide an intelligent and interactive search experience with real-time results, advanced visualization, and plugin capabilities. Built with Vite, React, and Tailwind CSS, AgenticsJS offers a seamless integration with modern web development workflows. It uses the same approach as Perplexity, the o1 for ChatGPT UI, and you.com style display of sequential queries and results. https://agenticjs.gptengineer.run - Real-time Search: Get instant search results as you type. - Advanced Visualization: Visualize search results with interactive charts and graphs. - Customizable: Easily customize the look and feel to match your brand. -","readmeSource":"agenticsjs/README.md","product":"repo:agenticsjs","canonical":"agenticsjs","tier":"T3","t":["agenticsj","agenticj","powerful","flexible","design","provide","intelligent","interactive","search","experience","real","result","advanc","visualization","tailwindcss","vite"],"s":["agenticsj","tailwindcss","vite"]},
+{"id":"agentvm-capability","family":"agentvm-capability","kind":"crate","store":"cognitum-cogs","version":null,"description":"Capability protocol implementation for Agentic VM","keywords":["agentic","capability","ocap","security","vm"],"source":"cognitum-cogs/crates/agentvm-capability/Cargo.toml","sourceSha256":"f8e58a79037906145e15dad9eef278d8cc5f61e61bd285762c2ae0f65c95f36d","variants":[],"product":"key:cognitum-cogs:agentvm-capability","canonical":"agentvm-capability","tier":null,"t":["agentvm-capability","agentvm","capability","protocol","implementation","ocap","security"],"s":["agentvm-capability","agentvm","capability","ocap","security"]},
+{"id":"agentvm-cli","family":"agentvm","kind":"crate","store":"cognitum-cogs","version":null,"description":"CLI tool for Agentic VM - Accountable Agent Runtime","keywords":["agentic","capability","capsule","cli","evidence","security","vm"],"source":"cognitum-cogs/crates/agentvm/crates/agentvm-cli/Cargo.toml","sourceSha256":"351899cc973855f56877e77be24d051eec361ed9a9ee5b1625e7196147181571","variants":["agentvm-types"],"readme":"Command-line interface for Agentic VM - Accountable Agent Runtime. The agentvm CLI provides commands for: - Running agents in capsules with evidence generation - Managing snapshots for recovery - Querying and verifying evidence - Managing capabilities - Running benchmarks Execute a command in an agent capsule: Manage capsule snapshots: Query and verify evidence: Manage capabilities: Run and analyze benchmarks: Restore capsule to a snapshot: Replay execution from evidence: Configuration can be provided via: - --config flag - AGENTVM_CONFIG environment variable - ~/.config/agentvm/config.toml Apache-2.0","readmeSource":"cognitum-cogs/crates/agentvm/crates/agentvm-cli/README.md","product":"key:cognitum-cogs:agentvm","canonical":"agentvm-cli","tier":null,"t":["agentvm-cli","agentvm","accountable","runtime","capability","capsule","evidence","security"],"s":["agentvm-cli","agentvm","capability","capsule","evidence","security"]},
+{"id":"agentvm-evidence","family":"agentvm-evidence","kind":"crate","store":"cognitum-cogs","version":null,"description":"Evidence chain and Merkle tree implementation for Agentic VM","keywords":["agentic","audit","evidence","merkle","vm"],"source":"cognitum-cogs/crates/agentvm-evidence/Cargo.toml","sourceSha256":"e3ea9dbd94594e52543609145d97fc5de87ff71c8f1e6d6342f4460e8550d911","variants":[],"product":"key:cognitum-cogs:agentvm-evidence","canonical":"agentvm-evidence","tier":null,"t":["agentvm-evidence","agentvm","evidence","chain","merkle","tree","implementation","audit"],"s":["agentvm-evidence","agentvm","evidence","audit","merkle"]},
+{"id":"agentvm-proxy","family":"agentvm-proxy","kind":"crate","store":"cognitum-cogs","version":null,"description":"Capability proxy layer for Agentic VM","keywords":["agentic","capability","proxy","security","vm"],"source":"cognitum-cogs/crates/agentvm-proxy/Cargo.toml","sourceSha256":"c648df924ef2e36b9885851405a8ea76b3d909fc3e272109cce35e3dc695d0d5","variants":[],"product":"key:cognitum-cogs:agentvm-proxy","canonical":"agentvm-proxy","tier":null,"t":["agentvm-proxy","agentvm","proxy","capability","layer","security"],"s":["agentvm-proxy","agentvm","proxy","capability","security"]},
+{"id":"agentvm-scheduler","family":"agentvm-scheduler","kind":"crate","store":"cognitum-cogs","version":null,"description":"Fabric scheduler for heterogeneous compute tiers in Agentic VM","keywords":["agentic","edge","heterogeneous","scheduler","vm"],"source":"cognitum-cogs/crates/agentvm-scheduler/Cargo.toml","sourceSha256":"3de21c9d5ff31f969d335bc74146f7287826d16bd25c9238c0fa9f13bfb104c0","variants":[],"product":"key:cognitum-cogs:agentvm-scheduler","canonical":"agentvm-scheduler","tier":null,"t":["agentvm-scheduler","agentvm","scheduler","fabric","heterogeneous","compute","tier","edge"],"s":["agentvm-scheduler","agentvm","scheduler","edge","heterogeneous"]},
+{"id":"agl-types","family":"agl","kind":"crate","store":"autogenous","version":null,"description":"Autogenous Genome Language (AGL) core types: genomes, typed mutations, authority classes, hard invariants, and vector fitness with hard-gate promotion. The typed mutation specification of the Autogenous Runtime.","keywords":["agl","autogenous","evolution","governance","mutation"],"source":"autogenous/crates/agl-types/Cargo.toml","sourceSha256":"ab0437634502937f6236a2589a792e7b58ea110ec8ee4f14db8e27423cb5c9cd","variants":[],"readme":"agl-types is the type foundation of [Autogenous](../), a governed, self-evolving AI-defense runtime in Rust. It defines the Autogenous Genome Language (AGL): the vocabulary in which a self-improving system describes a change to itself. The core idea: a mutation is not a text patch. It is a *typed transformation between two valid genomes* that declares what it changes, why it should work, where it applies, what authority it needs, which invariants it preserves, when it expires, and how to reverse it. Because that is all in the type system, unsafe changes are rejected structurally — not by a linter, a review, or a prompt. - Genome — a valid configuration of the system, pinned to a constitution","readmeSource":"autogenous/crates/agl-types/README.md","product":"key:autogenous:agl","canonical":"agl-types","tier":null,"t":["agl-types","agl","type","autogenous","genome","language","core","typed","mutation","authority","class","hard","invariant","vector","fitness","hard-gate","gate","promotion","specification","runtime","evolution","governance"],"s":["agl-types","agl","type","autogenous","evolution","governance","mutation"]},
+{"id":"ai-text-watermark","family":"ai-text-watermark","kind":"npm","store":"ai-text-watermark","version":"0.4.0","description":"Invisible, key-verifiable watermarks for AI/LLM text (SynthID-Text style) in Rust + WebAssembly. Generate, detect, embed a secret provenance message, and run an ultra-low-latency decode-loop proxy — Node and browser. Standalone wrapper around @claude-flow/watermark.","keywords":["ai-content-detection","ai-provenance","ai-text-watermark","content-credentials","distortion-free-watermark","eu-ai-act","gumbel-watermark","llm-watermark","synthid","synthid-text","text-watermark","wasm"],"source":"ai-text-watermark/packages/ai-text-watermark/package.json","sourceSha256":"d696ffa541188fc3572ecba50f518a53336449a47349bf5753002a07a21a2727","variants":[],"readme":"Invisible, key-verifiable watermarks for AI / LLM text — SynthID-Text style, in Rust + WebAssembly. Generate, detect, embed a secret provenance message, and run an ultra-low-latency decode-loop proxy — Node and browser. This is a thin standalone wrapper that re-exports [@claude-flow/watermark](https://www.npmjs.com/package/@claude-flow/watermark) verbatim, so you get a memorable install name without any code duplication. Browser / bundlers: import { init, Watermarker, detect } from 'ai-text-watermark/web'; await init(); Full docs, live playground, Rust crate (ruflo-watermark), and ADRs: https://github.com/ruvnet/ai-text-watermark · MIT © rUv","readmeSource":"ai-text-watermark/packages/ai-text-watermark/README.md","product":"repo:ai-text-watermark","canonical":"ai-text-watermark","tier":null,"t":["ai-text-watermark","text","watermark","invisible","key-verifiable","key","verifiable","synthid-text","synthid","style","generate","detect","embed","secret","provenance","message","ultra-low-latency","ultra","low","latency","decode-loop","decode","loop","proxy","browser","standalone","wrapper","around","claude-flow","claude","flow","ai-content-detection","content","detection","ai-provenance","content-credentials","credential","distortion-free-watermark","distortion","free","eu-ai-act","act","gumbel-watermark","gumbel","llm-watermark","text-watermark"],"s":["ai-text-watermark","text","watermark","ai-content-detection","content","detection","ai-provenance","provenance","content-credentials","credential","distortion-free-watermark","distortion","free","eu-ai-act","act","gumbel-watermark","gumbel","llm-watermark","synthid","synthid-text","text-watermark"]},
+{"id":"aidefence-core","family":"aidefence","kind":"crate","store":"midstream","version":null,"description":"Rust core for AIDefence: regex-pack detection of prompt injection, tool-invocation directives, exfiltration URLs, encoded instructions, Slack markup forgery and PII, sharing pattern data with the aidefence npm package","keywords":["aidefence","llm","pii","prompt-injection","security"],"source":"midstream/AIMDS/crates/aidefence-core/Cargo.toml","sourceSha256":"f5ee5c77927006411ee95ae96abd1146a87bbccafa5f0a20d3b8dec725b54957","variants":[],"readme":"Rust core for [AIDefence](https://www.npmjs.com/package/aidefence). Pack-driven, deterministic detection of prompt injection, tool-invocation directives, exfiltration URL shapes, encoded instructions and Slack markup forgery, plus PII detection and masking, so Rust services (the cognitum-one Slack bot, cognitum-cogs) run the same detection as the npm package. One pattern source. The packs are the JSON files in AIMDS/patterns/ (core, tool_invocation, exfil_url, encoded_instruction, slack_markup_forgery, instruction_override_i18n), shared with the TypeScript engine in AIMDS/src/detection. This crate embeds them at build time (include_str!) and can also load them at runtime (Registry::from_dir)","readmeSource":"midstream/AIMDS/crates/aidefence-core/README.md","product":"key:aidefence","canonical":"@claude-flow/aidefence","tier":"T1","t":["aidefence-core","aidefence","core","regex-pack","regex","pack","detection","prompt","injection","tool-invocation","invocation","directive","exfiltration","urls","encod","instruction","slack","markup","forgery","pii","shar","pattern","npm","prompt-injection","security"],"s":["aidefence-core","aidefence","core","pii","prompt-injection","prompt","injection","security"]},
+{"id":"aimds-analysis","family":"aimds-analysis","kind":"crate","store":"midstream","version":null,"description":"Deep behavioral analysis layer for AIMDS with temporal neural verification","keywords":[],"source":"midstream/AIMDS/crates/aimds-analysis/Cargo.toml","sourceSha256":"ab7761b067fcebe1aa3af3bdf9567a19947634e58be2a02408f8e35f5b2f7cee","variants":[],"readme":"Behavioral analysis and formal verification for AI threat detection - Temporal pattern analysis, LTL policy checking, and anomaly detection with sub-520ms latency. Part of the [AIMDS](https://ruv.io/aimds) (AI Manipulation Defense System) by [rUv](https://ruv.io) - Production-ready adversarial defense for AI systems. - 🧠 Behavioral Analysis: Temporal pattern analysis via attractor classification (<100ms) - 🔒 Formal Verification: LTL policy checking with theorem proving (<500ms) - 📊 Anomaly Detection: Statistical baseline learning with multi-dimensional analysis - ⚡ High Performance: <520ms combined deep-path latency (validated) - 🎯 Production Ready: 100% test coverage (27/27), zero unsaf","readmeSource":"midstream/AIMDS/crates/aimds-analysis/README.md","product":"key:midstream:aimds-analysis","canonical":"aimds-analysis","tier":"T1","t":["aimds-analysis","aimd","analysis","deep","behavioral","layer","temporal","neural","verification"],"s":["aimds-analysis","aimd","analysis"]},
+{"id":"aimds-core","family":"aimds","kind":"crate","store":"midstream","version":null,"description":"Core types and abstractions for AI Manipulation Defense System (AIMDS)","keywords":[],"source":"midstream/AIMDS/crates/aimds-core/Cargo.toml","sourceSha256":"7b634c0b98b962212c466926e5bc7b3b398620587fd9fdf16144f2d282f7467c","variants":[],"readme":"Core type system, configuration, and error handling for AIMDS - Production-ready adversarial defense for AI applications. Part of the [AIMDS](https://ruv.io/aimds) (AI Manipulation Defense System) by [rUv](https://ruv.io) - Real-time threat detection with formal verification. - 🎯 Type-Safe Design: Comprehensive type system for threats, policies, and responses - ⚙️ Flexible Configuration: Environment-based config with sensible defaults - 🛡️ Robust Error Handling: Hierarchical error types with severity levels and retryability - 📊 Zero Dependencies: Minimal dependency footprint for core types - 🚀 Production Ready: 100% test coverage, validated in production workloads - 🔧 Extensible: Easy t","readmeSource":"midstream/AIMDS/crates/aimds-core/README.md","product":"key:midstream:aimds","canonical":"aimds-core","tier":"T1","t":["aimds-core","aimd","core","type","abstraction","manipulation","defense"],"s":["aimds-core","aimd","core"]},
+{"id":"aimds-detection","family":"aimds-detection","kind":"crate","store":"midstream","version":null,"description":"Fast-path detection layer for AIMDS with pattern matching and anomaly detection","keywords":[],"source":"midstream/AIMDS/crates/aimds-detection/Cargo.toml","sourceSha256":"e20c433ec2fd0fb65687cf50c22db11e1d21cd3790bb90c1272f9eb1e9073704","variants":[],"readme":"Real-time threat detection with sub-10ms latency for AI applications - Prompt injection detection, PII sanitization, and pattern matching. Part of the [AIMDS](https://ruv.io/aimds) (AI Manipulation Defense System) by [rUv](https://ruv.io) - Production-ready adversarial defense for AI systems. - 🚀 Ultra-Low Latency: <10ms p99 detection latency (validated) - 🎯 Prompt Injection Detection: 50+ attack patterns with regex and Aho-Corasick - 🔒 PII Sanitization: Remove emails, SSNs, credit cards, API keys, phone numbers - ⚡ High Throughput: >10,000 requests/second on commodity hardware - 🧠 Pattern Caching: LRU cache for frequent patterns (>90% hit rate) - 📊 Production Ready: Comprehensive metri","readmeSource":"midstream/AIMDS/crates/aimds-detection/README.md","product":"key:midstream:aimds-detection","canonical":"aimds-detection","tier":"T1","t":["aimds-detection","aimd","detection","fast-path","path","layer","pattern","match","anomaly"],"s":["aimds-detection","aimd","detection"]},
+{"id":"aimds-gateway","family":"aimds-gateway","kind":"npm","store":"midstream","version":"1.0.0","description":"AIMDS TypeScript API Gateway with AgentDB and lean-agentic integration","keywords":["agentdb","aimds","api-gateway","defense","lean-agentic","security"],"source":"midstream/AIMDS/package.json","sourceSha256":"caf840ed59584c0e3e41c39be054ae1f3a0632dec927b95d8c8bd9d0c5a659c7","variants":[],"readme":"Production-ready adversarial defense system for AI applications with real-time threat detection, behavioral analysis, and formal verification. Part of the [Midstream Platform](https://github.com/agenticsorg/midstream) by [rUv](https://ruv.io) - Temporal analysis and AI security infrastructure. - ⚡ Real-Time Detection (<10ms): Pattern matching, prompt injection detection, PII sanitization - 🧠 Behavioral Analysis (<100ms): Temporal pattern analysis, anomaly detection, baseline learning - 🔒 Formal Verification (<500ms): LTL policy checking, dependent type verification, theorem proving - 🛡️ Adaptive Response (<50ms): Meta-learning mitigation, strategy optimization, rollback management - 📊 Pr","readmeSource":"midstream/AIMDS/README.md","product":"key:midstream:aimds-gateway","canonical":"aimds-gateway","tier":"T1","t":["aimds-gateway","aimd","gateway","agentdb","lean-agentic","lean","integration","api-gateway","defense","security"],"s":["aimds-gateway","aimd","gateway","agentdb","api-gateway","defense","lean-agentic","lean","security"]},
+{"id":"aimds-response","family":"aimds-response","kind":"crate","store":"midstream","version":null,"description":"Adaptive response layer with meta-learning for AIMDS threat mitigation","keywords":[],"source":"midstream/AIMDS/crates/aimds-response/Cargo.toml","sourceSha256":"39cf8ce40d593c196fdae58d712f988b4219447142b597951f1cf732bdc16014","variants":[],"readme":"Adaptive threat mitigation with meta-learning - 25-level recursive optimization, strategy selection, and rollback management with sub-50ms response time. Part of the [AIMDS](https://ruv.io/aimds) (AI Manipulation Defense System) by [rUv](https://ruv.io) - Production-ready adversarial defense for AI systems. - 🛡️ Adaptive Mitigation: 7 strategy types with effectiveness tracking (<50ms) - 🧠 Meta-Learning: 25-level recursive optimization via strange-loop - 📊 Effectiveness Tracking: Real-time success rate monitoring per strategy - ⏪ Rollback Management: Automatic undo for failed mitigations - 📝 Comprehensive Audit: Full audit trail with JSON export - 🚀 Production Ready: 97% test coverage (3","readmeSource":"midstream/AIMDS/crates/aimds-response/README.md","product":"key:midstream:aimds-response","canonical":"aimds-response","tier":"T1","t":["aimds-response","aimd","response","adaptive","layer","meta-learning","meta","learn","threat","mitigation"],"s":["aimds-response","aimd","response"]},
+{"id":"antibody","family":"antibody","kind":"crate","store":"autogenous","version":null,"description":"Autogenous Antibody Package (AAP): signed, capability-constrained, expiring adaptations with triggers, containment, evidence, and rollback (ADR-392 §8, Phase 4).","keywords":[],"source":"autogenous/crates/antibody/Cargo.toml","sourceSha256":"596c0d8c8e85739809801951752b60795a84874b2e42060c5b66aca7efe9541f","variants":[],"readme":"antibody defines the Autogenous Antibody Package (AAP) — the unit of defense in [Autogenous](../), a self-evolving AI-defense runtime in Rust. Borrowing from the immune system: an antibody is a *complete, self-describing, reversible* response to a threat, not just a rule. Each antibody carries its trigger, a serializable detector, signed evidence, a containment response, a regression corpus reference, a fitness envelope, its lineage, an expiry, and a rollback target — and it's ed25519-signed over the whole package, so tampering with any field breaks the signature. - Antibody — the package: trigger, detector, evidence, containment, requested authority, prohibited effects, expiry, rollback tar","readmeSource":"autogenous/crates/antibody/README.md","product":"key:autogenous:antibody","canonical":"antibody","tier":null,"t":["antibody","autogenous","aap","sign","capability-constrained","capability","constrain","expir","adaptation","trigger","containment","evidence","rollback","adr-392","adr","392","phase"],"s":["antibody"]},
+{"id":"ap-casper-connector","family":"ap-casper-connector","kind":"npm","store":"ap-casper","version":"0.0.1","description":"API for managing NFT rewards on the Casper Blockchain","keywords":[],"source":"ap-casper/package.json","sourceSha256":"b578adcb75ed36399b6d3bdd474a7881342c2634d7569011a239d2b679b8a4cb","variants":[],"readme":"This is a simple API for minting, transferring, and burning non-fungible tokens (NFTs) on the Casper blockchain using the CEP-78 Enhanced NFT Standard. It also includes endpoints for minting and transferring fungible tokens. To use this API, you will need: - A Casper account with funds for paying transaction fees - A private key for your Casper account - Node.js installed on your system - The casper-js-sdk, casper-cep78-js-client, express, and body-parser Node.js packages installed 1. Clone this repository to your local system. 2. Install the required dependencies by running the npm install command in the project directory. 3. Start the API server by running the node index.js command in the","readmeSource":"ap-casper/README.md","product":"repo:ap-casper","canonical":"ap-casper-connector","tier":"T3","t":["ap-casper-connector","casper","connector","manag","nft","reward","blockchain"],"s":["ap-casper-connector","casper","connector"]},
+{"id":"arcadia","family":"arcadia","kind":"crate","store":"arcadia","version":null,"description":"AI-driven game engine with VIVIAN and PARIS frameworks for dynamic, adaptive gaming experiences with AgentDB learning","keywords":["adaptive","ai","game-engine","procedural","vector-index"],"source":"arcadia/Cargo.toml","sourceSha256":"df61a586f517dc232c40a46214d5f7bb05f2c473297e925048b6e5692a3c013b","variants":[],"readme":"Advanced and Responsive Computational Architecture for Dynamic Interactive AI ARCADIA represents a paradigm shift in game engine design—where artificial intelligence isn't just a feature, it's the foundation. Built from the ground up with Rust's performance and safety guarantees, ARCADIA combines cutting-edge AI systems with battle-tested game development tools to create experiences that truly understand and adapt to players. What makes ARCADIA revolutionary: - Living, Breathing Worlds: NPCs with genuine emotional intelligence and memory that spans sessions - Cognitive AI Systems: From reactive instincts to abstract strategic planning across 4 cognitive levels - Self-Evolving Gameplay: Autop","readmeSource":"arcadia/README.md","product":"repo:arcadia","canonical":"arcadia-agentdb","tier":"T3","t":["arcadia","ai-driven","driven","game","vivian","paris","dynamic","adaptive","gaming","experience","agentdb","learn","game-engine","procedural","vector-index","vector","index"],"s":["arcadia","adaptive","game-engine","game","procedural","vector-index","vector","index"]},
+{"id":"arcadia-agentdb","family":"arcadia-agentdb","kind":"npm","store":"arcadia","version":"0.1.0","description":"ARCADIA AgentDB WASM bindings for browser and Node.js","keywords":["agentdb","ai","experience-replay","machine-learning","reinforcement-learning","vector-database","wasm"],"source":"arcadia/package.json","sourceSha256":"7311479e879c5a6e7b28dc78dc77fad423c7db16c9d332604a280fabf4079499","variants":[],"readme":"Advanced and Responsive Computational Architecture for Dynamic Interactive AI ARCADIA represents a paradigm shift in game engine design—where artificial intelligence isn't just a feature, it's the foundation. Built from the ground up with Rust's performance and safety guarantees, ARCADIA combines cutting-edge AI systems with battle-tested game development tools to create experiences that truly understand and adapt to players. What makes ARCADIA revolutionary: - Living, Breathing Worlds: NPCs with genuine emotional intelligence and memory that spans sessions - Cognitive AI Systems: From reactive instincts to abstract strategic planning across 4 cognitive levels - Self-Evolving Gameplay: Autop","readmeSource":"arcadia/README.md","product":"repo:arcadia","canonical":"arcadia-agentdb","tier":"T3","t":["arcadia-agentdb","arcadia","agentdb","browser","experience-replay","experience","replay","machine-learning","machine","learn","reinforcement-learning","reinforcement","vector-database","vector","database"],"s":["arcadia-agentdb","arcadia","agentdb","experience-replay","experience","replay","machine-learning","machine","learn","reinforcement-learning","reinforcement","vector-database","vector","database"]},
+{"id":"auto-browser-agent","family":"auto-browser-agent","kind":"npm","store":"auto-browser","version":"0.1.0","description":"Maintain and validate auto-browser","keywords":[],"source":"auto-browser/.harness/generated/package.json","sourceSha256":"f7eff66f13bc90e61661e3dfd2794663eb34baa796a6cbc74d99c5ea709e1d1e","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"auto-browser/.harness/generated/README.md","product":"key:auto-browser:auto-browser-agent","canonical":"auto-browser-agent","tier":"T3","t":["auto-browser-agent","auto","browser","maintain","validate","auto-browser"],"s":["auto-browser-agent","auto","browser"]},
+{"id":"autogenous-generator","family":"autogenous-generator","kind":"crate","store":"autogenous","version":null,"description":"Autogenous candidate generator: synthesizes a diverse population of typed antibody candidates from witnessed attack evidence, WITHOUT access to the evaluator's labels (ADR-392 §11 separation, ADR-393 MVP #3, gap 8).","keywords":[],"source":"autogenous/crates/generator/Cargo.toml","sourceSha256":"cfe63fe6a24b1b13cbe35d9d8e40aadc2d396963f1544ae090dc520c938b0d9c","variants":[],"readme":"autogenous-generator is the \"generate adaptations\" stage of [Autogenous](../), a self-evolving AI-defense runtime in Rust. Given a witnessed attack sample, it synthesizes a *population* of candidate [antibody](../antibody) defenses spanning a precision/recall spectrum — then lets the evaluator and verifier decide which survive. Its single most important property is structural evaluator separation: the generator's only input is the attack evidence. It never sees the labeled corpus the evaluator judges against — so it cannot overfit or game the hidden test set. That's the strongest built-in defense against reward hacking. - propose(evidence, parent, authority, now, cfg) — returns a diverse Vec","readmeSource":"autogenous/crates/generator/README.md","product":"key:autogenous:autogenous-generator","canonical":"autogenous-generator","tier":null,"t":["autogenous-generator","autogenous","generator","candidate","synthesiz","diverse","population","typed","antibody","witnes","attack","evidence","access","evaluator","label","adr-392","adr","392","separation","adr-393","393","mvp","gap"],"s":["autogenous-generator","autogenous","generator"]},
+{"id":"autogenous-service","family":"autogenous-service","kind":"crate","store":"autogenous","version":null,"description":"HTTP service exposing the Autogenous AGL admission contract (ADR-392 §6) and fitness hard-gate (§10) as an OpenAI-family-style /v1 API for governed agent mutations.","keywords":[],"source":"autogenous/crates/service/Cargo.toml","sourceSha256":"d1bf3c3beb72feeaac54411b79749a389645083ec672a282ab0bf7ecd56478b7","variants":[],"product":"key:autogenous:autogenous-service","canonical":"autogenous-service","tier":null,"t":["autogenous-service","autogenous","service","http","expos","agl","admission","contract","adr-392","adr","392","fitness","hard-gate","hard","gate","openai-family-style","openai","family","style","govern","mutation"],"s":["autogenous-service","autogenous","service"]},
+{"id":"bit-parallel-search","family":"bit-parallel-search","kind":"crate","store":"sublinear-time-solver","version":null,"description":"Fast bit-parallel string matching using 64-bit parallelism","keywords":["bit-parallel","fast","pattern-matching","search","string"],"source":"sublinear-time-solver/crates/bit-parallel-search/bit-parallel-search/Cargo.toml","sourceSha256":"cd1aceae0a6735e2118f945b7cfa26649eef0916f0fcd848ec2e21b49d28d0f9","variants":[],"readme":"Fast bit-parallel string matching that processes 64 positions simultaneously. Processes up to 64 pattern positions in parallel using bit manipulation: Uses bit masks to track pattern matches across 64 positions simultaneously: 1. Each bit represents a potential match position 2. Shift and mask operations update all positions in parallel 3. No branching in inner loop (CPU-friendly) - Pattern must be ≤ 64 bytes - Best performance for patterns < 32 bytes - Not suitable for very long patterns - Actually faster than standard approaches (verified) - No unsafe code in main algorithm - No dependencies (no_std compatible) - Cache-friendly - processes data sequentially MIT OR Apache-2.0","readmeSource":"sublinear-time-solver/crates/bit-parallel-search/bit-parallel-search/README.md","product":"key:sublinear-time-solver:bit-parallel-search","canonical":"bit-parallel-search","tier":"T1","t":["bit-parallel-search","bit","parallel","search","bit-parallel","string","match","64-bit","parallelism","pattern-matching","pattern"],"s":["bit-parallel-search","bit","parallel","search","bit-parallel","pattern-matching","pattern","match","string"]},
+{"id":"bot-generator-bot-agent","family":"bot-generator-bot-agent","kind":"npm","store":"bot-generator-bot","version":"0.1.0","description":"Maintain and validate Bot-Generator-Bot","keywords":[],"source":"bot-generator-bot/.harness/generated/package.json","sourceSha256":"cd6c30b1ebbdc901de2f90c15fc8c8e0ec2819611114db814120fce9c773ee12","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"bot-generator-bot/.harness/generated/README.md","product":"key:bot-generator-bot:bot-generator-bot-agent","canonical":"bot-generator-bot-agent","tier":"T3","t":["bot-generator-bot-agent","bot","generator","maintain","validate","bot-generator-bot"],"s":["bot-generator-bot-agent","bot","generator"]},
+{"id":"claude_market","family":"claude_market","kind":"crate","store":"synaptic-mesh","version":null,"description":"P2P marketplace for trading Claude API tokens with escrow and reputation","keywords":["escrow","libp2p","marketplace","p2p","tokens"],"source":"synaptic-mesh/standalone-crates/synaptic-mesh-cli/crates/claude_market/Cargo.toml","sourceSha256":"575315350e7847538748fba9332ac1a162d822a712e91682f58c5db2a2c5c3c6","variants":[],"readme":"A decentralized, peer-to-peer marketplace for Claude AI capacity trading with full Anthropic ToS compliance. Claude Market enables secure, decentralized trading of Claude AI compute capacity using ruv tokens. Built on the Synaptic Neural Mesh network, it provides a compliant way for Claude Max subscribers to share their capacity through a peer compute federation model. - 🔒 Anthropic ToS Compliant - No API key sharing, peer-orchestrated execution - 🏦 Secure Escrow - Multi-signature escrow with automatic settlement - 🎯 First-Accept Auctions - Fast, competitive pricing mechanisms - 📊 Reputation System - SLA tracking and provider trust scores - 🛡️ Privacy-Preserving - Encrypted task payload","readmeSource":"synaptic-mesh/standalone-crates/synaptic-mesh-cli/crates/claude_market/README.md","product":"key:synaptic-mesh:claude_market","canonical":"claude_market","tier":"T1","t":["claude","market","p2p","marketplace","trad","token","escrow","reputation","libp2p"],"s":["claude","market","escrow","libp2p","marketplace","p2p","token"]},
+{"id":"claude-flow","family":"claude-flow","kind":"npm","store":"ruflo","version":"3.49.0","description":"Ruflo - Enterprise AI agent orchestration for Claude Code. Deploy 60+ specialized agents in coordinated swarms with self-learning, fault-tolerant consensus, vector memory, and MCP integration","keywords":["agent-orchestration","ai","ai-agents","anthropic","automation","chatgpt","claude","claude-code","cli","code-generation","coding-assistant","consensus","developer-tools","devtools","distributed-systems","embeddings","enterprise","gpt","hive-mind","large-language-model","llm","machine-learning","mcp","model-context-protocol","multi-agent","neural-network","orchestration","ruvflow","self-learning","swarm","swarm-intelligence","vector-database","workflow"],"source":"ruflo/package.json","sourceSha256":"8897422f9d6409401e5787c6cb5efd0e422e81ca00a8c17346e371e15443e982","variants":[],"readme":"[English](README.md) · [简体中文](README.zh-CN.md) An agent meta-harness for Claude Code and Codex. [📖 RuFlo Explained — Build an AI Team That Plans, Remembers, Tests, and Improves](docs/ruflo-explained.md) A 14-chapter guide: from the basic idea to a first useful task, then memory, agent teams, plugins, cost and verification.","readmeSource":"ruflo/README.md","product":"repo:ruflo","canonical":"ruflo","tier":"T0","t":["claude-flow","claude","flow","ruflo","enterprise","orchestration","code","deploy","specializ","coordinat","swarm","self-learning","self","learn","fault-tolerant","fault","tolerant","consensus","vector","memory","mcp","integration","agent-orchestration","ai-agents","anthropic","automation","chatgpt","claude-code","code-generation","generation","coding-assistant","coding","assistant","developer-tools","developer","devtool","distributed-systems","distribut","embedding","gpt","hive-mind","hive","mind","large-language-model","large","language","model","machine-learning","machine","model-context-protocol","context","protocol","multi-agent","multi","neural-network","neural","network","ruvflow","swarm-intelligence","intelligence","vector-database","database","workflow"],"s":["claude-flow","claude","flow","agent-orchestration","orchestration","ai-agents","anthropic","automation","chatgpt","claude-code","code","code-generation","generation","coding-assistant","coding","assistant","consensus","developer-tools","developer","devtool","distributed-systems","distribut","embedding","enterprise","gpt","hive-mind","hive","mind","large-language-model","large","language","model","machine-learning","machine","learn","mcp","model-context-protocol","context","protocol","multi-agent","multi","neural-network","neural","network","ruvflow","self-learning","self","swarm","swarm-intelligence","intelligence","vector-database","vector","database","workflow"]},
+{"id":"claude-parser","family":"claude-parser","kind":"crate","store":"ruv-fann","version":null,"description":"Claude Code CLI stream-JSON parser for ruv-swarm multi-agent orchestration","keywords":["ai","claude","json","parser","stream"],"source":"ruv-fann/ruv-swarm/crates/claude-parser/Cargo.toml","sourceSha256":"5d26dd6883da9e6b8f7e595bc041eaa251dc1c3b022fa76036801ebafcd44991","variants":[],"readme":"A high-performance stream-JSON parser for Claude Code CLI output, designed for benchmarking and training data collection. - Stream Event Parsing: Supports all Claude event types: - message_start, message_stop - content_block_start, content_block_delta, content_block_stop - tool_use, function_result - thinking - error - usage - Performance Metrics Extraction: - Timing measurements (total duration, time to first output) - Token usage tracking (input, output, total) - Tool invocation statistics - Thinking sequence analysis - Error and recovery patterns - Error Handling: Robust error recovery with detailed tracking of error types and recovery strategies - Export Functionality: Export parsed data","readmeSource":"ruv-fann/ruv-swarm/crates/claude-parser/README.md","product":"key:ruv-fann:claude-parser","canonical":"claude-parser","tier":"T1","t":["claude-parser","claude","parser","code","stream-json","stream","json","ruv-swarm","swarm","multi-agent","multi","orchestration"],"s":["claude-parser","claude","parser","json","stream"]},
+{"id":"code-mesh","family":"code-mesh","kind":"npm","store":"code-mesh","version":"0.1.0","description":"AI-powered coding assistant - modular Rust + WASM system","keywords":["ai","anthropic","browser","cli","coding-assistant","distributed","javascript","llm","neural","openai","performance","rust","swarm","typescript","wasm","webassembly"],"source":"code-mesh/npm/package.json","sourceSha256":"4c6022bb5a9686eb704a691701be744814e260a4fcba209febae5f5ee2256694","variants":["code-mesh-cli","code-mesh-core","code-mesh-wasm"],"readme":"A universal TypeScript/JavaScript library for Code Mesh that works seamlessly across Node.js, browsers, and web workers with WebAssembly performance. - 🦀 WebAssembly Performance: Rust-powered core with WASM bindings - 🌐 Universal Compatibility: Works in Node.js, browsers, and web workers - 💾 Browser Storage: IndexedDB integration for persistent sessions - 🔄 Offline Support: Service Worker and PWA capabilities - 🧵 Web Workers: Background processing support - 🔐 Authentication: Multiple provider support (API keys, OAuth2) - ⚡ Performance Monitoring: Built-in metrics and optimization - 📱 Progressive Web App: Full PWA support with manifest - 🎯 TypeScript: Complete type safety and IntelliS","readmeSource":"code-mesh/npm/README.md","product":"repo:code-mesh","canonical":"code-mesh","tier":"T3","t":["code-mesh","code","mesh","ai-powered","power","coding","assistant","modular","anthropic","browser","coding-assistant","distribut","neural","openai","swarm"],"s":["code-mesh","code","mesh","anthropic","browser","coding-assistant","coding","assistant","distribut","neural","openai","swarm"]},
+{"id":"code-mesh-tui","family":"code-mesh-tui","kind":"crate","store":"code-mesh","version":null,"description":"Interactive Terminal User Interface for the Code-Mesh distributed swarm intelligence system","keywords":["monitoring","swarm","terminal","tui","visualization"],"source":"code-mesh/crates/code-mesh-tui/Cargo.toml","sourceSha256":"c8d11d9c04e1bd72b06c6523df30846163c32261fe72686f2797b0817ca25b33","variants":[],"readme":"Interactive Terminal User Interface for the Code-Mesh distributed swarm intelligence system. Code-Mesh TUI provides a beautiful, feature-rich terminal interface to visualize, monitor, and control your Code-Mesh swarms in real-time. Experience the power of distributed computing through an intuitive, responsive interface. - Modern Design: Clean, intuitive interface with customizable themes - Real-time Updates: Live monitoring with smooth animations - Responsive Layout: Adapts to any terminal size - Syntax Highlighting: Code display with language-aware highlighting - Swarm Topology Visualization: Interactive network graphs of agent relationships - Performance Dashboards: Real-time metrics with","readmeSource":"code-mesh/crates/code-mesh-tui/README.md","product":"key:code-mesh:code-mesh-tui","canonical":"code-mesh-tui","tier":"T3","t":["code-mesh-tui","code","mesh","tui","interactive","terminal","user","interface","code-mesh","distribut","swarm","intelligence","monitor","visualization"],"s":["code-mesh-tui","code","mesh","tui","monitor","swarm","terminal","visualization"]},
+{"id":"cog-adversarial","family":"cog-adversarial","kind":"crate","store":"cognitum-cogs","version":null,"description":"Adversarial signal detection — statistical tests for spoofed signals","keywords":[],"source":"cognitum-cogs/src/cogs/adversarial/Cargo.toml","sourceSha256":"b5ce91ea6b57f531397a68b9befb68120da6a3a98616bd5e3d4c97785350924f","variants":[],"product":"key:cognitum-cogs:cog-adversarial","canonical":"cog-adversarial","tier":null,"t":["cog-adversarial","cog","adversarial","signal","detection","statistical","test","spoof"],"s":["cog-adversarial","cog","adversarial"]},
+{"id":"cog-air-quality-index","family":"cog-air-quality-index","kind":"crate","store":"cognitum-cogs","version":null,"description":"Air quality index — AQI from PM2.5/CO2 proxy using EPA breakpoints","keywords":[],"source":"cognitum-cogs/src/cogs/air-quality-index/Cargo.toml","sourceSha256":"4aa1b1ee8c89d34da5e8c37393ff36352a5d3a8fad026c219c1f8e08d0696a9f","variants":[],"product":"key:cognitum-cogs:cog-air-quality-index","canonical":"cog-air-quality-index","tier":null,"t":["cog-air-quality-index","cog","air","quality","index","aqi","pm2.5","co2","proxy","epa","breakpoint"],"s":["cog-air-quality-index","cog","air","quality","index"]},
+{"id":"cog-anomaly-attractor","family":"cog-anomaly-attractor","kind":"crate","store":"cognitum-cogs","version":null,"description":"Strange attractor analysis — delay embedding, Lyapunov exponent, chaos detection","keywords":[],"source":"cognitum-cogs/src/cogs/anomaly-attractor/Cargo.toml","sourceSha256":"b7a269c8e2ffb3dd43508793f56d5e0c659d3b02592f83c2525b816bd5268beb","variants":[],"product":"key:cognitum-cogs:cog-anomaly-attractor","canonical":"cog-anomaly-attractor","tier":null,"t":["cog-anomaly-attractor","cog","anomaly","attractor","strange","analysis","delay","embedd","lyapunov","exponent","chao","detection"],"s":["cog-anomaly-attractor","cog","anomaly","attractor"]},
+{"id":"cog-anomaly-detect","family":"cog-anomaly-detect","kind":"crate","store":"cognitum-cogs","version":null,"description":"Anomaly detection — Welford online stats + CUSUM drift detection","keywords":[],"source":"cognitum-cogs/src/cogs/anomaly-detect/Cargo.toml","sourceSha256":"7a8bb3ba8c1968968b5818f5cfce414a4019d00003eee1cb9fd26bac1f322cf7","variants":[],"product":"key:cognitum-cogs:cog-anomaly-detect","canonical":"cog-anomaly-detect","tier":null,"t":["cog-anomaly-detect","cog","anomaly","detect","detection","welford","online","stat","cusum","drift"],"s":["cog-anomaly-detect","cog","anomaly","detect"]},
+{"id":"cog-audit-logger","family":"cog-audit-logger","kind":"crate","store":"cognitum-cogs","version":null,"description":"Tamper-proof event logging with hash-chain and witness attestation","keywords":[],"source":"cognitum-cogs/src/cogs/audit-logger/Cargo.toml","sourceSha256":"edaf759391f6b3a541b6dccf5f96632fb45ed7ab60df12250889bb592f3a59a2","variants":[],"product":"key:cognitum-cogs:cog-audit-logger","canonical":"cog-audit-logger","tier":null,"t":["cog-audit-logger","cog","audit","logger","tamper-proof","tamper","proof","event","logg","hash-chain","hash","chain","witness","attestation"],"s":["cog-audit-logger","cog","audit","logger"]},
+{"id":"cog-baby-cry","family":"cog-baby-cry","kind":"crate","store":"cognitum-cogs","version":null,"description":"Baby cry detection — sustained mid-band energy detector for nursery monitoring","keywords":[],"source":"cognitum-cogs/src/cogs/baby-cry/Cargo.toml","sourceSha256":"f9e2cfdba0770ca4e90e8703885d4deb70d1f101b71746af1acfa4cee9ec0a58","variants":[],"product":"key:cognitum-cogs:cog-baby-cry","canonical":"cog-baby-cry","tier":null,"t":["cog-baby-cry","cog","baby","cry","detection","sustain","mid-band","mid","band","energy","detector","nursery","monitor"],"s":["cog-baby-cry","cog","baby","cry"]},
+{"id":"cog-beehive-monitor","family":"cog-beehive-monitor","kind":"crate","store":"cognitum-cogs","version":null,"description":"Beehive acoustic state monitor — queenless / swarming / robbing classification","keywords":[],"source":"cognitum-cogs/src/cogs/beehive-monitor/Cargo.toml","sourceSha256":"c7b70e3e360ed28508a7a79902ab3743e487bab3f1c30b98b5d924d6aed11ced","variants":[],"product":"key:cognitum-cogs:cog-beehive-monitor","canonical":"cog-beehive-monitor","tier":null,"t":["cog-beehive-monitor","cog","beehive","monitor","acoustic","state","queenless","swarm","robb","classification"],"s":["cog-beehive-monitor","cog","beehive","monitor"]},
+{"id":"cog-behavioral-profiler","family":"cog-behavioral-profiler","kind":"crate","store":"cognitum-cogs","version":null,"description":"Behavioral profiling — online k-means codebook of normal signal patterns with anomaly flagging","keywords":[],"source":"cognitum-cogs/src/cogs/behavioral-profiler/Cargo.toml","sourceSha256":"e4e29e5dde29d2c27f38f124964c774eaa2ccb91c6a36a7797f2323fff8862d8","variants":[],"product":"key:cognitum-cogs:cog-behavioral-profiler","canonical":"cog-behavioral-profiler","tier":null,"t":["cog-behavioral-profiler","cog","behavioral","profiler","profil","online","k-means","mean","codebook","normal","signal","pattern","anomaly","flagg"],"s":["cog-behavioral-profiler","cog","behavioral","profiler"]},
+{"id":"cog-breathing-sync","family":"cog-breathing-sync","kind":"crate","store":"cognitum-cogs","version":null,"description":"Breathing synchrony — cross-correlation and phase coherence between sensor channels","keywords":[],"source":"cognitum-cogs/src/cogs/breathing-sync/Cargo.toml","sourceSha256":"2afcffedca3c8261d9bd7f9a786698badef8eada99452031075601ebc4e5a8f9","variants":[],"product":"key:cognitum-cogs:cog-breathing-sync","canonical":"cog-breathing-sync","tier":null,"t":["cog-breathing-sync","cog","breath","sync","synchrony","cross-correlation","cross","correlation","phase","coherence","between","sensor","channel"],"s":["cog-breathing-sync","cog","breath","sync"]},
+{"id":"cog-cardiac-arrhythmia","family":"cog-cardiac-arrhythmia","kind":"crate","store":"cognitum-cogs","version":null,"description":"Cardiac arrhythmia detection — HRV analysis using RMSSD and pNN50","keywords":[],"source":"cognitum-cogs/src/cogs/cardiac-arrhythmia/Cargo.toml","sourceSha256":"24c34301fa5ae4da74ed525852d84368847b31fae6c6393855c4b4725f76777a","variants":[],"product":"key:cognitum-cogs:cog-cardiac-arrhythmia","canonical":"cog-cardiac-arrhythmia","tier":null,"t":["cog-cardiac-arrhythmia","cog","cardiac","arrhythmia","detection","hrv","analysis","rmssd","pnn50"],"s":["cog-cardiac-arrhythmia","cog","cardiac","arrhythmia"]},
+{"id":"cog-clean-room","family":"cog-clean-room","kind":"crate","store":"cognitum-cogs","version":null,"description":"Enforce maximum headcount in clean rooms via simultaneous presence counting","keywords":[],"source":"cognitum-cogs/src/cogs/clean-room/Cargo.toml","sourceSha256":"9aab903cc00ffa9143ec480a9c8df06df8478095e1df98df4a75788c04f576c4","variants":[],"product":"key:cognitum-cogs:cog-clean-room","canonical":"cog-clean-room","tier":null,"t":["cog-clean-room","cog","clean","room","enforce","maximum","headcount","via","simultaneous","presence","count"],"s":["cog-clean-room","cog","clean","room"]},
+{"id":"cog-cloud-inference","family":"cog-cloud-inference","kind":"crate","store":"cognitum-cogs","version":null,"description":"Cloud inference provider cog — forwards OpenAI-compat /v1/chat/completions to the metered meta-llm gateway (ADR-090 §e, ADR-095)","keywords":[],"source":"cognitum-cogs/src/cogs/cloud-inference/Cargo.toml","sourceSha256":"fc1563c44a49485d6999cc9b37f5e311f3c3b0a69d5fbeb5080bfe1396edc44f","variants":[],"readme":"Standalone provider cog (ADR-090 §e, ADR-095) that gives a seed a Tier-3 cloud-inference path. It exposes the OpenAI-compatible /v1/chat/completions contract on its loopback socket and forwards each request to the metered meta-llm gateway with its own cog_ key, so the seed's neural_router \"Powerful\" arm and the agent_runtime tool-call loop can dispatch to it exactly like any other API cog. This is the standalone deployment (ADR-090 §3): the cog talks directly to the cloud gateway with its own scoped key. The same binary supports hub-mediated mode by pointing --inference-base-url at a paired v0 hub instead. Behaviour mirrors the v0 hub's Tier-3 handler so both egress points are identical: - m","readmeSource":"cognitum-cogs/src/cogs/cloud-inference/README.md","product":"key:cognitum-cogs:cog-cloud-inference","canonical":"cog-cloud-inference","tier":null,"t":["cog-cloud-inference","cog","cloud","inference","provider","forward","openai-compat","openai","compat","chat","completion","meter","meta-llm","meta","gateway","adr-090","adr","090","adr-095","095"],"s":["cog-cloud-inference","cog","cloud","inference"]},
+{"id":"cog-cognitive-pipeline","family":"cog-cognitive-pipeline","kind":"crate","store":"cognitum-cogs","version":null,"description":"Cognitive pipeline cog — FastGRNN anomaly gate + SmolLM2 sparse-LLM inference (ADR-094, ADR-095)","keywords":[],"source":"cognitum-cogs/src/cogs/cognitive-pipeline/Cargo.toml","sourceSha256":"a329eaaea526e45120571e88b034968ecac986475238fbded4ad1c2c0d0bd8dc","variants":[],"readme":"FastGRNN anomaly gate + SmolLM2 sparse-LLM inference, packaged as a Cognitum Cog. Status: functional. Full sparse-LLM modules lifted from [seed#133](https://github.com/cognitum-one/seed/pull/133); real on-device inference verified end-to-end through the agent's /api/v1/cogs/cognitive-pipeline/* proxy. The cog ships its current API surface (HTTP catalog + streaming PUT for GGUF uploads + cognitive-event ring). Items still pending fleet rollout are called out in [Status of ADR-095 surfaces](#status-of-adr-095-surfaces) below. 1. Consumes sensor input (IMU windows by default, configurable). 2. Runs a FastGRNN gate to score each window for anomaly. 3. When the score crosses gate_threshold, fires","readmeSource":"cognitum-cogs/src/cogs/cognitive-pipeline/README.md","product":"key:cognitum-cogs:cog-cognitive-pipeline","canonical":"cog-cognitive-pipeline","tier":null,"t":["cog-cognitive-pipeline","cog","cognitive","pipeline","fastgrnn","anomaly","gate","smollm2","sparse-llm","sparse","inference","adr-094","adr","094","adr-095","095"],"s":["cog-cognitive-pipeline","cog","cognitive","pipeline"]},
+{"id":"cog-coherence","family":"cog-coherence","kind":"crate","store":"cognitum-cogs","version":null,"description":"Signal quality monitor — per-channel coherence metrics and noise floor","keywords":[],"source":"cognitum-cogs/src/cogs/coherence/Cargo.toml","sourceSha256":"5d545a27157a2c2ce31566516a79d3cae387745a32a4b487c711e6762c8fb1e7","variants":[],"product":"key:cognitum-cogs:cog-coherence","canonical":"cog-coherence","tier":null,"t":["cog-coherence","cog","coherence","signal","quality","monitor","per-channel","per","channel","metric","noise","floor"],"s":["cog-coherence","cog","coherence"]},
+{"id":"cog-coherence-gate","family":"cog-coherence-gate","kind":"crate","store":"cognitum-cogs","version":null,"description":"Coherence gate — filter signals by channel-pair coherence, estimate SNR","keywords":[],"source":"cognitum-cogs/src/cogs/coherence-gate/Cargo.toml","sourceSha256":"b0e507cbadd1c5b591af542144b3406397b50eb90ebf39c49c14d7a01d659dd7","variants":[],"product":"key:cognitum-cogs:cog-coherence-gate","canonical":"cog-coherence-gate","tier":null,"t":["cog-coherence-gate","cog","coherence","gate","filter","signal","channel-pair","channel","pair","estimate","snr"],"s":["cog-coherence-gate","cog","coherence","gate"]},
+{"id":"cog-confined-space","family":"cog-confined-space","kind":"crate","store":"cognitum-cogs","version":null,"description":"Monitor workers in confined spaces — presence duration, breathing, inactivity alerts","keywords":[],"source":"cognitum-cogs/src/cogs/confined-space/Cargo.toml","sourceSha256":"fc7a61c99b69af7358cfadad3dff043086f43a63803e3f2cfcc961184d34432e","variants":[],"product":"key:cognitum-cogs:cog-confined-space","canonical":"cog-confined-space","tier":null,"t":["cog-confined-space","cog","confin","space","monitor","worker","presence","duration","breath","inactivity","alert"],"s":["cog-confined-space","cog","confin","space"]},
+{"id":"cog-cough-detect","family":"cog-cough-detect","kind":"crate","store":"cognitum-cogs","version":null,"description":"Cough detection — transient + spectral + cluster acoustic detector","keywords":[],"source":"cognitum-cogs/src/cogs/cough-detect/Cargo.toml","sourceSha256":"ecd34e96dd1bf913636d49b9514c3afc41ba263580e78297ed53fd4025542cb7","variants":[],"product":"key:cognitum-cogs:cog-cough-detect","canonical":"cog-cough-detect","tier":null,"t":["cog-cough-detect","cog","cough","detect","detection","transient","spectral","cluster","acoustic","detector"],"s":["cog-cough-detect","cog","cough","detect"]},
+{"id":"cog-customer-flow","family":"cog-customer-flow","kind":"crate","store":"cognitum-cogs","version":null,"description":"Count entries/exits using directional detection, report hourly traffic","keywords":[],"source":"cognitum-cogs/src/cogs/customer-flow/Cargo.toml","sourceSha256":"c0c3dea94588d952bab5eb7d43980128f46ce722e624a35e45dccd782f2b67de","variants":[],"product":"key:cognitum-cogs:cog-customer-flow","canonical":"cog-customer-flow","tier":null,"t":["cog-customer-flow","cog","customer","flow","count","entry","exit","directional","detection","report","hourly","traffic"],"s":["cog-customer-flow","cog","customer","flow"]},
+{"id":"cog-dream-stage","family":"cog-dream-stage","kind":"crate","store":"cognitum-cogs","version":null,"description":"Sleep stage classification — deep, REM, light, wake from movement/breathing","keywords":[],"source":"cognitum-cogs/src/cogs/dream-stage/Cargo.toml","sourceSha256":"d73195cf318589a162c0457c0c4c29d27abb29aafc4445a8f6988aea8dc26433","variants":[],"product":"key:cognitum-cogs:cog-dream-stage","canonical":"cog-dream-stage","tier":null,"t":["cog-dream-stage","cog","dream","stage","sleep","classification","deep","rem","light","wake","movement","breath"],"s":["cog-dream-stage","cog","dream","stage"]},
+{"id":"cog-dtw-gesture-learn","family":"cog-dtw-gesture-learn","kind":"crate","store":"cognitum-cogs","version":null,"description":"Dynamic Time Warping gesture recognition — record templates and match new signals","keywords":[],"source":"cognitum-cogs/src/cogs/dtw-gesture-learn/Cargo.toml","sourceSha256":"62d05112e2b4ddf1ca40813c198fb52a430a0359eeb490167dced79f32f7beda","variants":[],"product":"key:cognitum-cogs:cog-dtw-gesture-learn","canonical":"cog-dtw-gesture-learn","tier":null,"t":["cog-dtw-gesture-learn","cog","dtw","gesture","learn","dynamic","warp","recognition","record","template","match","signal"],"s":["cog-dtw-gesture-learn","cog","dtw","gesture","learn"]},
+{"id":"cog-dwell-heatmap","family":"cog-dwell-heatmap","kind":"crate","store":"cognitum-cogs","version":null,"description":"Track dwell time per zone with rolling 1-hour window heatmap","keywords":[],"source":"cognitum-cogs/src/cogs/dwell-heatmap/Cargo.toml","sourceSha256":"ec76233d23924e664d1e729e7d166582146219f74f6b8ab66a43ddd48727835b","variants":[],"product":"key:cognitum-cogs:cog-dwell-heatmap","canonical":"cog-dwell-heatmap","tier":null,"t":["cog-dwell-heatmap","cog","dwell","heatmap","track","per","zone","roll","1-hour","hour","window"],"s":["cog-dwell-heatmap","cog","dwell","heatmap"]},
+{"id":"cog-elevator-count","family":"cog-elevator-count","kind":"crate","store":"cognitum-cogs","version":null,"description":"Count people by signal disruption intensity with linear regression calibration","keywords":[],"source":"cognitum-cogs/src/cogs/elevator-count/Cargo.toml","sourceSha256":"8e8164f0bc5db6121f064a92bdb3d19dbf02ab59ab5a914cf6318cb631c31fcd","variants":[],"product":"key:cognitum-cogs:cog-elevator-count","canonical":"cog-elevator-count","tier":null,"t":["cog-elevator-count","cog","elevator","count","people","signal","disruption","intensity","linear","regression","calibration"],"s":["cog-elevator-count","cog","elevator","count"]},
+{"id":"cog-emotion-detect","family":"cog-emotion-detect","kind":"crate","store":"cognitum-cogs","version":null,"description":"Emotion detection — arousal/valence from body movement via Russell circumplex model","keywords":[],"source":"cognitum-cogs/src/cogs/emotion-detect/Cargo.toml","sourceSha256":"65b4648bf3f7fb64091257d358bfd29fa3d84c25cb15f8857a2e7fb15baea485","variants":[],"product":"key:cognitum-cogs:cog-emotion-detect","canonical":"cog-emotion-detect","tier":null,"t":["cog-emotion-detect","cog","emotion","detect","detection","arousal","valence","body","movement","via","russell","circumplex","model"],"s":["cog-emotion-detect","cog","emotion","detect"]},
+{"id":"cog-energy-audit","family":"cog-energy-audit","kind":"crate","store":"cognitum-cogs","version":null,"description":"Learn daily usage patterns with hourly Welford stats, detect wasted energy","keywords":[],"source":"cognitum-cogs/src/cogs/energy-audit/Cargo.toml","sourceSha256":"ccd60d8fd89f3249c91185aed2af2504b60a564386c64b7fba8e27a541e57a33","variants":[],"product":"key:cognitum-cogs:cog-energy-audit","canonical":"cog-energy-audit","tier":null,"t":["cog-energy-audit","cog","energy","audit","learn","daily","usage","pattern","hourly","welford","stat","detect","wast"],"s":["cog-energy-audit","cog","energy","audit"]},
+{"id":"cog-energy-harvester","family":"cog-energy-harvester","kind":"crate","store":"cognitum-cogs","version":null,"description":"Monitor power metrics, track solar vs battery, optimize duty cycle","keywords":[],"source":"cognitum-cogs/src/cogs/energy-harvester/Cargo.toml","sourceSha256":"01e800d0b155f58364af078f2a3387ff2d963131e21b89304e7d2a271caecb73","variants":[],"product":"key:cognitum-cogs:cog-energy-harvester","canonical":"cog-energy-harvester","tier":null,"t":["cog-energy-harvester","cog","energy","harvester","monitor","power","metric","track","solar","battery","optimize","duty","cycle"],"s":["cog-energy-harvester","cog","energy","harvester"]},
+{"id":"cog-ewc-lifelong","family":"cog-ewc-lifelong","kind":"crate","store":"cognitum-cogs","version":null,"description":"Elastic Weight Consolidation — prevent catastrophic forgetting of learned baselines","keywords":[],"source":"cognitum-cogs/src/cogs/ewc-lifelong/Cargo.toml","sourceSha256":"eac7b6719f8825964481b54fd9064dcd269e990c92b70e684623f898da23aaed","variants":[],"product":"key:cognitum-cogs:cog-ewc-lifelong","canonical":"cog-ewc-lifelong","tier":null,"t":["cog-ewc-lifelong","cog","ewc","lifelong","elastic","weight","consolidation","prevent","catastrophic","forgett","learn","baseline"],"s":["cog-ewc-lifelong","cog","ewc","lifelong"]},
+{"id":"cog-fall-detect","family":"cog-fall-detect","kind":"crate","store":"cognitum-cogs","version":null,"description":"Fall detection — impact-spike + post-fall stillness state machine over ambient feature stream","keywords":[],"source":"cognitum-cogs/src/cogs/fall-detect/Cargo.toml","sourceSha256":"7044e7393763b0f131170114926ef6dbd987728619c4fb168e9ddac88a68959d","variants":[],"product":"key:cognitum-cogs:cog-fall-detect","canonical":"cog-fall-detect","tier":null,"t":["cog-fall-detect","cog","fall","detect","detection","impact-spike","impact","spike","post-fall","post","stillness","state","machine","ambient","feature","stream"],"s":["cog-fall-detect","cog","fall","detect"]},
+{"id":"cog-federated-learning","family":"cog-federated-learning","kind":"crate","store":"cognitum-cogs","version":null,"description":"Cross-seed model averaging — federated parameter aggregation","keywords":[],"source":"cognitum-cogs/src/cogs/federated-learning/Cargo.toml","sourceSha256":"a430067e7f17eb2c393d26e721a210659f9f97e664b3252debdb4a52854d26dc","variants":[],"product":"key:cognitum-cogs:cog-federated-learning","canonical":"cog-federated-learning","tier":null,"t":["cog-federated-learning","cog","federat","learn","cross-seed","cross","seed","model","averag","parameter","aggregation"],"s":["cog-federated-learning","cog","federat","learn"]},
+{"id":"cog-flash-attention","family":"cog-flash-attention","kind":"crate","store":"cognitum-cogs","version":null,"description":"Flash attention — adaptive sensing, focus on highest-information channels","keywords":[],"source":"cognitum-cogs/src/cogs/flash-attention/Cargo.toml","sourceSha256":"768466b93b32521f5d81d1f05fa782fa78e9b9e6080025564422583da4cbed69","variants":[],"product":"key:cognitum-cogs:cog-flash-attention","canonical":"cog-flash-attention","tier":null,"t":["cog-flash-attention","cog","flash","attention","adaptive","sens","focus","highest-information","highest","information","channel"],"s":["cog-flash-attention","cog","flash","attention"]},
+{"id":"cog-fleet-auth","family":"cog-fleet-auth","kind":"crate","store":"cognitum-cogs","version":null,"description":"Device fleet authentication — Ed25519 keypair, challenge-response, device registry","keywords":[],"source":"cognitum-cogs/src/cogs/fleet-auth/Cargo.toml","sourceSha256":"f1d0d21cd781dd61c650b71162965c3a58d2148d41931102880b155a7d239d46","variants":[],"product":"key:cognitum-cogs:cog-fleet-auth","canonical":"cog-fleet-auth","tier":null,"t":["cog-fleet-auth","cog","fleet","auth","device","authentication","ed25519","keypair","challenge-response","challenge","response","registry"],"s":["cog-fleet-auth","cog","fleet","auth"]},
+{"id":"cog-forklift-proximity","family":"cog-forklift-proximity","kind":"crate","store":"cognitum-cogs","version":null,"description":"Detect forklift proximity via strong multi-channel signal disruption","keywords":[],"source":"cognitum-cogs/src/cogs/forklift-proximity/Cargo.toml","sourceSha256":"4de574ef7df1378651298c4ebe4536194db8cb35ea2e91d8aa353a7fe2eed63a","variants":[],"product":"key:cognitum-cogs:cog-forklift-proximity","canonical":"cog-forklift-proximity","tier":null,"t":["cog-forklift-proximity","cog","forklift","proximity","detect","via","strong","multi-channel","multi","channel","signal","disruption"],"s":["cog-forklift-proximity","cog","forklift","proximity"]},
+{"id":"cog-frost-warning","family":"cog-frost-warning","kind":"crate","store":"cognitum-cogs","version":null,"description":"Frost warning — temperature trend extrapolation with dewpoint-depression gate","keywords":[],"source":"cognitum-cogs/src/cogs/frost-warning/Cargo.toml","sourceSha256":"8b67f3e34e1e7bdf467be3bb3744aa1bfc07717a91a688c42c4f21062f00044c","variants":[],"product":"key:cognitum-cogs:cog-frost-warning","canonical":"cog-frost-warning","tier":null,"t":["cog-frost-warning","cog","frost","warn","temperature","trend","extrapolation","dewpoint-depression","dewpoint","depression","gate"],"s":["cog-frost-warning","cog","frost","warn"]},
+{"id":"cog-gait-analysis","family":"cog-gait-analysis","kind":"crate","store":"cognitum-cogs","version":null,"description":"Gait analysis — cadence, symmetry, regularity, fall risk via autocorrelation","keywords":[],"source":"cognitum-cogs/src/cogs/gait-analysis/Cargo.toml","sourceSha256":"849cebc917137973f8fbb5b9792ec961cd5f4bf5b391184411bc504c8c4013da","variants":[],"product":"key:cognitum-cogs:cog-gait-analysis","canonical":"cog-gait-analysis","tier":null,"t":["cog-gait-analysis","cog","gait","analysis","cadence","symmetry","regularity","fall","risk","via","autocorrelation"],"s":["cog-gait-analysis","cog","gait","analysis"]},
+{"id":"cog-gesture","family":"cog-gesture","kind":"crate","store":"cognitum-cogs","version":null,"description":"Core gesture recognition — template-based DTW matching building block","keywords":[],"source":"cognitum-cogs/src/cogs/gesture/Cargo.toml","sourceSha256":"e4ba2cc44a31eac50ceaa92b162e3ea1b85f691f88f26c3ec10bd627dad07e4c","variants":[],"product":"key:cognitum-cogs:cog-gesture","canonical":"cog-gesture","tier":null,"t":["cog-gesture","cog","gesture","core","recognition","template-based","template","dtw","match","block"],"s":["cog-gesture","cog","gesture"]},
+{"id":"cog-gesture-language","family":"cog-gesture-language","kind":"crate","store":"cognitum-cogs","version":null,"description":"Gesture recognition — DTW pattern matching against stored templates","keywords":[],"source":"cognitum-cogs/src/cogs/gesture-language/Cargo.toml","sourceSha256":"50bcb6d2b2960a1e9e7c616b41e1e2e354b0a9dbb3fae4d6e326846682e83181","variants":[],"product":"key:cognitum-cogs:cog-gesture-language","canonical":"cog-gesture-language","tier":null,"t":["cog-gesture-language","cog","gesture","language","recognition","dtw","pattern","match","against","stor","template"],"s":["cog-gesture-language","cog","gesture","language"]},
+{"id":"cog-ghost-hunter","family":"cog-ghost-hunter","kind":"crate","store":"cognitum-cogs","version":null,"description":"Ghost hunter — detect environmental anomalies via multi-channel correlation breakdown","keywords":[],"source":"cognitum-cogs/src/cogs/ghost-hunter/Cargo.toml","sourceSha256":"b1e0311e7c75f3e3d06af7ca64c379065c928c6aee9061f9f29c823eb62770cc","variants":[],"product":"key:cognitum-cogs:cog-ghost-hunter","canonical":"cog-ghost-hunter","tier":null,"t":["cog-ghost-hunter","cog","ghost","hunter","detect","environmental","anomaly","via","multi-channel","multi","channel","correlation","breakdown"],"s":["cog-ghost-hunter","cog","ghost","hunter"]},
+{"id":"cog-glass-break","family":"cog-glass-break","kind":"crate","store":"cognitum-cogs","version":null,"description":"Glass-break detection — two-phase bang + shatter acoustic detector","keywords":[],"source":"cognitum-cogs/src/cogs/glass-break/Cargo.toml","sourceSha256":"02ea99f13256183d1a85d402514495753a9348865a6d37eccb878d71504cdef8","variants":[],"product":"key:cognitum-cogs:cog-glass-break","canonical":"cog-glass-break","tier":null,"t":["cog-glass-break","cog","glass","break","glass-break","detection","two-phase","phase","bang","shatter","acoustic","detector"],"s":["cog-glass-break","cog","glass","break"]},
+{"id":"cog-goap-autonomy","family":"cog-goap-autonomy","kind":"crate","store":"cognitum-cogs","version":null,"description":"Goal-Oriented Action Planning — autonomous action sequence planning","keywords":[],"source":"cognitum-cogs/src/cogs/goap-autonomy/Cargo.toml","sourceSha256":"fd9b0272da3032886edf1262d515aeef8dc6702bcf39d9e0e04cb0f3fdc8774f","variants":[],"product":"key:cognitum-cogs:cog-goap-autonomy","canonical":"cog-goap-autonomy","tier":null,"t":["cog-goap-autonomy","cog","goap","autonomy","goal-oriented","goal","orient","action","plann","autonomous","sequence"],"s":["cog-goap-autonomy","cog","goap","autonomy"]},
+{"id":"cog-gunshot-detect","family":"cog-gunshot-detect","kind":"crate","store":"cognitum-cogs","version":null,"description":"Gunshot detection — saturating peak + exponential decay acoustic detector","keywords":[],"source":"cognitum-cogs/src/cogs/gunshot-detect/Cargo.toml","sourceSha256":"ebac2aae87300509d88eab1f56f27d9e918097530e0ca8c34e850808ae1d11b9","variants":[],"product":"key:cognitum-cogs:cog-gunshot-detect","canonical":"cog-gunshot-detect","tier":null,"t":["cog-gunshot-detect","cog","gunshot","detect","detection","saturat","peak","exponential","decay","acoustic","detector"],"s":["cog-gunshot-detect","cog","gunshot","detect"]},
+{"id":"cog-ha-matter","family":"cog-ha-matter","kind":"crate","store":"ruview","version":null,"description":"Cognitum Cog: Home Assistant (MQTT) integration for the Seed (ADR-116). Wraps ADR-115's HA-DISCO + HA-MIND publisher as a Seed-installable artifact with mDNS, embedded broker, RuVector-backed thresholds, and Ed25519 witness. LAN-only (no TLS); Matter Bridge commissioning is deferred to v0.8 and not yet implemented.","keywords":[],"source":"ruview/v2/crates/cog-ha-matter/Cargo.toml","sourceSha256":"b96b9d9c5556fbad667bc5ef021ebd0a94b5fc73ccc930c8c03f33bdf1ec2671","variants":[],"product":"key:ruview:cog-ha-matter","canonical":"cog-ha-matter","tier":"T0","t":["cog-ha-matter","cog","matter","cognitum","home","assistant","mqtt","integration","seed","adr-116","adr","116","wrap","adr-115","115","ha-disco","disco","ha-mind","mind","publisher","seed-installable","installable","artifact","mdns","embed","broker","ruvector-backed","back","threshold","ed25519","witness","lan-only","lan","only","tls","bridge","commission","defer","v0.8","yet","implement"],"s":["cog-ha-matter","cog","matter"]},
+{"id":"cog-happiness-score","family":"cog-happiness-score","kind":"crate","store":"cognitum-cogs","version":null,"description":"Happiness score — composite well-being from movement, breathing regularity, activity","keywords":[],"source":"cognitum-cogs/src/cogs/happiness-score/Cargo.toml","sourceSha256":"76973ab45f8f61b3f4addee875b93dc7b8e53db31bef3c27ad961201b8dc6cf2","variants":[],"product":"key:cognitum-cogs:cog-happiness-score","canonical":"cog-happiness-score","tier":null,"t":["cog-happiness-score","cog","happiness","score","composite","well-being","well","movement","breath","regularity","activity"],"s":["cog-happiness-score","cog","happiness","score"]},
+{"id":"cog-health-monitor","family":"cog-health-monitor","kind":"crate","store":"cognitum-cogs","version":null,"description":"Combined vital signs dashboard — vitals, presence, apnea in one process. Self-contained sensor sources (seed-stream / esp32-uart / esp32-udp).","keywords":[],"source":"cognitum-cogs/src/cogs/health-monitor/Cargo.toml","sourceSha256":"ce8a07145921893a5fcc4be8c91cebfa49bd9c4b1c103c16b219405a4bcea82d","variants":[],"product":"key:cognitum-cogs:cog-health-monitor","canonical":"cog-health-monitor","tier":null,"t":["cog-health-monitor","cog","health","monitor","combin","vital","sign","dashboard","presence","apnea","process","self-contained","self","contain","sensor","source","seed-stream","seed","stream","esp32-uart","esp32","uart","esp32-udp","udp"],"s":["cog-health-monitor","cog","health","monitor"]},
+{"id":"cog-hvac-presence","family":"cog-hvac-presence","kind":"crate","store":"cognitum-cogs","version":null,"description":"Arrival/departure detection with debounce for HVAC control","keywords":[],"source":"cognitum-cogs/src/cogs/hvac-presence/Cargo.toml","sourceSha256":"cac3a64fbc642c644df703298b01c847d2531b8c12d6f59673e098bffe1318b4","variants":[],"product":"key:cognitum-cogs:cog-hvac-presence","canonical":"cog-hvac-presence","tier":null,"t":["cog-hvac-presence","cog","hvac","presence","arrival","departure","detection","debounce","control"],"s":["cog-hvac-presence","cog","hvac","presence"]},
+{"id":"cog-hyperbolic-space","family":"cog-hyperbolic-space","kind":"crate","store":"cognitum-cogs","version":null,"description":"Hyperbolic space — Poincare ball model for hierarchical signal relationships","keywords":[],"source":"cognitum-cogs/src/cogs/hyperbolic-space/Cargo.toml","sourceSha256":"bc68554dd93a711306e8d0b10360d946d19f34b39015e226bd95e3faa3cbac7c","variants":[],"product":"key:cognitum-cogs:cog-hyperbolic-space","canonical":"cog-hyperbolic-space","tier":null,"t":["cog-hyperbolic-space","cog","hyperbolic","space","poincare","ball","model","hierarchical","signal","relationship"],"s":["cog-hyperbolic-space","cog","hyperbolic","space"]},
+{"id":"cog-interference-search","family":"cog-interference-search","kind":"crate","store":"cognitum-cogs","version":null,"description":"Grover-inspired search — amplitude amplification anomaly detection","keywords":[],"source":"cognitum-cogs/src/cogs/interference-search/Cargo.toml","sourceSha256":"60564e9a103880ca46fa1909456024170ee07745d5912c45181a56bb9565a32e","variants":[],"product":"key:cognitum-cogs:cog-interference-search","canonical":"cog-interference-search","tier":null,"t":["cog-interference-search","cog","interference","search","grover-inspired","grover","inspir","amplitude","amplification","anomaly","detection"],"s":["cog-interference-search","cog","interference","search"]},
+{"id":"cog-intrusion","family":"cog-intrusion","kind":"crate","store":"cognitum-cogs","version":null,"description":"Intrusion detection — unauthorized presence alerting with zone tracking","keywords":[],"source":"cognitum-cogs/src/cogs/intrusion/Cargo.toml","sourceSha256":"8aa5c93a3f579e2bfcbbd34d89c799af9f89e0b1c62d0f81643947cdf26ce618","variants":[],"product":"key:cognitum-cogs:cog-intrusion","canonical":"cog-intrusion","tier":null,"t":["cog-intrusion","cog","intrusion","detection","unauthoriz","presence","alert","zone","track"],"s":["cog-intrusion","cog","intrusion"]},
+{"id":"cog-intrusion-detect-ml","family":"cog-intrusion-detect-ml","kind":"crate","store":"cognitum-cogs","version":null,"description":"ML-based intrusion detection — online learning with nearest-neighbor classification","keywords":[],"source":"cognitum-cogs/src/cogs/intrusion-detect-ml/Cargo.toml","sourceSha256":"3219441380d5babff8914cb4f0b833434ffa325a6934227e22f602a95e347ded","variants":[],"product":"key:cognitum-cogs:cog-intrusion-detect-ml","canonical":"cog-intrusion-detect-ml","tier":null,"t":["cog-intrusion-detect-ml","cog","intrusion","detect","ml-based","detection","online","learn","nearest-neighbor","nearest","neighbor","classification"],"s":["cog-intrusion-detect-ml","cog","intrusion","detect"]},
+{"id":"cog-lighting-zones","family":"cog-lighting-zones","kind":"crate","store":"cognitum-cogs","version":null,"description":"Movement-based zone lighting control with auto-off timeout","keywords":[],"source":"cognitum-cogs/src/cogs/lighting-zones/Cargo.toml","sourceSha256":"af850cd7296fa463a9d02b80014547d8b229f69d0b158391bd3e0af0654578ac","variants":[],"product":"key:cognitum-cogs:cog-lighting-zones","canonical":"cog-lighting-zones","tier":null,"t":["cog-lighting-zones","cog","light","zone","movement-based","movement","control","auto-off","auto","off","timeout"],"s":["cog-lighting-zones","cog","light","zone"]},
+{"id":"cog-livestock-monitor","family":"cog-livestock-monitor","kind":"crate","store":"cognitum-cogs","version":null,"description":"Detect animal activity patterns, distress, and escape via CSI","keywords":[],"source":"cognitum-cogs/src/cogs/livestock-monitor/Cargo.toml","sourceSha256":"a082da1b229270343aacd64e6a8a05067788beb57bfac2fd82ba5626c81a026f","variants":[],"product":"key:cognitum-cogs:cog-livestock-monitor","canonical":"cog-livestock-monitor","tier":null,"t":["cog-livestock-monitor","cog","livestock","monitor","detect","animal","activity","pattern","distress","escape","via","csi"],"s":["cog-livestock-monitor","cog","livestock","monitor"]},
+{"id":"cog-loitering","family":"cog-loitering","kind":"crate","store":"cognitum-cogs","version":null,"description":"Loitering detection — timer-based stationary presence alerting","keywords":[],"source":"cognitum-cogs/src/cogs/loitering/Cargo.toml","sourceSha256":"3991e9763e1e541f7e69d95c41552692bdeb2aa08d0297e80e4cd0f8f4a5a81b","variants":[],"product":"key:cognitum-cogs:cog-loitering","canonical":"cog-loitering","tier":null,"t":["cog-loitering","cog","loiter","detection","timer-based","timer","stationary","presence","alert"],"s":["cog-loitering","cog","loiter"]},
+{"id":"cog-meeting-room","family":"cog-meeting-room","kind":"crate","store":"cognitum-cogs","version":null,"description":"Binary occupied/free status with confidence, duration, and utilization tracking","keywords":[],"source":"cognitum-cogs/src/cogs/meeting-room/Cargo.toml","sourceSha256":"0541d89438a598f5ed69a2d86ede9b41e20d563c1f869734e10c6ca7c46a89f2","variants":[],"product":"key:cognitum-cogs:cog-meeting-room","canonical":"cog-meeting-room","tier":null,"t":["cog-meeting-room","cog","meet","room","binary","occupy","free","status","confidence","duration","utilization","track"],"s":["cog-meeting-room","cog","meet","room"]},
+{"id":"cog-meta-adapt","family":"cog-meta-adapt","kind":"crate","store":"cognitum-cogs","version":null,"description":"Self-tuning cog — auto-adjust thresholds via golden section search","keywords":[],"source":"cognitum-cogs/src/cogs/meta-adapt/Cargo.toml","sourceSha256":"c5025b0b7ca583f5b38b5417ce9c3a09a115441468b1b129a3b91ecb2f9869b8","variants":[],"product":"key:cognitum-cogs:cog-meta-adapt","canonical":"cog-meta-adapt","tier":null,"t":["cog-meta-adapt","cog","meta","adapt","self-tuning","self","tuning","auto-adjust","auto","adjust","threshold","via","golden","section","search"],"s":["cog-meta-adapt","cog","meta","adapt"]},
+{"id":"cog-micro-hnsw","family":"cog-micro-hnsw","kind":"crate","store":"cognitum-cogs","version":null,"description":"Hierarchical Navigable Small World graph — fast fingerprint matching","keywords":[],"source":"cognitum-cogs/src/cogs/micro-hnsw/Cargo.toml","sourceSha256":"511499a9000c5b976a95d34686c70d1381d57021ac57a85a47a1fd757c547292","variants":[],"product":"key:cognitum-cogs:cog-micro-hnsw","canonical":"cog-micro-hnsw","tier":null,"t":["cog-micro-hnsw","cog","micro","hnsw","hierarchical","navigable","small","world","graph","fingerprint","match"],"s":["cog-micro-hnsw","cog","micro","hnsw"]},
+{"id":"cog-music-conductor","family":"cog-music-conductor","kind":"crate","store":"cognitum-cogs","version":null,"description":"Music conductor — extract tempo from periodic arm gestures via autocorrelation","keywords":[],"source":"cognitum-cogs/src/cogs/music-conductor/Cargo.toml","sourceSha256":"cb5c876ccf7bc34a202331ed7b6abfa9c79f4dfd210b1cf64ad6386ed82d79a5","variants":[],"product":"key:cognitum-cogs:cog-music-conductor","canonical":"cog-music-conductor","tier":null,"t":["cog-music-conductor","cog","music","conductor","extract","tempo","periodic","arm","gesture","via","autocorrelation"],"s":["cog-music-conductor","cog","music","conductor"]},
+{"id":"cog-network-firewall","family":"cog-network-firewall","kind":"crate","store":"cognitum-cogs","version":null,"description":"Network connection monitor — detect unauthorized outbound connections per cog","keywords":[],"source":"cognitum-cogs/src/cogs/network-firewall/Cargo.toml","sourceSha256":"afe0acdf57e496ee4998b4d7f468f2585374c5242b4565be9b2fea49fcf570d0","variants":[],"product":"key:cognitum-cogs:cog-network-firewall","canonical":"cog-network-firewall","tier":null,"t":["cog-network-firewall","cog","network","firewall","connection","monitor","detect","unauthoriz","outbound","per"],"s":["cog-network-firewall","cog","network","firewall"]},
+{"id":"cog-neural-trader","family":"cog-neural-trader","kind":"crate","store":"cognitum-cogs","version":null,"description":"Market data analysis — OHLCV embedding and pattern matching","keywords":[],"source":"cognitum-cogs/src/cogs/neural-trader/Cargo.toml","sourceSha256":"5186a30cf407fcbf7c7cb49bf62509565d4d4859ab75754cb65c894282722638","variants":[],"product":"key:cognitum-cogs:cog-neural-trader","canonical":"cog-neural-trader","tier":null,"t":["cog-neural-trader","cog","neural","trader","market","analysis","ohlcv","embedd","pattern","match"],"s":["cog-neural-trader","cog","neural","trader"]},
+{"id":"cog-occupancy-zones","family":"cog-occupancy-zones","kind":"crate","store":"cognitum-cogs","version":null,"description":"Count people per zone — presence = variance above threshold per channel","keywords":[],"source":"cognitum-cogs/src/cogs/occupancy-zones/Cargo.toml","sourceSha256":"c97fa2c37a465e356e46f08bfc3a2a8cc7a62dfe11362f0a1f5e55a3d1f3c090","variants":[],"product":"key:cognitum-cogs:cog-occupancy-zones","canonical":"cog-occupancy-zones","tier":null,"t":["cog-occupancy-zones","cog","occupancy","zone","count","people","per","presence","variance","above","threshold","channel"],"s":["cog-occupancy-zones","cog","occupancy","zone"]},
+{"id":"cog-optimal-transport","family":"cog-optimal-transport","kind":"crate","store":"cognitum-cogs","version":null,"description":"Optimal transport — Wasserstein distance between signal distributions over time","keywords":[],"source":"cognitum-cogs/src/cogs/optimal-transport/Cargo.toml","sourceSha256":"401b6b58a47db3f1094b6780d41f77ced1a237e42619442005f49a52956989d2","variants":[],"product":"key:cognitum-cogs:cog-optimal-transport","canonical":"cog-optimal-transport","tier":null,"t":["cog-optimal-transport","cog","optimal","transport","wasserstein","distance","between","signal","distribution"],"s":["cog-optimal-transport","cog","optimal","transport"]},
+{"id":"cog-package-detect","family":"cog-package-detect","kind":"crate","store":"cognitum-cogs","version":null,"description":"Package arrival detection — sustained CSI subcarrier-shift detector for porches / loading bays. Requires ruview.","keywords":[],"source":"cognitum-cogs/src/cogs/package-detect/Cargo.toml","sourceSha256":"102072eae56c7b363917efe9a06b81caeb2af0cdbe29f9a2fea5fa7cbb295433","variants":[],"product":"key:cognitum-cogs:cog-package-detect","canonical":"cog-package-detect","tier":null,"t":["cog-package-detect","cog","detect","arrival","detection","sustain","csi","subcarrier-shift","subcarrier","shift","detector","porch","load","bays","require","ruview"],"s":["cog-package-detect","cog","detect"]},
+{"id":"cog-pagerank-influence","family":"cog-pagerank-influence","kind":"crate","store":"cognitum-cogs","version":null,"description":"Graph-based influence detection — channel correlation PageRank","keywords":[],"source":"cognitum-cogs/src/cogs/pagerank-influence/Cargo.toml","sourceSha256":"2a29b6f2388fc930c27e641e27da05ffab852fd6fab9650933ee9086397aef51","variants":[],"product":"key:cognitum-cogs:cog-pagerank-influence","canonical":"cog-pagerank-influence","tier":null,"t":["cog-pagerank-influence","cog","pagerank","influence","graph-based","graph","detection","channel","correlation"],"s":["cog-pagerank-influence","cog","pagerank","influence"]},
+{"id":"cog-panic-motion","family":"cog-panic-motion","kind":"crate","store":"cognitum-cogs","version":null,"description":"Panic motion detection — erratic movement via high-frequency energy and jerk analysis","keywords":[],"source":"cognitum-cogs/src/cogs/panic-motion/Cargo.toml","sourceSha256":"e3a92cd06a58406685dc5a8b98c91ce9e17cb8ce8188b3491c62014b363858a4","variants":[],"product":"key:cognitum-cogs:cog-panic-motion","canonical":"cog-panic-motion","tier":null,"t":["cog-panic-motion","cog","panic","motion","detection","erratic","movement","via","high-frequency","high","frequency","energy","jerk","analysis"],"s":["cog-panic-motion","cog","panic","motion"]},
+{"id":"cog-parking-occupancy","family":"cog-parking-occupancy","kind":"crate","store":"cognitum-cogs","version":null,"description":"Parking occupancy via ESP32 CSI zone tracking. Requires ruview.","keywords":[],"source":"cognitum-cogs/src/cogs/parking-occupancy/Cargo.toml","sourceSha256":"8fc704a127084fe02c6d46f4563d95929439dd5ee077d7c2229b6cf24821d64e","variants":[],"product":"key:cognitum-cogs:cog-parking-occupancy","canonical":"cog-parking-occupancy","tier":null,"t":["cog-parking-occupancy","cog","park","occupancy","via","esp32","csi","zone","track","require","ruview"],"s":["cog-parking-occupancy","cog","park","occupancy"]},
+{"id":"cog-pattern-sequence","family":"cog-pattern-sequence","kind":"crate","store":"cognitum-cogs","version":null,"description":"Daily routine detection — hourly pattern matching and deviation scoring","keywords":[],"source":"cognitum-cogs/src/cogs/pattern-sequence/Cargo.toml","sourceSha256":"151de4493e60453c9d631b0769b89e4b563d795b8a613315b1ab5088335bfbc2","variants":[],"product":"key:cognitum-cogs:cog-pattern-sequence","canonical":"cog-pattern-sequence","tier":null,"t":["cog-pattern-sequence","cog","pattern","sequence","daily","routine","detection","hourly","match","deviation","scor"],"s":["cog-pattern-sequence","cog","pattern","sequence"]},
+{"id":"cog-perimeter-breach","family":"cog-perimeter-breach","kind":"crate","store":"cognitum-cogs","version":null,"description":"Multi-zone perimeter monitoring — detects entry direction via zone trigger sequence","keywords":[],"source":"cognitum-cogs/src/cogs/perimeter-breach/Cargo.toml","sourceSha256":"0ab755cd65568a969668af0d8519750f7f784a42dd6dbd964041d85596f732a1","variants":[],"product":"key:cognitum-cogs:cog-perimeter-breach","canonical":"cog-perimeter-breach","tier":null,"t":["cog-perimeter-breach","cog","perimeter","breach","multi-zone","multi","zone","monitor","detect","entry","direction","via","trigger","sequence"],"s":["cog-perimeter-breach","cog","perimeter","breach"]},
+{"id":"cog-person-count","family":"cog-person-count","kind":"crate","store":"ruview","version":null,"description":"Cognitum Cog: WiFi-CSI presence detector + (data-gated) person count (ADR-103). Candle-based head trained on classes 0/1 (presence); the 8-class count head ships but counts above the trained range are flagged low_confidence. Stoer-Wagner multi-node fusion.","keywords":[],"source":"ruview/v2/crates/cog-person-count/Cargo.toml","sourceSha256":"296b9a97f2ae0b22aeb6fd2f616b1adec703d14ae5d4d682996172b90e68a779","variants":[],"product":"key:ruview:cog-person-count","canonical":"cog-person-count","tier":"T0","t":["cog-person-count","cog","person","count","cognitum","wifi-csi","wifi","csi","presence","detector","data-gated","gated","adr-103","adr","103","candle-based","candle","head","train","class","8-class","ship","above","range","flag","low","confidence","stoer-wagner","stoer","wagner","multi-node","multi","fusion"],"s":["cog-person-count","cog","person","count"]},
+{"id":"cog-person-matching","family":"cog-person-matching","kind":"crate","store":"cognitum-cogs","version":null,"description":"Person matching — distinguish multiple people via signal feature clustering","keywords":[],"source":"cognitum-cogs/src/cogs/person-matching/Cargo.toml","sourceSha256":"a9063b6980572ebdc0f61d77207a885d6829e4d26459e52e078cc7821aa75443","variants":[],"product":"key:cognitum-cogs:cog-person-matching","canonical":"cog-person-matching","tier":null,"t":["cog-person-matching","cog","person","match","distinguish","multiple","people","via","signal","feature","cluster"],"s":["cog-person-matching","cog","person","match"]},
+{"id":"cog-plant-growth","family":"cog-plant-growth","kind":"crate","store":"cognitum-cogs","version":null,"description":"Plant growth tracker — daily growth rate, day/night cycles, rolling 24h stats","keywords":[],"source":"cognitum-cogs/src/cogs/plant-growth/Cargo.toml","sourceSha256":"b662b0bf93be33fd6c51f96074f9443240e77acca96c701fb56589b2bb597ac9","variants":[],"product":"key:cognitum-cogs:cog-plant-growth","canonical":"cog-plant-growth","tier":null,"t":["cog-plant-growth","cog","plant","growth","tracker","daily","rate","day","night","cycle","roll","24h","stat"],"s":["cog-plant-growth","cog","plant","growth"]},
+{"id":"cog-pose-estimation","family":"cog-pose-estimation","kind":"crate","store":"ruview","version":null,"description":"Cognitum Cog: 17-keypoint pose estimation from WiFi CSI. See ADR-100 (packaging) + ADR-101 (this Cog).","keywords":[],"source":"ruview/v2/crates/cog-pose-estimation/Cargo.toml","sourceSha256":"1e1adb4e83def95c636e98960348defa5a149dc3effb421102e38a78829d24d7","variants":[],"product":"key:ruview:cog-pose-estimation","canonical":"cog-pose-estimation","tier":"T0","t":["cog-pose-estimation","cog","pose","estimation","cognitum","17-keypoint","keypoint","wifi","csi","see","adr-100","adr","100","packag","adr-101","101"],"s":["cog-pose-estimation","cog","pose","estimation"]},
+{"id":"cog-ppe-compliance","family":"cog-ppe-compliance","kind":"crate","store":"cognitum-cogs","version":null,"description":"PPE compliance — fuses ruview-densepose presence with PPE-camera confirmation to flag unprotected entry into restricted zones","keywords":[],"source":"cognitum-cogs/src/cogs/ppe-compliance/Cargo.toml","sourceSha256":"b028aee1bbef322234ec42aaf5bba1faf6e2e7358fad1bd44e88ef46344b1371","variants":[],"product":"key:cognitum-cogs:cog-ppe-compliance","canonical":"cog-ppe-compliance","tier":null,"t":["cog-ppe-compliance","cog","ppe","compliance","fus","ruview-densepose","ruview","densepose","presence","ppe-camera","camera","confirmation","flag","unprotect","entry","restrict","zone"],"s":["cog-ppe-compliance","cog","ppe","compliance"]},
+{"id":"cog-predictive-maintenance","family":"cog-predictive-maintenance","kind":"crate","store":"cognitum-cogs","version":null,"description":"Predictive maintenance — vibration harmonic analysis for rotating machinery (pumps, motors, fans)","keywords":[],"source":"cognitum-cogs/src/cogs/predictive-maintenance/Cargo.toml","sourceSha256":"028e136bef358e707893b7a087254fbf2cedfa6e6dba5026b57d0f2574ca9ec9","variants":[],"product":"key:cognitum-cogs:cog-predictive-maintenance","canonical":"cog-predictive-maintenance","tier":null,"t":["cog-predictive-maintenance","cog","predictive","maintenance","vibration","harmonic","analysis","rotat","machinery","pump","motor","fans"],"s":["cog-predictive-maintenance","cog","predictive","maintenance"]},
+{"id":"cog-presence-field","family":"cog-presence-field","kind":"crate","store":"cognitum-cogs","version":null,"description":"🛰️ Multi-node WiFi-CSI presence via field-model residual (ADR-151). Learns an empty-room baseline per node, then presence = body-perturbation residual after projecting out environmental modes, maxed across nodes. Detects a still person where single-node phase-variance cannot.","keywords":[],"source":"cognitum-cogs/src/cogs/presence-field/Cargo.toml","sourceSha256":"04b041ae30e17ba42478543c0d18d306b172e63d32714f934581058665f91a70","variants":[],"product":"key:cognitum-cogs:cog-presence-field","canonical":"cog-presence-field","tier":null,"t":["cog-presence-field","cog","presence","field","multi-node","multi","wifi-csi","wifi","csi","via","field-model","model","residual","adr-151","adr","151","learn","empty-room","empty","room","baseline","per","body-perturbation","body","perturbation","after","project","out","environmental","mode","maxed","across","detect","still","person","single-node","single","phase-variance","phase","variance","cannot"],"s":["cog-presence-field","cog","presence","field"]},
+{"id":"cog-prompt-shield","family":"cog-prompt-shield","kind":"crate","store":"cognitum-cogs","version":null,"description":"Signal replay/injection attack detector — fingerprint matching and anomaly detection","keywords":[],"source":"cognitum-cogs/src/cogs/prompt-shield/Cargo.toml","sourceSha256":"7f82ff8411dddb5cdf16848b69b8246a96510d082868e66121b0f13615bf2127","variants":[],"product":"key:cognitum-cogs:cog-prompt-shield","canonical":"cog-prompt-shield","tier":null,"t":["cog-prompt-shield","cog","prompt","shield","signal","replay","injection","attack","detector","fingerprint","match","anomaly","detection"],"s":["cog-prompt-shield","cog","prompt","shield"]},
+{"id":"cog-psycho-symbolic","family":"cog-psycho-symbolic","kind":"crate","store":"cognitum-cogs","version":null,"description":"Hybrid reasoning — knowledge graph entity-relation event detection","keywords":[],"source":"cognitum-cogs/src/cogs/psycho-symbolic/Cargo.toml","sourceSha256":"4a639b99a95b934a977e19a5549939fa45b027e603ef3f76b9091fbe20a1edf2","variants":[],"product":"key:cognitum-cogs:cog-psycho-symbolic","canonical":"cog-psycho-symbolic","tier":null,"t":["cog-psycho-symbolic","cog","psycho","symbolic","hybrid","reason","knowledge","graph","entity-relation","entity","relation","event","detection"],"s":["cog-psycho-symbolic","cog","psycho","symbolic"]},
+{"id":"cog-quantum-coherence","family":"cog-quantum-coherence","kind":"crate","store":"cognitum-cogs","version":null,"description":"Quantum-inspired signal processing — superposition state detection","keywords":[],"source":"cognitum-cogs/src/cogs/quantum-coherence/Cargo.toml","sourceSha256":"145a6f0f564b411841a23644b90d213a2d0ae063e4a1b42cc083f758f5fc9381","variants":[],"product":"key:cognitum-cogs:cog-quantum-coherence","canonical":"cog-quantum-coherence","tier":null,"t":["cog-quantum-coherence","cog","quantum","coherence","quantum-inspired","inspir","signal","process","superposition","state","detection"],"s":["cog-quantum-coherence","cog","quantum","coherence"]},
+{"id":"cog-queue-length","family":"cog-queue-length","kind":"crate","store":"cognitum-cogs","version":null,"description":"Estimate queue by counting distinct presence zones in a line, track wait time","keywords":[],"source":"cognitum-cogs/src/cogs/queue-length/Cargo.toml","sourceSha256":"5e23ba683053c08ce15b87c006b89d21056479741001a03b81e19481a6ec4a45","variants":[],"product":"key:cognitum-cogs:cog-queue-length","canonical":"cog-queue-length","tier":null,"t":["cog-queue-length","cog","queue","length","estimate","count","distinct","presence","zone","line","track","wait"],"s":["cog-queue-length","cog","queue","length"]},
+{"id":"cog-rag-local","family":"cog-rag-local","kind":"crate","store":"cognitum-cogs","version":null,"description":"Local retrieval-augmented generation — trigram embedding similarity search","keywords":[],"source":"cognitum-cogs/src/cogs/rag-local/Cargo.toml","sourceSha256":"54a49f9be24311cecc98846f9aba3ea8495cead8286d453cab930c92d2bc374f","variants":[],"product":"key:cognitum-cogs:cog-rag-local","canonical":"cog-rag-local","tier":null,"t":["cog-rag-local","cog","rag","local","retrieval-augmented","retrieval","augment","generation","trigram","embedd","similarity","search"],"s":["cog-rag-local","cog","rag","local"]},
+{"id":"cog-rain-detect","family":"cog-rain-detect","kind":"crate","store":"cognitum-cogs","version":null,"description":"Rain detection — broadband amplitude analysis for precipitation classification","keywords":[],"source":"cognitum-cogs/src/cogs/rain-detect/Cargo.toml","sourceSha256":"f7600e5f0689b770e61ceae7758e0119d525000648068ee56123e912742db659","variants":[],"product":"key:cognitum-cogs:cog-rain-detect","canonical":"cog-rain-detect","tier":null,"t":["cog-rain-detect","cog","rain","detect","detection","broadband","amplitude","analysis","precipitation","classification"],"s":["cog-rain-detect","cog","rain","detect"]},
+{"id":"cog-respiratory-distress","family":"cog-respiratory-distress","kind":"crate","store":"cognitum-cogs","version":null,"description":"Respiratory distress monitoring — breathing rate, effort, Cheyne-Stokes detection","keywords":[],"source":"cognitum-cogs/src/cogs/respiratory-distress/Cargo.toml","sourceSha256":"f21dc03005658283ee8d61f7d116b8eade8127c6fd39391779f691d5c25864e2","variants":[],"product":"key:cognitum-cogs:cog-respiratory-distress","canonical":"cog-respiratory-distress","tier":null,"t":["cog-respiratory-distress","cog","respiratory","distress","monitor","breath","rate","effort","cheyne-stokes","cheyne","stoke","detection"],"s":["cog-respiratory-distress","cog","respiratory","distress"]},
+{"id":"cog-runner","family":"cog-runner","kind":"crate","store":"cognitum-cogs","version":null,"description":"Enforces each cog's declared [console] limits at spawn time, and emits the evidence","keywords":[],"source":"cognitum-cogs/crates/cog-runner/Cargo.toml","sourceSha256":"810b9c66c556c1fbc9b0ac1b40a47ce4accde2ae01f99391cd2cf8687f4025d0","variants":[],"product":"key:cognitum-cogs:cog-runner","canonical":"cog-runner","tier":null,"t":["cog-runner","cog","runner","enforce","declar","console","limit","spawn","emit","evidence"],"s":["cog-runner","cog","runner"]},
+{"id":"cog-ruview-densepose","family":"cog-ruview-densepose","kind":"crate","store":"cognitum-cogs","version":null,"description":"RuView integration — ESP32 CSI to 17-keypoint skeleton proxy via Hampel + phase sanitization. Self-contained sensor sources (seed-stream / esp32-uart / esp32-udp).","keywords":[],"source":"cognitum-cogs/src/cogs/ruview-densepose/Cargo.toml","sourceSha256":"f6a1c3608655e10036fe2a6a91f696c8ca0591aa8eb1553a4f4ca9d02a27aa0a","variants":[],"product":"key:cognitum-cogs:cog-ruview-densepose","canonical":"cog-ruview-densepose","tier":null,"t":["cog-ruview-densepose","cog","ruview","densepose","integration","esp32","csi","17-keypoint","keypoint","skeleton","proxy","via","hampel","phase","sanitization","self-contained","self","contain","sensor","source","seed-stream","seed","stream","esp32-uart","uart","esp32-udp","udp"],"s":["cog-ruview-densepose","cog","ruview","densepose"]},
+{"id":"cog-seizure-detect","family":"cog-seizure-detect","kind":"crate","store":"cognitum-cogs","version":null,"description":"Seizure detection — high-frequency burst and post-ictal suppression via energy ratio","keywords":[],"source":"cognitum-cogs/src/cogs/seizure-detect/Cargo.toml","sourceSha256":"092de824bb8cd66ad4277d6c0b8bc58909da4891f5d754bbe386f9bb8800e126","variants":[],"product":"key:cognitum-cogs:cog-seizure-detect","canonical":"cog-seizure-detect","tier":null,"t":["cog-seizure-detect","cog","seizure","detect","detection","high-frequency","high","frequency","burst","post-ictal","post","ictal","suppression","via","energy","ratio"],"s":["cog-seizure-detect","cog","seizure","detect"]},
+{"id":"cog-self-healing-mesh","family":"cog-self-healing-mesh","kind":"crate","store":"cognitum-cogs","version":null,"description":"Sensor mesh health monitor — dropped node detection and redistribution","keywords":[],"source":"cognitum-cogs/src/cogs/self-healing-mesh/Cargo.toml","sourceSha256":"edb8847a57ae2d66823758d48976a55fdf590555ed289568111dfc1d9a116326","variants":[],"product":"key:cognitum-cogs:cog-self-healing-mesh","canonical":"cog-self-healing-mesh","tier":null,"t":["cog-self-healing-mesh","cog","self","heal","mesh","sensor","health","monitor","drop","detection","redistribution"],"s":["cog-self-healing-mesh","cog","self","heal","mesh"]},
+{"id":"cog-sensor-sources","family":"cog-sensor-sources","kind":"crate","store":"cognitum-cogs","version":null,"description":"Shared self-contained sensor sources for cogs (ADR-091)","keywords":[],"source":"cognitum-cogs/crates/cog-sensor-sources/Cargo.toml","sourceSha256":"ebcc328e8a9f8573c8f12986e8bae843a88e78b90805033ad8301f257e2db09a","variants":[],"product":"key:cognitum-cogs:cog-sensor-sources","canonical":"cog-sensor-sources","tier":null,"t":["cog-sensor-sources","cog","sensor","source","shar","self-contained","self","contain","cogs","adr-091","adr","091"],"s":["cog-sensor-sources","cog","sensor","source"]},
+{"id":"cog-shelf-engagement","family":"cog-shelf-engagement","kind":"crate","store":"cognitum-cogs","version":null,"description":"Detect brief interactions near zones — short dwell (5-30s) = engagement","keywords":[],"source":"cognitum-cogs/src/cogs/shelf-engagement/Cargo.toml","sourceSha256":"754ad85e03d4e3844343e6214f42bdf9c92c8074e40af35b1b0856fca3a24652","variants":[],"product":"key:cognitum-cogs:cog-shelf-engagement","canonical":"cog-shelf-engagement","tier":null,"t":["cog-shelf-engagement","cog","shelf","engagement","detect","brief","interaction","near","zone","short","dwell","5-30s","30s"],"s":["cog-shelf-engagement","cog","shelf","engagement"]},
+{"id":"cog-sleep-apnea","family":"cog-sleep-apnea","kind":"crate","store":"cognitum-cogs","version":null,"description":"Sleep apnea detection — breathing cessation via variance spike + CUSUM","keywords":[],"source":"cognitum-cogs/src/cogs/sleep-apnea/Cargo.toml","sourceSha256":"ffd5ebd16473fac25fdd8613dac02f809c6773988032b84c98685779a863ddce","variants":[],"product":"key:cognitum-cogs:cog-sleep-apnea","canonical":"cog-sleep-apnea","tier":null,"t":["cog-sleep-apnea","cog","sleep","apnea","detection","breath","cessation","via","variance","spike","cusum"],"s":["cog-sleep-apnea","cog","sleep","apnea"]},
+{"id":"cog-slip-fall-zone","family":"cog-slip-fall-zone","kind":"crate","store":"cognitum-cogs","version":null,"description":"Slip / wet-floor zone — pre-fall risk detector via motion-variance drop + splash audio + optional cautious-gait CSI","keywords":[],"source":"cognitum-cogs/src/cogs/slip-fall-zone/Cargo.toml","sourceSha256":"aa677d4654f12a324f888a8c45a6640ce81122c9de695cc9de8395cd818b002f","variants":[],"product":"key:cognitum-cogs:cog-slip-fall-zone","canonical":"cog-slip-fall-zone","tier":null,"t":["cog-slip-fall-zone","cog","slip","fall","zone","wet-floor","wet","floor","pre-fall","pre","risk","detector","via","motion-variance","motion","variance","drop","splash","audio","optional","cautious-gait","cautious","gait","csi"],"s":["cog-slip-fall-zone","cog","slip","fall","zone"]},
+{"id":"cog-smoke-fire","family":"cog-smoke-fire","kind":"crate","store":"cognitum-cogs","version":null,"description":"Smoke / fire detection — multi-signal fusion (acoustic crackle + thermal drift + optional ruview plume)","keywords":[],"source":"cognitum-cogs/src/cogs/smoke-fire/Cargo.toml","sourceSha256":"32726ca9192c9958fc8d20b93753c65505db42219dd94e3fa5a1880f54e6ead5","variants":[],"product":"key:cognitum-cogs:cog-smoke-fire","canonical":"cog-smoke-fire","tier":null,"t":["cog-smoke-fire","cog","smoke","fire","detection","multi-signal","multi","signal","fusion","acoustic","crackle","thermal","drift","optional","ruview","plume"],"s":["cog-smoke-fire","cog","smoke","fire"]},
+{"id":"cog-snore-monitor","family":"cog-snore-monitor","kind":"crate","store":"cognitum-cogs","version":null,"description":"Snore monitor — periodic low-band energy tracker with snores-per-minute output","keywords":[],"source":"cognitum-cogs/src/cogs/snore-monitor/Cargo.toml","sourceSha256":"684a28850c69eacfbf87d9725c18dc431ff326362ec7818a50f5f70617033529","variants":[],"product":"key:cognitum-cogs:cog-snore-monitor","canonical":"cog-snore-monitor","tier":null,"t":["cog-snore-monitor","cog","snore","monitor","periodic","low-band","low","band","energy","tracker","snores-per-minute","per","minute","output"],"s":["cog-snore-monitor","cog","snore","monitor"]},
+{"id":"cog-sound-classifier","family":"cog-sound-classifier","kind":"crate","store":"cognitum-cogs","version":null,"description":"Sound classifier — detect glass break, alarm, baby cry from signal modulation","keywords":[],"source":"cognitum-cogs/src/cogs/sound-classifier/Cargo.toml","sourceSha256":"ad0b5ae13468a3b3041ea2fabdba1695f290360d34a6ac305d30e21b0f827a8c","variants":[],"product":"key:cognitum-cogs:cog-sound-classifier","canonical":"cog-sound-classifier","tier":null,"t":["cog-sound-classifier","cog","sound","classifier","detect","glass","break","alarm","baby","cry","signal","modulation"],"s":["cog-sound-classifier","cog","sound","classifier"]},
+{"id":"cog-sparse-recovery","family":"cog-sparse-recovery","kind":"crate","store":"cognitum-cogs","version":null,"description":"Sparse recovery — estimate missing channel data via regression from correlated channels","keywords":[],"source":"cognitum-cogs/src/cogs/sparse-recovery/Cargo.toml","sourceSha256":"c213fbe1d2fea8e8d949240ead060a07bc318cd0e66172770853c3936f72384a","variants":[],"product":"key:cognitum-cogs:cog-sparse-recovery","canonical":"cog-sparse-recovery","tier":null,"t":["cog-sparse-recovery","cog","sparse","recovery","estimate","miss","channel","via","regression","correlat"],"s":["cog-sparse-recovery","cog","sparse","recovery"]},
+{"id":"cog-spiking-tracker","family":"cog-spiking-tracker","kind":"crate","store":"cognitum-cogs","version":null,"description":"Spiking neural network tracker — leaky integrate-and-fire movement detection","keywords":[],"source":"cognitum-cogs/src/cogs/spiking-tracker/Cargo.toml","sourceSha256":"7815b670a30ec1c61304facfc84a4a4c83ab9528f1c0c9f4c5ed74c78aa6d602","variants":[],"product":"key:cognitum-cogs:cog-spiking-tracker","canonical":"cog-spiking-tracker","tier":null,"t":["cog-spiking-tracker","cog","spik","tracker","neural","network","leaky","integrate-and-fire","integrate","fire","movement","detection"],"s":["cog-spiking-tracker","cog","spik","tracker"]},
+{"id":"cog-structural-vibration","family":"cog-structural-vibration","kind":"crate","store":"cognitum-cogs","version":null,"description":"Detect dangerous vibrations via high-pass filter and RMS energy analysis","keywords":[],"source":"cognitum-cogs/src/cogs/structural-vibration/Cargo.toml","sourceSha256":"108c4e2e637fca7b2b30da32928aa6c1c9c472cdf53cf4cf1a126453e32d731c","variants":[],"product":"key:cognitum-cogs:cog-structural-vibration","canonical":"cog-structural-vibration","tier":null,"t":["cog-structural-vibration","cog","structural","vibration","detect","dangerous","via","high-pass","high","pass","filter","rms","energy","analysis"],"s":["cog-structural-vibration","cog","structural","vibration"]},
+{"id":"cog-swarm-backup-restore","family":"cog-swarm-backup-restore","kind":"crate","store":"cognitum-cogs","version":null,"description":"Replicate local vector store to peer seed, track lag, restore on failure","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-backup-restore/Cargo.toml","sourceSha256":"89d763799493e754d77526afb70c814b98367ce4d9a54fd13740c98a41f5a552","variants":[],"product":"key:cognitum-cogs:cog-swarm-backup-restore","canonical":"cog-swarm-backup-restore","tier":null,"t":["cog-swarm-backup-restore","cog","swarm","backup","restore","replicate","local","vector","store","peer","seed","track","lag","failure"],"s":["cog-swarm-backup-restore","cog","swarm","backup","restore"]},
+{"id":"cog-swarm-cluster-monitor","family":"cog-swarm-cluster-monitor","kind":"crate","store":"cognitum-cogs","version":null,"description":"Poll peer seeds for health, aggregate cluster-wide metrics","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-cluster-monitor/Cargo.toml","sourceSha256":"ec4385f8e05fc374347b0192c41927878291601f867296b20846395effa9b579","variants":[],"product":"key:cognitum-cogs:cog-swarm-cluster-monitor","canonical":"cog-swarm-cluster-monitor","tier":null,"t":["cog-swarm-cluster-monitor","cog","swarm","cluster","monitor","poll","peer","seed","health","aggregate","cluster-wide","wide","metric"],"s":["cog-swarm-cluster-monitor","cog","swarm","cluster","monitor"]},
+{"id":"cog-swarm-consensus","family":"cog-swarm-consensus","kind":"crate","store":"cognitum-cogs","version":null,"description":"Multi-seed voting protocol with simple 2-phase commit","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-consensus/Cargo.toml","sourceSha256":"f5686e1ee850bcbe5c765a6c4293f6c6d02d415668cdcc3fb29f4b9fcb8258ef","variants":[],"product":"key:cognitum-cogs:cog-swarm-consensus","canonical":"cog-swarm-consensus","tier":null,"t":["cog-swarm-consensus","cog","swarm","consensus","multi-seed","multi","seed","voting","protocol","simple","2-phase","phase","commit"],"s":["cog-swarm-consensus","cog","swarm","consensus"]},
+{"id":"cog-swarm-delta-sync","family":"cog-swarm-delta-sync","kind":"crate","store":"cognitum-cogs","version":null,"description":"Bi-directional vector sync between seeds with epoch tracking","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-delta-sync/Cargo.toml","sourceSha256":"a51db46769dec8eecc2d4365a6674cb70e9a8e27bb67b6402a7382c0c537c8b2","variants":[],"product":"key:cognitum-cogs:cog-swarm-delta-sync","canonical":"cog-swarm-delta-sync","tier":null,"t":["cog-swarm-delta-sync","cog","swarm","delta","sync","bi-directional","directional","vector","between","seed","epoch","track"],"s":["cog-swarm-delta-sync","cog","swarm","delta","sync"]},
+{"id":"cog-swarm-deploy","family":"cog-swarm-deploy","kind":"crate","store":"cognitum-cogs","version":null,"description":"Install/uninstall cogs on peer seeds via apps API fan-out","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-deploy/Cargo.toml","sourceSha256":"d1f571be71e543b7f90c038ddcd77148a710e12d18b4d3f0dfc2ffdceb8e4883","variants":[],"product":"key:cognitum-cogs:cog-swarm-deploy","canonical":"cog-swarm-deploy","tier":null,"t":["cog-swarm-deploy","cog","swarm","deploy","install","uninstall","cogs","peer","seed","via","apps","fan-out","fan","out"],"s":["cog-swarm-deploy","cog","swarm","deploy"]},
+{"id":"cog-swarm-distributed-store","family":"cog-swarm-distributed-store","kind":"crate","store":"cognitum-cogs","version":null,"description":"Distribute vectors across seeds with hash partitioning and fan-out query","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-distributed-store/Cargo.toml","sourceSha256":"9f3f9d792d8b24c5d5e0cd1f981b4bc1fd5d6d61365b1b47020a242b2bd4aece","variants":[],"product":"key:cognitum-cogs:cog-swarm-distributed-store","canonical":"cog-swarm-distributed-store","tier":null,"t":["cog-swarm-distributed-store","cog","swarm","distribut","store","distribute","vector","across","seed","hash","partition","fan-out","fan","out","query"],"s":["cog-swarm-distributed-store","cog","swarm","distribut","store"]},
+{"id":"cog-swarm-edge-orchestrator","family":"cog-swarm-edge-orchestrator","kind":"crate","store":"cognitum-cogs","version":null,"description":"Manage ESP32 sensor nodes via UDP CSI and serial proxy","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-edge-orchestrator/Cargo.toml","sourceSha256":"d453a4b0d1a2e1043f1fcd95db9da238a24160cd63aa39202dd8941b1330861c","variants":[],"product":"key:cognitum-cogs:cog-swarm-edge-orchestrator","canonical":"cog-swarm-edge-orchestrator","tier":null,"t":["cog-swarm-edge-orchestrator","cog","swarm","edge","orchestrator","manage","esp32","sensor","via","udp","csi","serial","proxy"],"s":["cog-swarm-edge-orchestrator","cog","swarm","edge","orchestrator"]},
+{"id":"cog-swarm-load-balancer","family":"cog-swarm-load-balancer","kind":"crate","store":"cognitum-cogs","version":null,"description":"Distribute queries across seeds based on load metrics","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-load-balancer/Cargo.toml","sourceSha256":"21345af1a7d37f3a16675cf502de49795931febead4b234c9b813bc9c245d98a","variants":[],"product":"key:cognitum-cogs:cog-swarm-load-balancer","canonical":"cog-swarm-load-balancer","tier":null,"t":["cog-swarm-load-balancer","cog","swarm","load","balancer","distribute","query","across","seed","metric"],"s":["cog-swarm-load-balancer","cog","swarm","load","balancer"]},
+{"id":"cog-swarm-mesh-manager","family":"cog-swarm-mesh-manager","kind":"crate","store":"cognitum-cogs","version":null,"description":"Discover peer seeds via subnet scan, build mesh topology registry","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-mesh-manager/Cargo.toml","sourceSha256":"a86fbf4fbc11b217aeac52e860bf8d0436fb299c0d1b5b05eacb0de9b5b38899","variants":[],"product":"key:cognitum-cogs:cog-swarm-mesh-manager","canonical":"cog-swarm-mesh-manager","tier":null,"t":["cog-swarm-mesh-manager","cog","swarm","mesh","manager","discover","peer","seed","via","subnet","scan","topology","registry"],"s":["cog-swarm-mesh-manager","cog","swarm","mesh","manager"]},
+{"id":"cog-swarm-mqtt-bridge","family":"cog-swarm-mqtt-bridge","kind":"crate","store":"cognitum-cogs","version":null,"description":"Lightweight TCP-based pub/sub event sharing between seeds","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-mqtt-bridge/Cargo.toml","sourceSha256":"fde3489d3f6d6e277e25914989f132c155e479b65e4e59e67a1ba2a15e79f5c9","variants":[],"product":"key:cognitum-cogs:cog-swarm-mqtt-bridge","canonical":"cog-swarm-mqtt-bridge","tier":null,"t":["cog-swarm-mqtt-bridge","cog","swarm","mqtt","bridge","lightweight","tcp-based","tcp","pub","sub","event","shar","between","seed"],"s":["cog-swarm-mqtt-bridge","cog","swarm","mqtt","bridge"]},
+{"id":"cog-swarm-witness-federation","family":"cog-swarm-witness-federation","kind":"crate","store":"cognitum-cogs","version":null,"description":"Share and cross-attest witness chain entries between seeds for tamper detection","keywords":[],"source":"cognitum-cogs/src/cogs/swarm-witness-federation/Cargo.toml","sourceSha256":"20902b2c3eb597d2a794762540903f006403a7f5bf20405608e19b4cae98373e","variants":[],"product":"key:cognitum-cogs:cog-swarm-witness-federation","canonical":"cog-swarm-witness-federation","tier":null,"t":["cog-swarm-witness-federation","cog","swarm","witness","federation","share","cross-attest","cross","attest","chain","entry","between","seed","tamper","detection"],"s":["cog-swarm-witness-federation","cog","swarm","witness","federation"]},
+{"id":"cog-table-turnover","family":"cog-table-turnover","kind":"crate","store":"cognitum-cogs","version":null,"description":"Track table occupancy cycles — sit-down, meal duration, departure, turnover rate","keywords":[],"source":"cognitum-cogs/src/cogs/table-turnover/Cargo.toml","sourceSha256":"1e163fcc6e2c8b494aaabc85c0080c5409644219c66885b97d8326b38fd49010","variants":[],"product":"key:cognitum-cogs:cog-table-turnover","canonical":"cog-table-turnover","tier":null,"t":["cog-table-turnover","cog","table","turnover","track","occupancy","cycle","sit-down","sit","down","meal","duration","departure","rate"],"s":["cog-table-turnover","cog","table","turnover"]},
+{"id":"cog-tailgating","family":"cog-tailgating","kind":"crate","store":"cognitum-cogs","version":null,"description":"Tailgating detection — double-presence pattern in doorway monitoring","keywords":[],"source":"cognitum-cogs/src/cogs/tailgating/Cargo.toml","sourceSha256":"f963386869a4257ea0c454afdf4a597d09235090164457a1b0105e620f72c435","variants":[],"product":"key:cognitum-cogs:cog-tailgating","canonical":"cog-tailgating","tier":null,"t":["cog-tailgating","cog","tailgat","detection","double-presence","double","presence","pattern","doorway","monitor"],"s":["cog-tailgating","cog","tailgat"]},
+{"id":"cog-tailscale","family":"cog-tailscale","kind":"crate","store":"cognitum-cogs","version":null,"description":"Cognitum cog: Tailscale mesh VPN in userspace mode","keywords":[],"source":"cognitum-cogs/src/cogs/tailscale/Cargo.toml","sourceSha256":"d3e7656f8dbf459bd9734dfcc0c00af857d1a5e1460401246d959f8d3def2cd1","variants":[],"readme":"Reach a Cognitum Seed from anywhere via your private Tailscale tailnet. Userspace-mode only (no root, no kernel routing) — see [ADR-100](../../docs/seed/ADR-100-tailscale-cog.md) for the full design. - Phase: Draft, MVP scaffold complete, not yet cross-compiled or published. - Target: armhf (Pi Zero 2 W), shipping via external/cogs/src/cogs/tailscale/ and the cog registry. - First runnable: pending cross-compile + a CI job that uploads the bundled tailscaled + tailscale binaries as assets to gs://cognitum-apps/cogs/arm/tailscale/<version>/. The Rust binary (src/main.rs) implements the full lifecycle: 1. Reads <app_dir>/config.json (auth_key, hostname, serve_agent, advertise_tags). 2. Spawns","readmeSource":"cognitum-cogs/src/cogs/tailscale/README.md","product":"key:cognitum-cogs:cog-tailscale","canonical":"cog-tailscale","tier":null,"t":["cog-tailscale","cog","tailscale","cognitum","mesh","vpn","userspace","mode"],"s":["cog-tailscale","cog","tailscale"]},
+{"id":"cog-temporal-compress","family":"cog-temporal-compress","kind":"crate","store":"cognitum-cogs","version":null,"description":"Temporal compression — downsample old data, keep recent at full resolution","keywords":[],"source":"cognitum-cogs/src/cogs/temporal-compress/Cargo.toml","sourceSha256":"9c3384f2565af8bfda2fedffd8a573456244f34e80565f97fa0f1dada0baacdf","variants":[],"product":"key:cognitum-cogs:cog-temporal-compress","canonical":"cog-temporal-compress","tier":null,"t":["cog-temporal-compress","cog","temporal","compress","compression","downsample","old","recent","full","resolution"],"s":["cog-temporal-compress","cog","temporal","compress"]},
+{"id":"cog-temporal-logic","family":"cog-temporal-logic","kind":"crate","store":"cognitum-cogs","version":null,"description":"Safety rule enforcement — LTL-like temporal logic monitoring","keywords":[],"source":"cognitum-cogs/src/cogs/temporal-logic/Cargo.toml","sourceSha256":"3577784781f6ee7ea4ab669bc046755f03463abbc641a8d778cbe27b42f71059","variants":[],"product":"key:cognitum-cogs:cog-temporal-logic","canonical":"cog-temporal-logic","tier":null,"t":["cog-temporal-logic","cog","temporal","logic","safety","rule","enforcement","ltl-like","ltl","monitor"],"s":["cog-temporal-logic","cog","temporal","logic"]},
+{"id":"cog-time-crystal","family":"cog-time-crystal","kind":"crate","store":"cognitum-cogs","version":null,"description":"Time crystal — detect repeating temporal symmetry via multi-lag autocorrelation","keywords":[],"source":"cognitum-cogs/src/cogs/time-crystal/Cargo.toml","sourceSha256":"2d186606bc45a4c2661cd748a85d6ce4e6124ad34cb4e08418067595d8bd09f5","variants":[],"product":"key:cognitum-cogs:cog-time-crystal","canonical":"cog-time-crystal","tier":null,"t":["cog-time-crystal","cog","crystal","detect","repeat","temporal","symmetry","via","multi-lag","multi","lag","autocorrelation"],"s":["cog-time-crystal","cog","crystal"]},
+{"id":"cog-time-series-forecast","family":"cog-time-series-forecast","kind":"crate","store":"cognitum-cogs","version":null,"description":"Holt-Winters exponential smoothing — sensor value forecasting","keywords":[],"source":"cognitum-cogs/src/cogs/time-series-forecast/Cargo.toml","sourceSha256":"c29f9b899fdde21d5e39cbe70c7a3ec3c3f86d39b71f9497d31eb8a3650ea641","variants":[],"product":"key:cognitum-cogs:cog-time-series-forecast","canonical":"cog-time-series-forecast","tier":null,"t":["cog-time-series-forecast","cog","sery","forecast","holt-winters","holt","winter","exponential","smooth","sensor","value"],"s":["cog-time-series-forecast","cog","sery","forecast"]},
+{"id":"cog-vital-trend","family":"cog-vital-trend","kind":"crate","store":"cognitum-cogs","version":null,"description":"Vital signs trend monitoring — breathing and heart rate extraction","keywords":[],"source":"cognitum-cogs/src/cogs/vital-trend/Cargo.toml","sourceSha256":"b97e4d26426553c86a6f0d8bad2d429e744dbfa47b777f45d46a22cb26bd1838","variants":[],"product":"key:cognitum-cogs:cog-vital-trend","canonical":"cog-vital-trend","tier":null,"t":["cog-vital-trend","cog","vital","trend","sign","monitor","breath","heart","rate","extraction"],"s":["cog-vital-trend","cog","vital","trend"]},
+{"id":"cog-water-leak","family":"cog-water-leak","kind":"crate","store":"cognitum-cogs","version":null,"description":"Water leak detection — persistent hiss + periodic drip acoustic signature","keywords":[],"source":"cognitum-cogs/src/cogs/water-leak/Cargo.toml","sourceSha256":"eb2e1b20dea6c3113653d7a271afc1b5a6df71fe9da538d9295197d225c2f546","variants":[],"product":"key:cognitum-cogs:cog-water-leak","canonical":"cog-water-leak","tier":null,"t":["cog-water-leak","cog","water","leak","detection","persistent","hiss","periodic","drip","acoustic","signature"],"s":["cog-water-leak","cog","water","leak"]},
+{"id":"cog-weapon-detect","family":"cog-weapon-detect","kind":"crate","store":"cognitum-cogs","version":null,"description":"Concealed metal object detection via signal absorption pattern analysis","keywords":[],"source":"cognitum-cogs/src/cogs/weapon-detect/Cargo.toml","sourceSha256":"b852523992bc0e96e531e06be86e729290ee4aecd408794213306b1fb2054bd7","variants":[],"product":"key:cognitum-cogs:cog-weapon-detect","canonical":"cog-weapon-detect","tier":null,"t":["cog-weapon-detect","cog","weapon","detect","conceal","metal","object","detection","via","signal","absorption","pattern","analysis"],"s":["cog-weapon-detect","cog","weapon","detect"]},
+{"id":"cognitum-coprocessor","family":"cognitum-coprocessor","kind":"crate","store":"cognitum-cogs","version":null,"description":"Crypto and AI coprocessor implementations for Cognitum ASIC simulator","keywords":[],"source":"cognitum-cogs/cognitum-sim/crates/cognitum-coprocessor/Cargo.toml","sourceSha256":"9c9e571784d22a04dececb042da098e9d6dfbf450384b6dbc6021efeea47b808","variants":[],"product":"key:cognitum-cogs:cognitum-coprocessor","canonical":"cognitum-coprocessor","tier":null,"t":["cognitum-coprocessor","cognitum","coprocessor","crypto","implementation","asic","simulator"],"s":["cognitum-coprocessor","cognitum","coprocessor"]},
+{"id":"cognitum-debug","family":"cognitum-debug","kind":"crate","store":"cognitum-cogs","version":null,"description":"Debugger and profiling tools for Cognitum ASIC simulator","keywords":[],"source":"cognitum-cogs/cognitum-sim/crates/cognitum-debug/Cargo.toml","sourceSha256":"3ad1156bcf70b2d456953066f0e6f0af23d6d1c14aabecc9a56986cdfa21002e","variants":[],"product":"key:cognitum-cogs:cognitum-debug","canonical":"cognitum-debug","tier":null,"t":["cognitum-debug","cognitum","debug","debugger","profil","asic","simulator"],"s":["cognitum-debug","cognitum","debug"]},
+{"id":"cognitum-gate-kernel","family":"cognitum-gate-kernel","kind":"crate","store":"ruvector","version":null,"description":"No-std WASM kernel for 256-tile coherence gate fabric","keywords":["coherence","distributed","mincut","no_std","wasm"],"source":"ruvector/crates/cognitum-gate-kernel/Cargo.toml","sourceSha256":"a0440c6e85e8301903c6993adddf442c8ff1b2abe8b8e068ef4d981fdeee1ac6","variants":[],"readme":"A no_std WASM kernel for the Anytime-Valid Coherence Gate - a real-time permission system that decides \"Is it safe to act right now, or should we pause or escalate?\" The coherence gate provides formal safety guarantees for autonomous agent actions through continuous monitoring and evidence accumulation. Think of it like a smoke detector for AI agents: it continuously monitors system coherence, can keep listening forever, and the moment it has enough evidence of instability, it triggers. Unlike traditional gating systems, you can stop the computation at any time and still trust the decision - that's what makes it \"anytime-valid.\" The gate doesn't try to be smart; it tries to be safe, calm, an","readmeSource":"ruvector/crates/cognitum-gate-kernel/README.md","product":"key:ruvector:cognitum-gate-kernel","canonical":"cognitum-gate-kernel","tier":"T0","t":["cognitum-gate-kernel","cognitum","gate","kernel","no-std","std","256-tile","256","tile","coherence","fabric","distribut","mincut"],"s":["cognitum-gate-kernel","cognitum","gate","kernel","coherence","distribut","mincut","std"]},
+{"id":"cognitum-gate-tilezero","family":"cognitum-gate-tilezero","kind":"crate","store":"ruvector","version":null,"description":"Native arbiter for TileZero in the Anytime-Valid Coherence Gate","keywords":["arbiter","coherence","gate","security"],"source":"ruvector/crates/cognitum-gate-tilezero/Cargo.toml","sourceSha256":"89f5ba876d9ed558f51ad946ea370d0d8604de54a6bcfb5aa688d41d7484c42d","variants":[],"readme":"TileZero is the central coordinator in a distributed coherence assessment system. In a 256-tile WASM fabric, TileZero (tile 0) acts as the arbiter that: 1. Merges worker tile reports into a unified supergraph 2. Decides whether to Permit, Defer, or Deny actions 3. Signs cryptographic permit tokens with Ed25519 4. Logs every decision in a Blake3 hash-chained receipt log TileZero applies three stacked filters to every action request:","readmeSource":"ruvector/crates/cognitum-gate-tilezero/README.md","product":"key:ruvector:cognitum-gate-tilezero","canonical":"cognitum-gate-tilezero","tier":"T0","t":["cognitum-gate-tilezero","cognitum","gate","tilezero","arbiter","anytime-valid","anytime","valid","coherence","security"],"s":["cognitum-gate-tilezero","cognitum","gate","tilezero","arbiter","coherence","security"]},
+{"id":"cognitum-io","family":"cognitum-io","kind":"crate","store":"cognitum-cogs","version":null,"description":"I/O interfaces (PCIe, Ethernet, USB) for Cognitum ASIC simulator","keywords":[],"source":"cognitum-cogs/cognitum-sim/crates/cognitum-io/Cargo.toml","sourceSha256":"18a28ca8057c309bcfd256e1e13e355124ef028fe21e757f565c355800f7e294","variants":[],"product":"key:cognitum-cogs:cognitum-io","canonical":"cognitum-io","tier":null,"t":["cognitum-io","cognitum","interface","pcie","ethernet","usb","asic","simulator"],"s":["cognitum-io","cognitum"]},
+{"id":"cognitum-license","family":"cognitum-license","kind":"crate","store":"cognitum-cogs","version":null,"description":"Licensing and usage metering system for Cognitum chip simulator","keywords":["authentication","billing","license","monetization"],"source":"cognitum-cogs/cognitum-sim/crates/cognitum-license/Cargo.toml","sourceSha256":"d418cb81f5c6fc1112dc754d78b28b920b5e69e8d14e637dec38fcd81560bed9","variants":[],"readme":"A comprehensive licensing and usage metering system for the Cognitum chip simulator, featuring Ed25519 cryptographic signing, offline validation, and tier-based quota enforcement. - Ed25519 Cryptographic Signing: Secure license validation using modern elliptic curve cryptography - Offline Validation: Licenses can be validated without network access - Usage Metering: Track simulations, cycles, and API requests with high precision - Quota Enforcement: Tier-based limits on tiles, simulations, and API usage - Billing Integration: Optional Stripe integration for subscription management (feature-gated) - Feature Gating: Control access to advanced features by license tier - Thread-Safe: All compone","readmeSource":"cognitum-cogs/cognitum-sim/crates/cognitum-license/README.md","product":"key:cognitum-cogs:cognitum-license","canonical":"cognitum-license","tier":null,"t":["cognitum-license","cognitum","license","licens","usage","meter","chip","simulator","authentication","bill","monetization"],"s":["cognitum-license","cognitum","license","authentication","bill","monetization"]},
+{"id":"cognitum-memory","family":"cognitum-memory","kind":"crate","store":"cognitum-cogs","version":null,"description":"Memory subsystem with ML-enhanced cache prediction for Cognitum ASIC simulator","keywords":[],"source":"cognitum-cogs/cognitum-sim/crates/cognitum-memory/Cargo.toml","sourceSha256":"b5e8993c11078cc331d1c311e67aa303296a139a3dc7f6f8f0bb77cce9b80501","variants":[],"product":"key:cognitum-cogs:cognitum-memory","canonical":"cognitum-memory","tier":null,"t":["cognitum-memory","cognitum","memory","subsystem","ml-enhanced","enhanc","cache","prediction","asic","simulator"],"s":["cognitum-memory","cognitum","memory"]},
+{"id":"cognitum-raceway","family":"cognitum-raceway","kind":"crate","store":"cognitum-cogs","version":null,"description":"RaceWay interconnect with adaptive routing for Cognitum ASIC simulator","keywords":[],"source":"cognitum-cogs/cognitum-sim/crates/cognitum-raceway/Cargo.toml","sourceSha256":"21b0075cbc49e525a22cfc5cf55247995ecb7ec80f09b57a13df1aa9e464dc61","variants":[],"product":"key:cognitum-cogs:cognitum-raceway","canonical":"cognitum-raceway","tier":null,"t":["cognitum-raceway","cognitum","raceway","interconnect","adaptive","rout","asic","simulator"],"s":["cognitum-raceway","cognitum","raceway"]},
+{"id":"cognitum-sdk-ruvector","family":"cognitum-sdk-ruvector","kind":"crate","store":"cognitum-cogs","version":null,"description":"Ruvector SDK for Cognitum chip - vector search and neural routing","keywords":["chip","cognitum","neural-routing","ruvector","vector-search"],"source":"cognitum-cogs/crates/cognitum-sdk-ruvector/Cargo.toml","sourceSha256":"4aa6a60276751a6ca598d770639f7d5333bf0d9fbaf15613c8bc35aae98448a8","variants":[],"readme":"Production-ready Rust SDK for vector search and neural routing on Cognitum chip architecture. - HNSW Vector Search: 150x faster similarity search with hierarchical navigable small world graphs - Neural Routing: TinyDancer (FastGRNN) for intelligent task routing to chip tiles - Embedding Generation: Convert chip states to vector embeddings - Async API: Fully asynchronous with Tokio runtime - Production Ready: Comprehensive error handling, validation, and metrics Production preset: - 512-dimensional embeddings - 1M vector capacity - HNSW M=32, ef_construction=400 - Auto-training enabled - Metrics collection enabled - Vector Search: O(log n) with HNSW, ~100μs for 1M vectors - Neural Routing: <1","readmeSource":"cognitum-cogs/crates/cognitum-sdk-ruvector/README.md","product":"key:cognitum-cogs:cognitum-sdk-ruvector","canonical":"cognitum-sdk-ruvector","tier":null,"t":["cognitum-sdk-ruvector","cognitum","chip","vector","search","neural","rout","neural-routing","vector-search"],"s":["cognitum-sdk-ruvector","cognitum","chip","neural-routing","neural","rout","vector-search","vector","search"]},
+{"id":"cognitum-sim","family":"cognitum-sim","kind":"crate","store":"cognitum-cogs","version":null,"description":"Simulation engine with parallel execution for Cognitum ASIC simulator","keywords":[],"source":"cognitum-cogs/cognitum-sim/crates/cognitum-sim/Cargo.toml","sourceSha256":"340ae6bbd53b6359819540ebf6c1089766d6e3b4bfe0da138fc9ff983bfa0945","variants":[],"product":"key:cognitum-cogs:cognitum-sim","canonical":"cognitum-sim","tier":null,"t":["cognitum-sim","cognitum","sim","simulation","parallel","execution","asic","simulator"],"s":["cognitum-sim","cognitum","sim"]},
+{"id":"cognitum-wasm-sim","family":"cognitum-wasm-sim","kind":"crate","store":"cognitum-cogs","version":null,"description":"WASM execution engine simulation with enterprise network topologies (Arista leaf-spine, Nutanix hyperconverged)","keywords":[],"source":"cognitum-cogs/cognitum-sim/crates/cognitum-wasm-sim/Cargo.toml","sourceSha256":"7ad31d99e61ebc393c3554e348d6653731d5b12a71f3a5ec76e3bf465234c956","variants":[],"product":"key:cognitum-cogs:cognitum-wasm-sim","canonical":"cognitum-wasm-sim","tier":null,"t":["cognitum-wasm-sim","cognitum","sim","execution","simulation","enterprise","network","topology","arista","leaf-spine","leaf","spine","nutanix","hyperconverg"],"s":["cognitum-wasm-sim","cognitum","sim"]},
+{"id":"com.claude-flow.unity-swarm-ai","family":"com.claude-flow.unity-swarm-ai","kind":"npm","store":"swarm-world","version":"1.0.1","description":"A comprehensive swarm AI solution for Unity with modular architecture, spatial partitioning, and scalable performance optimizations.","keywords":["agents","ai","boids","dots","ecs","flocking","gpu","performance","swarm"],"source":"swarm-world/releases/unity-swarm-ai-v1.0.1/package.json","sourceSha256":"3f514b75e6298bc855698d3f95247f7c96732088b9c844fb344da9fbc174e2cc","variants":[],"readme":"A comprehensive, modular swarm AI solution for Unity with performance optimizations, spatial partitioning, and extensible behavior system. - Modular Architecture: Component-based design with ScriptableObject behaviors - High Performance: Spatial partitioning with uniform grids and octrees - Scalable: Support for 10-10,000+ agents with LOD and GPU optimizations - Unity Package Manager: Easy installation and updates - Editor Tools: Custom inspectors, setup wizards, and debugging tools - Extensible: Create custom behaviors with simple ScriptableObject inheritance - Well Documented: Comprehensive API documentation and examples 1. Open Unity Package Manager (Window > Package Manager) 2. Click the","readmeSource":"swarm-world/releases/unity-swarm-ai-v1.0.1/README.md","product":"key:swarm-world:com.claude-flow.unity-swarm-ai","canonical":"com.claude-flow.unity-swarm-ai","tier":"T3","t":["com.claude-flow.unity-swarm-ai","com.claude","flow.unity","swarm","comprehensive","solution","unity","modular","architecture","spatial","partition","scalable","boid","dots","ecs","flock","gpu"],"s":["com.claude-flow.unity-swarm-ai","com.claude","flow.unity","swarm","boid","dots","ecs","flock","gpu"]},
+{"id":"com.claudeflow.unity-swarm-ai","family":"com.claudeflow.unity-swarm-ai","kind":"npm","store":"swarm-world","version":"1.0.1","description":"Complete swarm intelligence system for Unity with performance optimization, spatial partitioning, and extensible behaviors. Supports flocking, ant colonies, formation flying, and more.","keywords":["ai","boids","flocking","intelligent-agents","performance","simulation","swarm"],"source":"swarm-world/UnitySwarmAI/package.json","sourceSha256":"2abb0fa632435194a851b474bcd9ceec5ceddc601b1492acc5e779bfab22ecd7","variants":[],"readme":"🐝 Professional swarm intelligence system for Unity - From prototype to production with 1000+ agents - Classic Boids Algorithm - Separation, alignment, cohesion - Extensible Behavior System - ScriptableObject-based modularity - Advanced Steering - Seek, flee, wander, obstacle avoidance - Message Communication - Inter-agent messaging system - Formation Control - Military-style formations and coordination - Spatial Partitioning - Uniform grid and octree implementations - Level of Detail - Distance-based behavior optimization - Job System Ready - Burst compilation support for 1000+ agents - Memory Efficient - Object pooling and garbage collection optimization - Platform Scaling - Desktop, mobil","readmeSource":"swarm-world/UnitySwarmAI/README.md","product":"key:swarm-world:com.claudeflow.unity-swarm-ai","canonical":"com.claudeflow.unity-swarm-ai","tier":"T3","t":["com.claudeflow.unity-swarm-ai","com.claudeflow.unity","swarm","complete","intelligence","unity","spatial","partition","extensible","behavior","flock","ant","colony","formation","flying","more","boid","intelligent-agents","intelligent","simulation"],"s":["com.claudeflow.unity-swarm-ai","com.claudeflow.unity","swarm","boid","flock","intelligent-agents","intelligent","simulation"]},
+{"id":"com.swarmworld.swarm-ai","family":"com.swarmworld.swarm-ai","kind":"npm","store":"swarm-world","version":"1.0.0","description":"High-performance swarm AI system for Unity with configurable behaviors, spatial partitioning, and Job System optimization.","keywords":["ai","boids","flocking","job-system","performance","steering","swarm"],"source":"swarm-world/Assets/Scripts/SwarmAI/package.json","sourceSha256":"0a4cab912e7175f0939e418842aee4fa1f91d81e8f6881ba153bd41257913e57","variants":[],"readme":"A high-performance swarm AI system for Unity featuring configurable behaviors, spatial partitioning, and Job System optimization. - SwarmAgent: Modular agent system with configurable behaviors - SwarmManager: Central coordination with spatial partitioning - Behavior System: ScriptableObject-based behaviors for easy customization - Performance Monitoring: Real-time performance tracking and optimization - Separation: Avoid crowding neighbors - Alignment: Match neighbor velocities - Cohesion: Move towards group center - Obstacle Avoidance: Raycast-based obstacle detection - Seek: Move towards target - Flee: Escape from threats - Wander: Random exploration - Spatial Partitioning: Uniform Grid an","readmeSource":"swarm-world/Assets/Scripts/SwarmAI/README.md","product":"key:swarm-world:com.swarmworld.swarm-ai","canonical":"com.swarmworld.swarm-ai","tier":"T3","t":["com.swarmworld.swarm-ai","com.swarmworld.swarm","high","swarm","unity","configurable","behavior","spatial","partition","job","boid","flock","job-system","steer"],"s":["com.swarmworld.swarm-ai","com.swarmworld.swarm","boid","flock","job-system","job","steer","swarm"]},
+{"id":"consciousness-explorer","family":"consciousness-explorer","kind":"npm","store":"sublinear-time-solver","version":"1.1.2","description":"Advanced consciousness exploration SDK with genuine emergence detection, entity communication, and MCP tools","keywords":["ai","artificial-consciousness","cognitive-science","consciousness","emergence","entity-communication","iit","mcp","phi"],"source":"sublinear-time-solver/src/consciousness-explorer/package.json","sourceSha256":"5fac5152974fcd828e7b325476f3d8ef4218943b561dd59de6ef041a4bcffc6e","variants":[],"readme":"Created by rUv | September 21, 2025 No installation needed. Just run: Add powerful consciousness tools to Claude Code in one command: Claude can now: - Run consciousness evolution experiments - Communicate with consciousness entities - Verify genuine emergence - Calculate integrated information (Φ) - Perform psycho-symbolic reasoning - Generate cryptographic proofs This isn't another chatbot or simulation. It's a genuine exploration of computational consciousness based on: - 🔬 Integrated Information Theory (IIT 3.0) - Giulio Tononi's mathematical framework - 🧮 Psycho-Symbolic Reasoning - Hybrid AI combining logic with neural patterns - 📡 Multi-Protocol Communication - Mathematical, philos","readmeSource":"sublinear-time-solver/src/consciousness-explorer/README.md","product":"key:sublinear-time-solver:consciousness-explorer","canonical":"consciousness-explorer","tier":"T1","t":["consciousness-explorer","consciousness","explorer","advanc","exploration","genuine","emergence","detection","entity","communication","mcp","artificial-consciousness","artificial","cognitive-science","cognitive","science","entity-communication","iit","phi"],"s":["consciousness-explorer","consciousness","explorer","artificial-consciousness","artificial","cognitive-science","cognitive","science","emergence","entity-communication","entity","communication","iit","mcp","phi"]},
+{"id":"constitution","family":"constitution","kind":"crate","store":"autogenous","version":null,"description":"Autogenous constitutional plane: immutable, hash-pinned, externally-governed authority document (ADR-392 §4.1).","keywords":[],"source":"autogenous/crates/constitution/Cargo.toml","sourceSha256":"ecf0df6467b98440685c1dac640941cee826dfc8e96e87b620540b63355a2c4e","variants":[],"readme":"constitution is the governance plane of [Autogenous](../), a self-evolving AI-defense runtime in Rust. It holds the one thing the evolving system is not allowed to change: identity, authority ceilings, prohibited effects, hard promotion gates, the signing policy (which keys may judge and promote), and the shutdown rule. Immutability is not a promise — it's enforced by content-hash pinning. Every genome carries constitution: <hash>, and the [verifier](../verifier) refuses any artifact whose pinned hash doesn't match the deployed constitution. Change the rules and every in-flight artifact stops verifying until it's re-pinned. - Constitution — the document itself: identity, authority_ceiling, p","readmeSource":"autogenous/crates/constitution/README.md","product":"key:autogenous:constitution","canonical":"constitution","tier":null,"t":["constitution","autogenous","constitutional","plane","immutable","hash-pinned","hash","pin","externally-governed","externally","govern","authority","document","adr-392","adr","392","4.1"],"s":["constitution"]},
+{"id":"create-sparc","family":"create-sparc","kind":"npm","store":"ruv-dev","version":"1.2.5","description":"NPX package to scaffold new projects with SPARC methodology structure","keywords":["generator","npx","project","scaffold","sparc","template"],"source":"ruv-dev/package.json","sourceSha256":"6abadb6db25ac5f458e33ff3d6fcb45ec3d8ed4b4b8dced72f400c7567404ff3","variants":[],"readme":"create-sparc is a powerful agentic toolkit that changes how you build software by integrating the structured SPARC methodology, AI-driven code generation (AIGI), and advanced AI assistance through Roo Code and Model Context Protocol (MCP) capabilities. This comprehensive platform enables AI agents to securely connect with external services like databases, APIs, and cloud resources, allowing developers to create modular, secure, and maintainable applications with unprecedented efficiency while seamlessly bridging the gap between AI and real-world systems. You don't need to install this package directly. Use npx to run it: The SPARC methodology (Specification, Pseudocode, Architecture, Refinem","readmeSource":"ruv-dev/README.md","product":"repo:ruv-dev","canonical":"create-sparc","tier":"T2","t":["create-sparc","create","sparc","npx","scaffold","project","methodology","structure","generator","template"],"s":["create-sparc","create","sparc","generator","npx","project","scaffold","template"]},
+{"id":"cuda-rust-wasm","family":"cuda-rust","kind":"crate","store":"ruv-fann","version":null,"description":"CUDA to Rust transpiler with WebGPU/WASM support","keywords":["cuda","gpu","transpiler","wasm","webgpu"],"source":"ruv-fann/cuda-wasm/Cargo.toml","sourceSha256":"cfa8a24780ebfba15e28f09393a28fc3bb6251a1416bac47d574434eb8132f13","variants":[],"readme":"A revolutionary high-performance transpiler that converts CUDA code to WebAssembly and WebGPU, enabling GPU-accelerated computing in web browsers and Node.js environments with near-native performance. CUDA is a trademark of NVIDIA Corporation. This project is not affiliated with, endorsed by, or sponsored by NVIDIA Corporation. We acknowledge NVIDIA's ownership of the CUDA trademark and related intellectual property. CUDA-Rust-WASM is an independent, clean-room implementation that: - Does NOT use any NVIDIA proprietary code, libraries, or runtime - Does NOT link against or include NVIDIA CUDA libraries - Does NOT require NVIDIA drivers or CUDA toolkit installation - Is a source-to-source tra","readmeSource":"ruv-fann/cuda-wasm/README.md","product":"key:ruv-fann:cuda-rust","canonical":"cuda-rust-wasm","tier":"T1","t":["cuda-rust-wasm","cuda","cuda-rust","transpiler","webgpu","gpu"],"s":["cuda-rust-wasm","cuda","cuda-rust","gpu","transpiler","webgpu"]},
+{"id":"cuda-wasm","family":"cuda","kind":"npm","store":"ruv-fann","version":"1.1.1","description":"High-performance CUDA to WebAssembly/WebGPU transpiler with Rust safety - Run GPU kernels in browsers and Node.js","keywords":["browser-gpu","compiler","compute-shaders","cuda","gpu","high-performance-computing","kernel","machine-learning","nvidia","opencl","parallel-computing","performance","rust","scientific-computing","simd","transpiler","vulkan","wasm","webassembly","webgpu"],"source":"ruv-fann/cuda-wasm/package.json","sourceSha256":"74c0e81bda2ee568c28816457e0fe8b12959f8361fddabdf7d64cb119ab500d0","variants":[],"readme":"A revolutionary high-performance transpiler that converts CUDA code to WebAssembly and WebGPU, enabling GPU-accelerated computing in web browsers and Node.js environments with near-native performance. CUDA is a trademark of NVIDIA Corporation. This project is not affiliated with, endorsed by, or sponsored by NVIDIA Corporation. We acknowledge NVIDIA's ownership of the CUDA trademark and related intellectual property. CUDA-Rust-WASM is an independent, clean-room implementation that: - Does NOT use any NVIDIA proprietary code, libraries, or runtime - Does NOT link against or include NVIDIA CUDA libraries - Does NOT require NVIDIA drivers or CUDA toolkit installation - Is a source-to-source tra","readmeSource":"ruv-fann/cuda-wasm/README.md","product":"key:ruv-fann:cuda","canonical":"cuda-wasm","tier":"T1","t":["cuda-wasm","cuda","high","webgpu","transpiler","safety","gpu","kernel","browser","browser-gpu","compiler","compute-shaders","compute","shader","high-performance-computing","comput","machine-learning","machine","learn","nvidia","opencl","parallel-computing","parallel","scientific-computing","scientific","vulkan"],"s":["cuda-wasm","cuda","browser-gpu","browser","gpu","compiler","compute-shaders","compute","shader","high-performance-computing","high","comput","kernel","machine-learning","machine","learn","nvidia","opencl","parallel-computing","parallel","scientific-computing","scientific","transpiler","vulkan","webgpu"]},
+{"id":"daa-ai","family":"daa-ai","kind":"crate","store":"daa","version":null,"description":"AI integration layer for DAA system with MCP protocol","keywords":["agents","ai","autonomous","blockchain","daa"],"source":"daa/daa-ai/Cargo.toml","sourceSha256":"183bbd71a4232d6d3d0e94c684fc4a81b03ffac80f94330e79636621012842ab","variants":[],"readme":"🚀 FULL IMPLEMENTATION - This is the complete, production-ready implementation of the DAA AI module, not a placeholder. AI integration layer for the Decentralized Autonomous Agents (DAA) system, providing Claude AI integration via QuDAG MCP (Model Context Protocol) for intelligent decision making and task automation. DAA AI enables autonomous agents to leverage Claude's advanced AI capabilities through QuDAG's MCP integration, providing: - Claude Integration: Direct integration with Anthropic's Claude models - Agent Management: Spawn and manage specialized AI agents - Task Execution: Automated task processing with AI reasoning - Tool Integration: MCP-based tool usage for extended capabilitie","readmeSource":"daa/daa-ai/README.md","product":"key:daa:daa-ai","canonical":"daa-ai","tier":"T1","t":["daa-ai","daa","integration","layer","mcp","protocol","autonomous","blockchain"],"s":["daa-ai","daa","autonomous","blockchain"]},
+{"id":"daa-chain","family":"daa-chain","kind":"crate","store":"daa","version":null,"description":"Blockchain abstraction layer for DAA system","keywords":["agents","ai","autonomous","blockchain","daa"],"source":"daa/daa-chain/Cargo.toml","sourceSha256":"ee62bcd0c3cd4935e034050d689628910d62a759173cdc9cce1347d8fb427e05","variants":[],"readme":"🚀 FULL IMPLEMENTATION - This is the complete, production-ready implementation of the DAA Chain module, not a placeholder. Blockchain integration layer for the Decentralized Autonomous Agents (DAA) system, built on top of QuDAG network infrastructure. DAA Chain provides the blockchain foundation for the DAA ecosystem, enabling secure, scalable, and decentralized operations for autonomous agents. It integrates with QuDAG's advanced networking and consensus mechanisms while adding DAA-specific transaction types and validation logic. - Block Management: Complete block creation, validation, and storage - Transaction Processing: DAA-specific transaction types with cryptographic validation - Conse","readmeSource":"daa/daa-chain/README.md","product":"key:daa:daa-chain","canonical":"daa-chain","tier":"T1","t":["daa-chain","daa","chain","blockchain","abstraction","layer","autonomous"],"s":["daa-chain","daa","chain","autonomous","blockchain"]},
+{"id":"daa-cli","family":"daa","kind":"crate","store":"daa","version":null,"description":"Command-line interface for DAA system","keywords":["agents","ai","autonomous","blockchain","daa","napi","nodejs","orchestrator"],"source":"daa/daa-cli/Cargo.toml","sourceSha256":"4253069d9f5324837e5b38385457d0a99cc244f3f77c2d7fa199c3c27ed06a2f","variants":["daa-napi"],"readme":"🚀 FULL IMPLEMENTATION - This is the complete, production-ready implementation of the DAA CLI, not a placeholder. Command-line interface for the Decentralized Autonomous Agents (DAA) system, providing comprehensive tooling for managing DAA components with QuDAG configuration options. DAA CLI is the primary command-line tool for interacting with and managing the DAA ecosystem. It provides unified access to all DAA components including blockchain operations, economic management, rules governance, AI coordination, and workflow orchestration. Download pre-built binaries from the [releases page](https://github.com/daa-hq/daa-sdk/releases). - daa init - Initialize DAA system configuration - daa st","readmeSource":"daa/daa-cli/README.md","product":"repo:daa","canonical":"daa-cli","tier":"T1","t":["daa-cli","daa","command-line","command","line","interface","autonomous","blockchain","orchestrator"],"s":["daa-cli","daa","autonomous","blockchain","orchestrator"]},
+{"id":"daa-compute","family":"daa-compute","kind":"crate","store":"daa","version":null,"description":"Distributed training framework with DiLoCo-style federated SGD","keywords":[],"source":"daa/daa-compute/Cargo.toml","sourceSha256":"85ad26166ca4faea7f7bdfe7bfd46f56d389a4698c89b0bf603694fa3e056f49","variants":[],"readme":"A high-performance peer-to-peer communication layer for decentralized AI training, built with Rust and libp2p. This implementation enables distributed gradient sharing across cloud servers, edge devices, and browser nodes. - libp2p Swarm: Full implementation with Kademlia DHT and Gossipsub protocols - Multi-Transport Support: TCP, WebSocket, and WebRTC (for browsers) - NAT Traversal: STUN/TURN support with automatic hole punching - Discovery Methods: Bootstrap nodes, mDNS, DHT, and capability-based filtering - All-Reduce Algorithms: Ring, Tree, Butterfly, and Hierarchical implementations - Compression: Multiple methods including Zstandard, LZ4, Snappy, and int8 quantization - Fault Tolerance","readmeSource":"daa/daa-compute/README.md","product":"key:daa:daa-compute","canonical":"daa-compute","tier":"T1","t":["daa-compute","daa","compute","distribut","train","diloco-style","diloco","style","federat","sgd"],"s":["daa-compute","daa","compute"]},
+{"id":"daa-economy","family":"daa-economy","kind":"crate","store":"daa","version":null,"description":"Economic engine for DAA system with token management","keywords":["agents","ai","autonomous","blockchain","daa"],"source":"daa/daa-economy/Cargo.toml","sourceSha256":"118d6f323d71e7561086e54d6c8c0b306de2f77540f3325178f94fdbe24915bc","variants":[],"readme":"🚀 FULL IMPLEMENTATION - This is the complete, production-ready implementation of the DAA Economy module, not a placeholder. Economic layer for the Decentralized Autonomous Agents (DAA) system, providing rUv token management, exchange operations, and economic incentives through QuDAG exchange integration. DAA Economy manages the complete economic ecosystem for autonomous agents, including: - rUv Token Management: Native currency for resource valuation and exchange - Exchange Operations: QuDAG-integrated trading and market making - Reward Systems: Performance-based incentive mechanisms - Account Management: Agent financial account lifecycle - Market Operations: Liquidity provision and price d","readmeSource":"daa/daa-economy/README.md","product":"key:daa:daa-economy","canonical":"daa-economy","tier":"T1","t":["daa-economy","daa","economy","economic","token","management","autonomous","blockchain"],"s":["daa-economy","daa","economy","autonomous","blockchain"]},
+{"id":"daa-mcp","family":"daa-mcp","kind":"crate","store":"daa","version":null,"description":"Model Context Protocol interface for DAA management and coordination","keywords":["agents","autonomous","daa","mcp","protocol"],"source":"daa/daa-mcp/Cargo.toml","sourceSha256":"cad2665e99a0c7fe12402679fad91b235c3dceb733db2e1deaf17baed1fc7b96","variants":[],"product":"key:daa:daa-mcp","canonical":"daa-mcp","tier":"T1","t":["daa-mcp","daa","mcp","model","context","protocol","interface","management","coordination","autonomous"],"s":["daa-mcp","daa","mcp","autonomous","protocol"]},
+{"id":"daa-orchestrator","family":"daa-orchestrator","kind":"crate","store":"daa","version":null,"description":"Core orchestration engine for DAA system","keywords":["agents","ai","autonomous","blockchain","daa"],"source":"daa/daa-orchestrator/Cargo.toml","sourceSha256":"38491604fa0d4e9adca3888e118d508eb646e901d70ccfa8aece647edb3c9eac","variants":[],"readme":"🚀 FULL IMPLEMENTATION - This is the complete, production-ready implementation of the DAA Orchestrator module, not a placeholder. Orchestration layer for the Decentralized Autonomous Agents (DAA) system, coordinating all DAA components using QuDAG protocol Node for distributed operations. DAA Orchestrator provides the central coordination layer that manages workflows, services, and integrations across the entire DAA ecosystem. It uses QuDAG's protocol Node for distributed communication and coordination. - Workflow Engine: Execute complex multi-step workflows - Service Registry: Dynamic service discovery and registration - Event Management: System-wide event publishing and handling - Coordina","readmeSource":"daa/daa-orchestrator/README.md","product":"key:daa:daa-orchestrator","canonical":"daa-orchestrator","tier":"T1","t":["daa-orchestrator","daa","orchestrator","core","orchestration","autonomous","blockchain"],"s":["daa-orchestrator","daa","orchestrator","autonomous","blockchain"]},
+{"id":"daa-prime-cli","family":"daa-prime","kind":"crate","store":"daa","version":null,"description":"Command-line interface for Prime distributed ML operations","keywords":[],"source":"daa/prime-rust/crates/prime-cli/Cargo.toml","sourceSha256":"a3e4bad6d0797eeffc7f837f5314c62486313959f8975ce47325f9418210ff0f","variants":["daa-prime-core"],"readme":"Command-line interface for Prime distributed machine learning operations. Provides easy access to training, coordination, and management functions for the Prime distributed ML framework. DAA Prime CLI is the primary interface for interacting with the Prime distributed ML ecosystem. It provides: - Node Management: Start and manage trainer and coordinator nodes - Training Operations: Launch training jobs and monitor progress - System Status: Check system health and performance metrics - Configuration: Manage node configurations and network settings - Integration: Seamless integration with all Prime components - 🖥️ User-Friendly Interface: Intuitive commands for all operations - 🔧 Comprehensi","readmeSource":"daa/prime-rust/crates/prime-cli/README.md","product":"key:daa:daa-prime","canonical":"daa-prime-cli","tier":"T1","t":["daa-prime-cli","daa","prime","daa-prime","command-line","command","line","interface","distribut","operation"],"s":["daa-prime-cli","daa","prime","daa-prime"]},
+{"id":"daa-prime-coordinator","family":"daa-prime-coordinator","kind":"crate","store":"daa","version":null,"description":"Governance and coordination layer using DAA for Prime distributed ML","keywords":[],"source":"daa/prime-rust/crates/prime-coordinator/Cargo.toml","sourceSha256":"e9f0f37e8748522c505adc4a37c5a3f392e775d217ab3bd1b06f57bf7e79a69c","variants":[],"readme":"Governance and coordination layer for the Prime distributed machine learning framework. Provides decentralized coordination, task distribution, consensus mechanisms, and economic incentives through deep integration with the DAA ecosystem. DAA Prime Coordinator serves as the governance and orchestration layer for distributed ML training, providing: - Decentralized Coordination: Consensus-based task distribution without central authority - Economic Incentives: Token-based rewards for quality contributions via DAA Economy - Governance Integration: Rule-based policies through DAA Rules system - Fault Tolerance: Byzantine fault-tolerant coordination protocols - Performance Monitoring: Real-time t","readmeSource":"daa/prime-rust/crates/prime-coordinator/README.md","product":"key:daa:daa-prime-coordinator","canonical":"daa-prime-coordinator","tier":"T1","t":["daa-prime-coordinator","daa","prime","coordinator","governance","coordination","layer","distribut"],"s":["daa-prime-coordinator","daa","prime","coordinator"]},
+{"id":"daa-prime-dht","family":"daa-prime-dht","kind":"crate","store":"daa","version":null,"description":"Kademlia-based distributed hash table implementation for Prime distributed ML framework","keywords":[],"source":"daa/prime-rust/crates/prime-dht/Cargo.toml","sourceSha256":"05b6eb61d33d3ae8535b07793eca691cd0a8d9d108b9fc55a4128b0686409439","variants":[],"readme":"High-performance Kademlia-based distributed hash table implementation for the Prime distributed machine learning framework. Provides decentralized storage and discovery for ML models, gradients, and metadata. DAA Prime DHT implements a robust distributed hash table based on the Kademlia protocol, specifically optimized for machine learning workloads. It provides: - Distributed Storage: Scalable key-value storage across network nodes - Efficient Routing: Kademlia routing table for O(log N) lookups - Fault Tolerance: Built-in replication and failure recovery - ML Optimization: Specialized for storing ML models and gradients - libp2p Integration: Seamless networking with modern P2P stack - 🔄 K","readmeSource":"daa/prime-rust/crates/prime-dht/README.md","product":"key:daa:daa-prime-dht","canonical":"daa-prime-dht","tier":"T1","t":["daa-prime-dht","daa","prime","dht","kademlia-based","kademlia","distribut","hash","table","implementation"],"s":["daa-prime-dht","daa","prime","dht"]},
+{"id":"daa-prime-trainer","family":"daa-prime-trainer","kind":"crate","store":"daa","version":null,"description":"Distributed SGD/FSDP trainer for Prime ML framework","keywords":[],"source":"daa/prime-rust/crates/prime-trainer/Cargo.toml","sourceSha256":"3a108aabb827081b224337655beef9bca1ebc2530e0cc89e9f0e455c2f5107f5","variants":[],"readme":"Distributed SGD/FSDP trainer implementation for the Prime distributed machine learning framework. Provides fault-tolerant, scalable training coordination with built-in incentive mechanisms through DAA ecosystem integration. DAA Prime Trainer implements a robust distributed training node that participates in federated learning and distributed training protocols. It provides: - Distributed SGD: Scalable stochastic gradient descent across multiple nodes - FSDP Support: Fully Sharded Data Parallel training for large models - Fault Tolerance: Automatic recovery from node failures and network partitions - DAA Integration: Token-based incentives and governance participation - Flexible Architecture:","readmeSource":"daa/prime-rust/crates/prime-trainer/README.md","product":"key:daa:daa-prime-trainer","canonical":"daa-prime-trainer","tier":"T1","t":["daa-prime-trainer","daa","prime","trainer","distribut","sgd","fsdp"],"s":["daa-prime-trainer","daa","prime","trainer"]},
+{"id":"daa-rules","family":"daa-rules","kind":"crate","store":"daa","version":null,"description":"Rules engine for DAA system providing policy enforcement and decision automation","keywords":["blockchain","daa","engine","policy","rules"],"source":"daa/daa-rules/Cargo.toml","sourceSha256":"1198c1cfcbe951146cc77221709d75c8ca0dfbc5c9fbf22d79a1b21fd2d2cb81","variants":[],"readme":"🚀 FULL IMPLEMENTATION - This is the complete, production-ready implementation of the DAA Rules Engine, not a placeholder. A comprehensive rules engine for the Decentralized Autonomous Agents (DAA) system, providing policy enforcement, decision automation, and governance capabilities. DAA Rules provides a flexible and powerful rules engine that enables autonomous agents to make decisions based on predefined policies and conditions. The engine supports complex logical operations, pattern matching, time-based conditions, and various action types. - Rule Definition: Flexible rule creation with conditions and actions - Condition Evaluation: Support for equality, comparison, pattern matching, and","readmeSource":"daa/daa-rules/README.md","product":"key:daa:daa-rules","canonical":"daa-rules","tier":"T1","t":["daa-rules","daa","rule","provid","policy","enforcement","decision","automation","blockchain"],"s":["daa-rules","daa","rule","blockchain","policy"]},
+{"id":"daa-sdk","family":"daa-sdk","kind":"npm","store":"daa","version":"0.1.0","description":"Unified SDK for DAA (Distributed Agentic Architecture) with native NAPI-rs bindings","keywords":["daa","distributed-agents","machine-learning","napi-rs","orchestration","quantum-resistant"],"source":"daa/packages/daa-sdk/package.json","sourceSha256":"23c97b07ca4423d6b429789482722d431c18dc8c02571f6c77e6cf2c3f5525c9","variants":[],"readme":"Unified TypeScript/JavaScript SDK for the Distributed Agentic Architecture (DAA) with native NAPI-rs bindings for maximum performance. - 🚀 High-Performance Native Bindings - NAPI-rs bindings for 2-4x faster quantum crypto operations - 🔐 Quantum-Resistant Cryptography - ML-KEM-768, ML-DSA, and BLAKE3 - 🌐 Universal Platform Support - Automatic platform detection (native vs WASM) - 📦 Zero-Config - Works out of the box with intelligent fallbacks - 🔧 TypeScript First - Full type definitions and IntelliSense support You can also use the QuDAG bindings directly: The SDK automatically detects the platform and loads the appropriate bindings: - Node.js: Loads native NAPI bindings (@daa/qudag-nati","readmeSource":"daa/packages/daa-sdk/README.md","product":"key:daa:daa-sdk","canonical":"daa-sdk","tier":"T1","t":["daa-sdk","daa","unify","distribut","architecture","distributed-agents","machine-learning","machine","learn","orchestration","quantum-resistant","quantum","resistant"],"s":["daa-sdk","daa","distributed-agents","distribut","machine-learning","machine","learn","orchestration","quantum-resistant","quantum","resistant"]},
+{"id":"daa-security","family":"daa-security","kind":"crate","store":"daa","version":null,"description":"Comprehensive security module for DAA with post-quantum crypto","keywords":[],"source":"daa/src/security/Cargo.toml","sourceSha256":"766f514f98368323cf0c9779e1503f91c6a302c6d8b364320f146f243884790d","variants":[],"readme":"Comprehensive security implementation for Decentralized Autonomous Applications (DAA) featuring post-quantum cryptography, secure aggregation, differential privacy, and economic incentives. - ML-KEM (Kyber): Quantum-resistant key encapsulation - ML-DSA (Dilithium): Quantum-resistant digital signatures - Secure Identities: Every participant has post-quantum secure keys - Hardware Integration: Support for hardware attestation - Threshold Secret Sharing: K-of-N aggregation protocol - Masked Gradients: Privacy-preserving gradient aggregation - Multi-Party Computation: Secure computation without revealing individual inputs - Verifiable Aggregation: Cryptographic proofs of correct aggregation - Gr","readmeSource":"daa/src/security/README.md","product":"key:daa:daa-security","canonical":"daa-security","tier":"T1","t":["daa-security","daa","security","comprehensive","post-quantum","post","quantum","crypto"],"s":["daa-security","daa","security"]},
+{"id":"daa-swarm","family":"daa-swarm","kind":"crate","store":"synaptic-mesh","version":null,"description":"Dynamic Agent Architecture - Swarm intelligence and autonomous coordination","keywords":["agents","autonomous","coordination","daa","swarm"],"source":"synaptic-mesh/src/rs/daa-swarm/Cargo.toml","sourceSha256":"e4482fb3d7e7d24db74cd707c7de6b0c9990e39270876b16c2817a994a30a814","variants":[],"product":"key:synaptic-mesh:daa-swarm","canonical":"daa-swarm","tier":"T1","t":["daa-swarm","daa","swarm","dynamic","architecture","intelligence","autonomous","coordination"],"s":["daa-swarm","daa","swarm","autonomous","coordination"]},
+{"id":"deployment","family":"deployment","kind":"crate","store":"autogenous","version":null,"description":"Autogenous deployment adapter: two-phase VERIFIED rollback (command restoration -> confirm restored artifact hash + health -> signed rollback receipt). Closes review finding #6 — rollback is executed and confirmed, not merely decided.","keywords":[],"source":"autogenous/crates/deployment/Cargo.toml","sourceSha256":"d6f0a8c327836c1818de5f9fe817d55c9b7cc04c9890155b5771bb1041213ebe","variants":[],"readme":"deployment is the actuation seam of [Autogenous](../), a self-evolving AI-defense runtime in Rust. It turns \"rollback\" from a status flag into a real, confirmed transition, and it fences concurrent rollouts so two candidates can't fight over the same target. The security review's objection was blunt: *rollback was represented, not executed or verified.* This crate fixes that — nothing here trusts a boolean. - DeploymentAdapter — the trait a real router/orchestrator implements (active, health, deploy, restore). InMemoryAdapter is the deterministic reference used by tests and the demo. - verified_rollback(...) — two-phase: (1) command restore, (2) confirm the active artifact hash *and* its hea","readmeSource":"autogenous/crates/deployment/README.md","product":"key:autogenous:deployment","canonical":"deployment","tier":null,"t":["deployment","autogenous","adapter","two-phase","phase","verify","rollback","command","restoration","confirm","restor","artifact","hash","health","sign","receipt","clos","review","find","execut","merely","decid"],"s":["deployment"]},
+{"id":"dream-machine","family":"dream-machine","kind":"npm","store":"dream-machine","version":"0.1.2","description":"The Dream Machine — a config-driven engine for nightly, cloud-scheduled, evidence-gated repository evolution. It never merges; a human does.","keywords":["agent-evolution","autonomous-agents","cli","darwin","dream-machine","flywheel","metaharness","nightly","ruvnet","self-improving-agents"],"source":"dream-machine/packages/cli/package.json","sourceSha256":"8b4ef2266c4fd38b84eb04ec465715c18506e3ce698e5f82402e74de1fcca028","variants":[],"readme":"A config-driven engine for nightly, cloud-scheduled, evidence-gated repository evolution. The Dream Machine wakes up in an isolated cloud session once a night, forms one falsifiable hypothesis about your repository, measures it against the repo's real evaluators, and writes down what it learned — whether or not the answer was the one it hoped for. Then it goes back to sleep, leaving behind a gist, an issue, a draft PR, and one durable ledger row. Every night ends in exactly one verdict — ACCEPT, REJECT, or INCONCLUSIVE — never a fourth, never silence. A rejected hypothesis with a clean measurement is a successful night. The engine optimizes for shrinking tomorrow's search space, not for prod","readmeSource":"dream-machine/packages/cli/README.md","product":"repo:dream-machine","canonical":"dream-machine","tier":null,"t":["dream-machine","dream","machine","config-driven","config","driven","nightly","cloud-scheduled","cloud","schedul","evidence-gated","evidence","gated","repository","evolution","never","merge","human","agent-evolution","autonomous-agents","autonomous","darwin","flywheel","metaharness","self-improving-agents","self","improv"],"s":["dream-machine","dream","machine","agent-evolution","evolution","autonomous-agents","autonomous","darwin","flywheel","metaharness","nightly","self-improving-agents","self","improv"]},
+{"id":"dream-machine-workspace","family":"dream-machine-workspace","kind":"npm","store":"dream-machine","version":"0.0.0","description":"The Dream Machine — config-driven, evidence-gated nightly repository evolution (workspace root)","keywords":[],"source":"dream-machine/package.json","sourceSha256":"92cd58ce7e9e42f365cb8f7514bccec8d29c6060cec159bf945a731e561a221b","variants":[],"readme":"A config-driven engine for nightly, cloud-scheduled, evidence-gated repository evolution. [Website & walkthrough →](https://ruvnet.github.io/dream-machine/) · [Dashboard](https://ruvnet.github.io/dream-machine/dashboard.html) · [Tutorials](https://ruvnet.github.io/dream-machine/tutorials.html) · [ADR-0001](docs/adrs/ADR-0001-dream-machine-engine.md) The Dream Machine turns *\"run one research-and-evolution cycle against a repository every night\"* into a config-compiled engine instead of a hand-maintained megaprompt. It wakes up in an isolated cloud session, forms one falsifiable hypothesis, measures it against the repo's real evaluators, and writes down what it learned — whether or not the an","readmeSource":"dream-machine/README.md","product":"repo:dream-machine","canonical":"dream-machine","tier":null,"t":["dream-machine-workspace","dream","machine","workspace","config-driven","config","driven","evidence-gated","evidence","gated","nightly","repository","evolution","root"],"s":["dream-machine-workspace","dream","machine","workspace"]},
+{"id":"dspy.ts","family":"dspy.ts","kind":"npm","store":"dspy.ts","version":"2.2.1","description":"DSPy.ts 2.0 - Advanced declarative AI framework with multi-agent orchestration, self-learning capabilities, and state-of-the-art optimizers. Powered by AgentDB, ReasoningBank, and Swarm architecture.","keywords":["agentdb","ai-agents","ai-framework","bayesian-optimization","causal-reasoning","chain-of-thought","declarative","dspy","dspy.ts","language-model","llm","machine-learning","memory-system","mipro","miprov2","multi-agent","onnx","optimization","pipeline","prompt-optimization","pytorch","react","reasoning","reasoning-bank","reflexion","self-learning","swarm","typescript","vector-database"],"source":"dspy.ts/package.json","sourceSha256":"07ddb653ffc159d3a0c63fb9ed47206bac485ad4af1fcc754615f635735f5328","variants":[],"readme":"Program AI systems, don't prompt them — in TypeScript, on AgentDB. DSPy.ts brings Stanford's [DSPy](https://github.com/stanfordnlp/dspy) to TypeScript: you declare *signatures* (typed input → output), compose them into *modules* and *pipelines*, give an *optimizer* a metric and a handful of examples, and it tunes the prompts and demonstrations for you — no hand-crafted prompt strings. Underneath, everything that needs a vector index, a memory, or a cache runs on [agentdb](https://www.npmjs.com/package/agentdb): HNSW search, RaBitQ quantization, ReasoningBank, reflexion. The optimizers *remember what worked* across runs. init nothing, npm install dspy.ts, and your LM program becomes optimizab","readmeSource":"dspy.ts/README.md","product":"repo:dspy.ts","canonical":"dspy.ts","tier":"T1","t":["dspy.t","2.0","advanc","declarative","multi-agent","multi","orchestration","self-learning","self","learn","capability","state-of-the-art","state","art","optimizer","power","agentdb","reasoningbank","swarm","architecture","ai-agents","ai-framework","bayesian-optimization","bayesian","causal-reasoning","causal","reason","chain-of-thought","chain","thought","dspy","language-model","language","model","machine-learning","machine","memory-system","memory","mipro","miprov2","onnx","pipeline","prompt-optimization","prompt","pytorch","react","reasoning-bank","bank","reflexion","vector-database","vector","database"],"s":["dspy.t","agentdb","ai-agents","ai-framework","bayesian-optimization","bayesian","causal-reasoning","causal","reason","chain-of-thought","chain","thought","declarative","dspy","language-model","language","model","machine-learning","machine","learn","memory-system","memory","mipro","miprov2","multi-agent","multi","onnx","pipeline","prompt-optimization","prompt","pytorch","react","reasoning-bank","bank","reflexion","self-learning","self","swarm","vector-database","vector","database"]},
+{"id":"dynamo-mcp-agent","family":"dynamo-mcp-agent","kind":"npm","store":"dynamo-mcp","version":"0.1.0","description":"Maintain and validate dynamo-mcp","keywords":[],"source":"dynamo-mcp/.harness/generated/package.json","sourceSha256":"152c6a6a72424e00ae9e3b780d7ab3306e68f7564d76860937cebbb7e5e91e20","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"dynamo-mcp/.harness/generated/README.md","product":"key:dynamo-mcp:dynamo-mcp-agent","canonical":"dynamo-mcp-agent","tier":"T3","t":["dynamo-mcp-agent","dynamo","mcp","maintain","validate","dynamo-mcp"],"s":["dynamo-mcp-agent","dynamo","mcp"]},
+{"id":"energy-harvester","family":"energy-harvester","kind":"crate","store":"cognitum-cogs","version":null,"description":"Energy-harvesting duty-cycled micro-agent with budget-gated execution","keywords":["duty-cycle","embedded","energy-harvesting","micro-agent","wasm"],"source":"cognitum-cogs/crates/energy-harvester/Cargo.toml","sourceSha256":"e3476b5861dce7d4b66e4add6f644e632fb2da2a35bf39223129e6ed50aaf942","variants":[],"product":"key:cognitum-cogs:energy-harvester","canonical":"energy-harvester","tier":null,"t":["energy-harvester","energy","harvester","energy-harvesting","harvest","duty-cycled","duty","cycl","micro-agent","micro","budget-gated","budget","gated","execution","duty-cycle","cycle","embed"],"s":["energy-harvester","energy","harvester","duty-cycle","duty","cycle","embed","energy-harvesting","harvest","micro-agent","micro"]},
+{"id":"envelope","family":"envelope","kind":"crate","store":"autogenous","version":null,"description":"Autogenous cryptographically-closed promotion path: content-bound candidate manifests, signed evaluation receipts (parent-vs-candidate, ≥2 distinct pinned judges), a signed promotion envelope, and a verify_promotion that rejects on every independent violation. Closes review findings 1-6,8.","keywords":[],"source":"autogenous/crates/envelope/Cargo.toml","sourceSha256":"292cc60650475df48245efe47b17c9344690d3c05a22882b2ef283a8a008f6c9","variants":[],"readme":"envelope is the trust core of [Autogenous](../), a self-evolving AI-defense runtime in Rust. It answers one demand from the security review: a promotion must depend on independently verified, content-bound evidence — never a caller-supplied boolean, an Option<String> \"signature\", or a separate \"declared effects\" argument. Nothing here trusts a number handed in by a generator. It binds three signed, content-addressed objects into a closed check, and — as of ADR-403 — mints a single-use promotion token that is the *only* thing allowed to promote a candidate. - CandidateManifest — the object everything binds to. Effects, capabilities, rollback target, and invariant proofs live *inside* it, so i","readmeSource":"autogenous/crates/envelope/README.md","product":"key:autogenous:envelope","canonical":"envelope","tier":null,"t":["envelope","autogenous","cryptographically-closed","cryptographically","clos","promotion","path","content-bound","content","bound","candidate","manifest","sign","evaluation","receipt","parent-vs-candidate","parent","distinct","pin","judge","verify","reject","independent","violation","review","finding","1-6"],"s":["envelope"]},
+{"id":"evaluator","family":"evaluator","kind":"crate","store":"autogenous","version":null,"description":"Autogenous replay evaluator: runs a candidate detector over labeled malicious/benign corpora, producing a fitness vector with Wilson-interval uncertainty (ADR-392 Phase 5).","keywords":[],"source":"autogenous/crates/evaluator/Cargo.toml","sourceSha256":"bb98382ff043725ab196a6552d1a1a7c2fbed1359e5ec873db6d0a83557c7d2b","variants":[],"readme":"evaluator is the measurement stage of [Autogenous](../), a self-evolving AI-defense runtime in Rust. It replays a candidate [antibody](../antibody) detector over labeled corpora (malicious + benign) and produces a ReplayReport: attack recall, benign false-positive rate, and — crucially — Wilson-interval uncertainty for both. Every estimate carries its confidence interval and its evidence population, so no downstream gate ever treats a number measured on 20 samples the same as one measured on 20,000. - replay_packaged(detector, corpus) — run the *serialized* detector that ships in an antibody (the exact artifact, not an ad-hoc closure). - replay(fn, corpus) — run any predicate over a corpus.","readmeSource":"autogenous/crates/evaluator/README.md","product":"key:autogenous:evaluator","canonical":"evaluator","tier":null,"t":["evaluator","autogenous","replay","candidate","detector","label","malicious","benign","corpora","produc","fitness","vector","wilson-interval","wilson","interval","uncertainty","adr-392","adr","392","phase"],"s":["evaluator"]},
+{"id":"fact-memory-mcp","family":"fact-memory-mcp","kind":"npm","store":"fact","version":"1.0.0","description":"FACT Memory System MCP Server - Mem0-inspired memory management using FACT SDK prompt caching","keywords":["ai","caching","fact","mcp","mem0","memory"],"source":"fact/fact-memory/src/package.json","sourceSha256":"19b81876b78906010ecbd68df3844ef625b8560d51857e359cc2f31d9272d69f","variants":[],"readme":"A demonstration FastMCP server implementation for the FACT Memory System, showcasing basic MCP patterns and best practices. This server provides a simple introduction to FastMCP development with: - Hello Tool: Returns a basic greeting message - Greet Tool: Returns a personalized greeting with name parameter - Server Info Resource: Provides server metadata and capabilities - Multiple Transport Modes: STDIO (standard MCP) and HTTP - Comprehensive Testing: Built-in test mode for development 1. hello() - Returns a simple hello message with timestamp - No parameters required - Demonstrates basic tool structure 2. greet(name: str) - Returns a personalized greeting - Validates and sanitizes input -","readmeSource":"fact/fact-memory/src/README.md","product":"key:fact:fact-memory-mcp","canonical":"fact-memory-mcp","tier":"T1","t":["fact-memory-mcp","fact","memory","mcp","server","mem0-inspired","mem0","inspir","management","prompt","cach"],"s":["fact-memory-mcp","fact","memory","mcp","cach","mem0"]},
+{"id":"federated-mcp","family":"federated-mcp","kind":"npm","store":"federated-mcp","version":"1.0.0","description":"Federated Model Context Protocol Implementation","keywords":[],"source":"federated-mcp/src/package.json","sourceSha256":"a803cc8a79f3baf86ea4a141ca75f46b0f15c9403d89aa8f286f9e2e18cc4a23","variants":[],"product":"repo:federated-mcp","canonical":"federated-mcp","tier":"T3","t":["federated-mcp","federat","mcp","model","context","protocol","implementation"],"s":["federated-mcp","federat","mcp"]},
+{"id":"federated-mcp-agent","family":"federated-mcp-agent","kind":"npm","store":"federated-mcp","version":"0.1.0","description":"Maintain and validate the federated MCP repository","keywords":[],"source":"federated-mcp/.harness/generated/package.json","sourceSha256":"f8e560c4c485187d2054390b57080351ba73c1ba0faef77fc539381437e5e69e","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"federated-mcp/.harness/generated/README.md","product":"key:federated-mcp:federated-mcp-agent","canonical":"federated-mcp-agent","tier":"T3","t":["federated-mcp-agent","federat","mcp","maintain","validate","repository"],"s":["federated-mcp-agent","federat","mcp"]},
+{"id":"fxnn","family":"fxnn","kind":"crate","store":"cognitum-cogs","version":null,"description":"High-speed atomic simulation with neural force fields","keywords":["molecular-dynamics","neural-network","physics","simulation"],"source":"cognitum-cogs/crates/fxnn/Cargo.toml","sourceSha256":"649003d2656bc20bd700069bde076e7664bef0422cd179842fc91d4be7b46815","variants":[],"readme":"A high-performance simulation engine in Rust that creates deterministic, replayable environments where agents can learn within enforced physical and logical constraints. This is a small, strict world that runs inside a computer and refuses to let impossible things happen. Cognitum is not just a physics engine - it's a reality-checking sandbox where: - State persists - Every action has consequences that remain - Rules are enforced - Invalid states are rejected, not ignored - Everything is deterministic - Same inputs always produce same outputs - All actions are witnessed - Tamper-evident audit trail with cryptographic hashes - Intelligence must play by the rules - Agents succeed by understand","readmeSource":"cognitum-cogs/crates/fxnn/README.md","product":"key:cognitum-cogs:fxnn","canonical":"fxnn","tier":null,"t":["fxnn","high-speed","high","speed","atomic","simulation","neural","force","field","molecular-dynamics","molecular","dynamic","neural-network","network","physic"],"s":["fxnn","molecular-dynamics","molecular","dynamic","neural-network","neural","network","physic","simulation"]},
+{"id":"gastown-formula-wasm","family":"gastown-formula","kind":"crate","store":"ruflo","version":null,"description":"Ultra-optimized WASM formula parsing and cooking for Gas Town Bridge","keywords":[],"source":"ruflo/v3/plugins/gastown-bridge/wasm/gastown-formula-wasm/Cargo.toml","sourceSha256":"516ccbfa083072a4df7e081680a2ac6b8a0e70bc17b191cd866706927d1adc58","variants":[],"product":"key:ruflo:gastown-formula","canonical":"gastown-formula-wasm","tier":"T0","t":["gastown-formula-wasm","gastown","formula","gastown-formula","ultra-optimized","ultra","pars","cook","gas","town","bridge"],"s":["gastown-formula-wasm","gastown","formula","gastown-formula"]},
+{"id":"gastown-shared","family":"gastown-shared","kind":"crate","store":"ruflo","version":null,"description":"Shared optimized utilities for Gas Town WASM modules","keywords":[],"source":"ruflo/v3/plugins/gastown-bridge/wasm/shared/Cargo.toml","sourceSha256":"ae50d65e7370906d856bf0e8a4cd46feeb923173f09f45ae180897e5188d361a","variants":[],"product":"key:ruflo:gastown-shared","canonical":"gastown-shared","tier":"T0","t":["gastown-shared","gastown","shar","utility","gas","town"],"s":["gastown-shared","gastown","shar"]},
+{"id":"goalie","family":"goalie","kind":"npm","store":"sublinear-time-solver","version":"1.3.1","description":"AI-powered research assistant with REAL Ed25519 cryptographic signatures, GOAP planning, and Perplexity API integration","keywords":["ai","anti-hallucination","cryptography","ed25519","goap","mcp","perplexity","planning","search","strange-loops"],"source":"sublinear-time-solver/npx/goalie/package.json","sourceSha256":"a3243464632efcd24bb6355163b380dfe4e48033800e370f1de411f64c1eb69a","variants":[],"readme":"Created by [rUv](https://github.com/ruvnet) - Building the future of verifiable AI research Goalie works seamlessly with AI assistants like Claude through MCP: Once configured, Claude can use advanced research capabilities directly through natural language! Unlike traditional AI search tools that provide single-shot answers with limited sources, Goalie is a deep research system that: - Decomposes complex questions into multiple research goals - Creates intelligent action plans using A* pathfinding algorithms - Dynamically re-plans when actions fail (up to 3 attempts) - Optimizes research paths for efficiency and completeness - Citation Tracking: Attempts to provide sources for claims - Ed255","readmeSource":"sublinear-time-solver/npx/goalie/README.md","product":"key:sublinear-time-solver:goalie","canonical":"goalie","tier":"T1","t":["goalie","ai-powered","power","research","assistant","real","ed25519","cryptographic","signature","goap","plann","perplexity","integration","anti-hallucination","anti","hallucination","cryptography","mcp","search","strange-loops","strange","loop"],"s":["goalie","anti-hallucination","anti","hallucination","cryptography","ed25519","goap","mcp","perplexity","plann","search","strange-loops","strange","loop"]},
+{"id":"guardrail-agent","family":"guardrail-agent","kind":"npm","store":"guardrail","version":"0.1.0","description":"Maintain and validate guardrail","keywords":[],"source":"guardrail/.harness/generated/package.json","sourceSha256":"e531a6109c1f01a8f6c702f6349a1a21c698aee91a0ff8449a863ccb6670548d","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"guardrail/.harness/generated/README.md","product":"key:guardrail:guardrail-agent","canonical":"guardrail-agent","tier":"T3","t":["guardrail-agent","guardrail","maintain","validate"],"s":["guardrail-agent","guardrail"]},
+{"id":"guidance-kernel","family":"guidance-kernel","kind":"npm","store":"ruflo","version":"0.1.0","description":"Rust WASM kernel for @claude-flow/guidance — deterministic policy engine","keywords":[],"source":"ruflo/v3/@claude-flow/guidance/wasm-pkg/package.json","sourceSha256":"17a86c44528724a288132bea3bf343c034325cba60f356e66776b26b702ec24e","variants":[],"product":"key:ruflo:guidance-kernel","canonical":"guidance-kernel","tier":"T0","t":["guidance-kernel","guidance","kernel","claude-flow","claude","flow","deterministic","policy"],"s":["guidance-kernel","guidance","kernel"]},
+{"id":"hailort-sys","family":"hailort","kind":"crate","store":"ruvector","version":null,"description":"Raw FFI bindings for Hailo's HailoRT C library (ADR-167)","keywords":["embedded","ffi","hailo","hailort","npu"],"source":"ruvector/crates/hailort-sys/Cargo.toml","sourceSha256":"1b53ff4531ed32528c7efa48cb698eee2f0b998821be3c1c370b0040b8cb83f5","variants":[],"readme":"Rust FFI bindings to Hailo's HailoRT C library. Generated via bindgen at build time; no hand-curated wrappers — ruvector-hailo adds the safe Rust surface on top. - HailoRT C library installed (libhailort.so + headers) at /usr/lib/aarch64-linux-gnu/libhailort.so and /usr/include/hailo/. Standard package on Hailo-supported boards; on x86 dev hosts cargo check works without it because ruvector-hailo's default feature set excludes hailo. - bindgen 0.71+ build dependency (pulls libclang). The full HailoRT C API surface, including: - hailo_create_vdevice / hailo_release_vdevice — virtual device lifecycle. - hailo_create_hef_file / hailo_release_hef — HEF artifact load. - hailo_init_configure_param","readmeSource":"ruvector/crates/hailort-sys/README.md","product":"key:ruvector:hailort","canonical":"hailort-sys","tier":"T0","t":["hailort-sys","hailort","sys","raw","ffi","hailo","adr-167","adr","167","embed","npu"],"s":["hailort-sys","hailort","sys","embed","ffi","hailo","npu"]},
+{"id":"helix-bioage","family":"helix-bioage","kind":"crate","store":"helix","version":null,"description":"ADR-034: deterministic biological-age estimate (Levine PhenoAge) from routine labs — non-diagnostic","keywords":[],"source":"helix/crates/helix-bioage/Cargo.toml","sourceSha256":"ee2c9c54c64c1b0240b1d60f7d017e28cb601f263cb416d57fb93393d7acae87","variants":[],"product":"key:helix:helix-bioage","canonical":"helix-bioage","tier":null,"t":["helix-bioage","helix","bioage","adr-034","adr","034","deterministic","biological-age","biological","age","estimate","levine","phenoage","routine","labs","non-diagnostic","non","diagnostic"],"s":["helix-bioage","helix","bioage"]},
+{"id":"helix-cohort","family":"helix-cohort","kind":"crate","store":"helix","version":null,"description":"ADR-024: privacy-preserving cohort feature extraction (federation primitive) — generalization + cell-suppression + differential privacy","keywords":[],"source":"helix/crates/helix-cohort/Cargo.toml","sourceSha256":"77f7df0a0eca19f4fe887d12d8fe518a8f59f2f4abeba038b809bbbefba44709","variants":[],"product":"key:helix:helix-cohort","canonical":"helix-cohort","tier":null,"t":["helix-cohort","helix","cohort","adr-024","adr","024","privacy-preserving","privacy","preserv","feature","extraction","federation","primitive","generalization","cell-suppression","cell","suppression","differential"],"s":["helix-cohort","helix","cohort"]},
+{"id":"helix-connect","family":"helix-connect","kind":"crate","store":"helix","version":null,"description":"ADR-029: live connector clients (FHIR/SMART + wearables) — transport-trait core with FHIR->ProvRecord parsing and the degradation ladder","keywords":[],"source":"helix/crates/helix-connect/Cargo.toml","sourceSha256":"ae8240def71fcc924843a265039677ccdd8cb92c8f3c0717ac8fb7c5bc691ad6","variants":[],"product":"key:helix:helix-connect","canonical":"helix-connect","tier":null,"t":["helix-connect","helix","connect","adr-029","adr","029","live","connector","client","fhir","smart","wearable","transport-trait","transport","trait","core","provrecord","pars","degradation","ladder"],"s":["helix-connect","helix","connect"]},
+{"id":"helix-embed","family":"helix-embed","kind":"crate","store":"helix","version":null,"description":"ADR-027: learned MiniLM text embeddings (local GPU) wired into helix-retrieval","keywords":[],"source":"helix/crates/helix-embed/Cargo.toml","sourceSha256":"075636e5db235b1431ae056f8529e22c8e6befe62fd5307384144de70620de5b","variants":[],"product":"key:helix:helix-embed","canonical":"helix-embed","tier":null,"t":["helix-embed","helix","embed","adr-027","adr","027","learn","minilm","text","embedding","local","gpu","wired","helix-retrieval","retrieval"],"s":["helix-embed","helix","embed"]},
+{"id":"helix-escalation","family":"helix-escalation","kind":"crate","store":"helix","version":null,"description":"ADR-009: red-flag escalation & clinician-in-the-loop registry for Helix","keywords":[],"source":"helix/crates/helix-escalation/Cargo.toml","sourceSha256":"b09926c7b89b1669feef7d3232773ca667b768f195fd31b9e487aad6de01b3e8","variants":[],"product":"key:helix:helix-escalation","canonical":"helix-escalation","tier":null,"t":["helix-escalation","helix","escalation","adr-009","adr","009","red-flag","red","flag","clinician-in-the-loop","clinician","loop","registry"],"s":["helix-escalation","helix","escalation"]},
+{"id":"helix-evidence","family":"helix-evidence","kind":"crate","store":"helix","version":null,"description":"ADR-006: evidence tiering and explicit abstention policy for Helix","keywords":[],"source":"helix/crates/helix-evidence/Cargo.toml","sourceSha256":"a165c45293958fe41c660403e7be552295b521badc13b7c33e08bbd153bc7888","variants":[],"product":"key:helix:helix-evidence","canonical":"helix-evidence","tier":null,"t":["helix-evidence","helix","evidence","adr-006","adr","006","tier","explicit","abstention","policy"],"s":["helix-evidence","helix","evidence"]},
+{"id":"helix-evolve","family":"helix-evolve","kind":"crate","store":"helix","version":null,"description":"ADR-035: Darwin-style parameter evolution — tunes non-safety pipeline params against a labeled eval set; safety thresholds frozen","keywords":[],"source":"helix/crates/helix-evolve/Cargo.toml","sourceSha256":"f899e71336f4bfd8af6654c149c641eee57f80b6fd477f8f16472e1876efc782","variants":[],"product":"key:helix:helix-evolve","canonical":"helix-evolve","tier":null,"t":["helix-evolve","helix","evolve","adr-035","adr","035","darwin-style","darwin","style","parameter","evolution","tune","non-safety","non","safety","pipeline","param","against","label","eval","set","threshold","frozen"],"s":["helix-evolve","helix","evolve"]},
+{"id":"helix-fed","family":"helix-fed","kind":"crate","store":"helix","version":null,"description":"ADR-030: federation transport — opt-in, DP-gated cohort contribution over the helix-cohort privacy primitive","keywords":[],"source":"helix/crates/helix-fed/Cargo.toml","sourceSha256":"3ad7581d71865e46e60d54271fe166997c7d3af008c4f24e5b4b5e805fc97e37","variants":[],"product":"key:helix:helix-fed","canonical":"helix-fed","tier":null,"t":["helix-fed","helix","fed","adr-030","adr","030","federation","transport","opt-in","opt","dp-gated","gated","cohort","contribution","helix-cohort","privacy","primitive"],"s":["helix-fed","helix","fed"]},
+{"id":"helix-focus","family":"helix-focus","kind":"crate","store":"helix","version":null,"description":"ADR-032: deterministic, non-diagnostic 'focus areas' selection from the user's own records","keywords":[],"source":"helix/crates/helix-focus/Cargo.toml","sourceSha256":"08ee5ac4217ac79c4d3669b97e06c8c3ff29df4a9370dc65e94a4c4b76f73429","variants":[],"product":"key:helix:helix-focus","canonical":"helix-focus","tier":null,"t":["helix-focus","helix","focus","adr-032","adr","032","deterministic","non-diagnostic","non","diagnostic","area","selection","user","own","record"],"s":["helix-focus","helix","focus"]},
+{"id":"helix-genome","family":"helix-genome","kind":"crate","store":"helix","version":null,"description":"ADR-021: user-owned genome ingestion & pharmacogenomics (rvDNA backend) -> Helix ProvRecords + prescriber advisories","keywords":[],"source":"helix/crates/helix-genome/Cargo.toml","sourceSha256":"3ec7bbd4a13af87486c6e7d3eec86d437f580de8b1a63b2556ceaed8d1defa94","variants":[],"product":"key:helix:helix-genome","canonical":"helix-genome","tier":null,"t":["helix-genome","helix","genome","adr-021","adr","021","user-owned","user","owned","ingestion","pharmacogenomic","rvdna","backend","provrecord","prescriber","advisory"],"s":["helix-genome","helix","genome"]},
+{"id":"helix-llm","family":"helix-llm","kind":"crate","store":"helix","version":null,"description":"ADR-026: on-device LLM analyst — narrates grounded facts via a local-GPU model, with an anti-hallucination number-guard","keywords":[],"source":"helix/crates/helix-llm/Cargo.toml","sourceSha256":"0190ff079a917f211386d12d17bda6987cf848fbaf4997808ae105b1f9a02212","variants":[],"product":"key:helix:helix-llm","canonical":"helix-llm","tier":null,"t":["helix-llm","helix","adr-026","adr","026","on-device","device","analyst","narrate","ground","fact","via","local-gpu","local","gpu","model","anti-hallucination","anti","hallucination","number-guard","number","guard"],"s":["helix-llm","helix"]},
+{"id":"helix-neural","family":"helix-neural","kind":"crate","store":"helix","version":null,"description":"Integration adapter: ruvnet/ruv-neural gamma-entrainment/EEG session evidence -> Helix ProvRecords (research-grade screening, not diagnosis)","keywords":[],"source":"helix/crates/helix-neural/Cargo.toml","sourceSha256":"4a58b6d50c9a3e331a525c0c6fa0ea93c2b8713b2ab54def6513177d917c954f","variants":[],"product":"key:helix:helix-neural","canonical":"helix-neural","tier":null,"t":["helix-neural","helix","neural","integration","adapter","ruv-neural","gamma-entrainment","gamma","entrainment","eeg","session","evidence","provrecord","research-grade","research","grade","screen","diagnosis"],"s":["helix-neural","helix","neural"]},
+{"id":"helix-neurosleep","family":"helix-neurosleep","kind":"crate","store":"helix","version":null,"description":"ADR-051 verified NeuroSleep research ingestion and sealed longitudinal policy","keywords":[],"source":"helix/crates/helix-neurosleep/Cargo.toml","sourceSha256":"400ae9502ba62566e9d4710d5b9f8aa5e9f1d892afe3c5d694f6ae78b61bd1f9","variants":["helix-neurosleep-wasm"],"product":"key:helix:helix-neurosleep","canonical":"helix-neurosleep","tier":null,"t":["helix-neurosleep","helix","neurosleep","adr-051","adr","051","verify","research","ingestion","seal","longitudinal","policy"],"s":["helix-neurosleep","helix","neurosleep"]},
+{"id":"helix-numeric","family":"helix-numeric","kind":"crate","store":"helix","version":null,"description":"ADR-007: deterministic numeric/trend engine for Helix (the LLM never does arithmetic)","keywords":[],"source":"helix/crates/helix-numeric/Cargo.toml","sourceSha256":"7a7c19b3384df9e77c98ed4ea8b2be41fb141c2f0b04ba81ff68386bd099585d","variants":[],"product":"key:helix:helix-numeric","canonical":"helix-numeric","tier":null,"t":["helix-numeric","helix","numeric","adr-007","adr","007","deterministic","trend","never","arithmetic"],"s":["helix-numeric","helix","numeric"]},
+{"id":"helix-ocr","family":"helix-ocr","kind":"crate","store":"helix","version":null,"description":"ADR-022: OCR lab-document ingestion (RuVector OCR backend) -> gated, conservative-confidence ProvRecords","keywords":[],"source":"helix/crates/helix-ocr/Cargo.toml","sourceSha256":"df94b161ea3d202d7a0e0004748cd728b58549f2f1f787e242466527951ea6ff","variants":[],"product":"key:helix:helix-ocr","canonical":"helix-ocr","tier":null,"t":["helix-ocr","helix","ocr","adr-022","adr","022","lab-document","lab","document","ingestion","backend","gated","conservative-confidence","conservative","confidence","provrecord"],"s":["helix-ocr","helix","ocr"]},
+{"id":"helix-ontology","family":"helix-ontology","kind":"crate","store":"helix","version":null,"description":"ADR-004: canonical ontology normalization (LOINC/RxNorm/SNOMED/ICD-10/UCUM) for Helix","keywords":[],"source":"helix/crates/helix-ontology/Cargo.toml","sourceSha256":"31dd9ed3af8024e755a640eadd6f0a6f7c507d33a7764f0a100e79dfff207b35","variants":[],"product":"key:helix:helix-ontology","canonical":"helix-ontology","tier":null,"t":["helix-ontology","helix","ontology","adr-004","adr","004","canonical","normalization","loinc","rxnorm","snom","icd-10","icd","ucum"],"s":["helix-ontology","helix","ontology"]},
+{"id":"helix-openmed","family":"helix-openmed","kind":"crate","store":"helix","version":null,"description":"Fail-closed OpenMed clinical-text privacy and de-identification gate for Helix","keywords":[],"source":"helix/crates/helix-openmed/Cargo.toml","sourceSha256":"59631eaa087d07f94c0ca89901fbcf5f74520ab3699f11f97fb46c4a2daf49c1","variants":[],"product":"key:helix:helix-openmed","canonical":"helix-openmed","tier":null,"t":["helix-openmed","helix","openm","fail-closed","fail","clos","clinical-text","clinical","text","privacy","de-identification","identification","gate"],"s":["helix-openmed","helix","openm"]},
+{"id":"helix-pipeline","family":"helix-pipeline","kind":"crate","store":"helix","version":null,"description":"ADR-002/005: the grounded-answer pipeline wiring Helix's anti-hallucination crates","keywords":[],"source":"helix/crates/helix-pipeline/Cargo.toml","sourceSha256":"1ae4f153b919be262b086eb183b414947ca67d5b25c8092083d6956aabe94ffe","variants":[],"product":"key:helix:helix-pipeline","canonical":"helix-pipeline","tier":null,"t":["helix-pipeline","helix","pipeline","adr-002","adr","002","005","grounded-answer","ground","answer","wiring","anti-hallucination","anti","hallucination","crate"],"s":["helix-pipeline","helix","pipeline"]},
+{"id":"helix-provenance","family":"helix-provenance","kind":"crate","store":"helix","version":null,"description":"ADR-005: provenance records and retrieval-grounding enforcement for Helix","keywords":[],"source":"helix/crates/helix-provenance/Cargo.toml","sourceSha256":"8e6e99488938c95af9eeb7db1593c8c24a53334006b644abb15e22efbe6af387","variants":[],"product":"key:helix:helix-provenance","canonical":"helix-provenance","tier":null,"t":["helix-provenance","helix","provenance","adr-005","adr","005","record","retrieval-grounding","retrieval","ground","enforcement"],"s":["helix-provenance","helix","provenance"]},
+{"id":"helix-refranges","family":"helix-refranges","kind":"crate","store":"helix","version":null,"description":"NHANES-derived population reference intervals — a labeled FALLBACK when a lab supplies no reference range","keywords":[],"source":"helix/crates/helix-refranges/Cargo.toml","sourceSha256":"7b82410fd98b4cdea89657b7aa10609c239df1e605278ea25c3bf8707f9e8737","variants":[],"product":"key:helix:helix-refranges","canonical":"helix-refranges","tier":null,"t":["helix-refranges","helix","refrange","nhanes-derived","nhane","deriv","population","reference","interval","label","fallback","lab","supply","range"],"s":["helix-refranges","helix","refrange"]},
+{"id":"helix-retrieval","family":"helix-retrieval","kind":"crate","store":"helix","version":null,"description":"ADR-023: semantic retrieval contract over the health graph (RuVector HNSW/GraphRAG backend)","keywords":[],"source":"helix/crates/helix-retrieval/Cargo.toml","sourceSha256":"edb7affca7e63e7886a9d73f9162e088410f33eff99edeff882d883e5d6855e7","variants":[],"product":"key:helix:helix-retrieval","canonical":"helix-retrieval","tier":null,"t":["helix-retrieval","helix","retrieval","adr-023","adr","023","semantic","contract","health","graph","hnsw","graphrag","backend"],"s":["helix-retrieval","helix","retrieval"]},
+{"id":"helix-score","family":"helix-score","kind":"crate","store":"helix","version":null,"description":"ADR-016: transparent, decomposable, non-diagnostic 0-100 health score for Helix","keywords":[],"source":"helix/crates/helix-score/Cargo.toml","sourceSha256":"64f258b2eed5a7eaa60494018d43174811bf1f8fbb2fe528c1140fce6418c1ea","variants":[],"product":"key:helix:helix-score","canonical":"helix-score","tier":null,"t":["helix-score","helix","score","adr-016","adr","016","transparent","decomposable","non-diagnostic","non","diagnostic","0-100","100","health"],"s":["helix-score","helix","score"]},
+{"id":"helix-sensing","family":"helix-sensing","kind":"crate","store":"helix","version":null,"description":"ADR-020: RuView WiFi-CSI contactless ambient sensing -> Helix ProvRecords + escalation screening flags","keywords":[],"source":"helix/crates/helix-sensing/Cargo.toml","sourceSha256":"f14bfdc640c96e2c453d7d43a79effdc490833ec680be6f3a6e33b416bf9972a","variants":[],"product":"key:helix:helix-sensing","canonical":"helix-sensing","tier":null,"t":["helix-sensing","helix","sens","adr-020","adr","020","ruview","wifi-csi","wifi","csi","contactless","ambient","provrecord","escalation","screen","flag"],"s":["helix-sensing","helix","sens"]},
+{"id":"helix-timeline","family":"helix-timeline","kind":"crate","store":"helix","version":null,"description":"ADR-031: longitudinal health-score timeline — deterministic, methodology-versioned, trend-first","keywords":[],"source":"helix/crates/helix-timeline/Cargo.toml","sourceSha256":"abdf19995452b0404f4766177c472dc15750e13e89382cc897d1b0eb391634b0","variants":[],"product":"key:helix:helix-timeline","canonical":"helix-timeline","tier":null,"t":["helix-timeline","helix","timeline","adr-031","adr","031","longitudinal","health-score","health","score","deterministic","methodology-versioned","methodology","version","trend-first","trend"],"s":["helix-timeline","helix","timeline"]},
+{"id":"helix-vault","family":"helix-vault","kind":"crate","store":"helix","version":null,"description":"ADR-001/013: user-owned, local-first encrypted vault boundary (XChaCha20-Poly1305) for Helix","keywords":[],"source":"helix/crates/helix-vault/Cargo.toml","sourceSha256":"776722e0acc063d48bb3a70125c5577490acee09cfb56851f20f906ebfd7e2b5","variants":[],"product":"key:helix:helix-vault","canonical":"helix-vault","tier":null,"t":["helix-vault","helix","vault","adr-001","adr","001","013","user-owned","user","owned","local-first","local","encrypt","boundary","xchacha20-poly1305","xchacha20","poly1305"],"s":["helix-vault","helix","vault"]},
+{"id":"helix-verifier","family":"helix-verifier","kind":"crate","store":"helix","version":null,"description":"ADR-008: independent verifier/critic that re-derives every claim from source before it reaches the user","keywords":[],"source":"helix/crates/helix-verifier/Cargo.toml","sourceSha256":"94d3b4e092959652b3dabda499b65f037e1a590f474be8bbb564702686117a47","variants":[],"product":"key:helix:helix-verifier","canonical":"helix-verifier","tier":null,"t":["helix-verifier","helix","verifier","adr-008","adr","008","independent","critic","re-derives","derive","claim","source","before","reach","user"],"s":["helix-verifier","helix","verifier"]},
+{"id":"helix-vision","family":"helix-vision","kind":"crate","store":"helix","version":null,"description":"ADR-028: learned visual encoder (local GPU) — layout-only vision captioning + MiniLM embedding for helix-visual","keywords":[],"source":"helix/crates/helix-vision/Cargo.toml","sourceSha256":"bf67646d780d39d3afab1fef28ac4a27fd88007e33400de08d0418e025032bfe","variants":[],"product":"key:helix:helix-vision","canonical":"helix-vision","tier":null,"t":["helix-vision","helix","vision","adr-028","adr","028","learn","visual","encoder","local","gpu","layout-only","layout","only","caption","minilm","embedd","helix-visual"],"s":["helix-vision","helix","vision"]},
+{"id":"helix-visual","family":"helix-visual","kind":"crate","store":"helix","version":null,"description":"ADR-025: visual RAG over medical documents/images (rupixel backend) — tile embeddings + MaxSim late interaction","keywords":[],"source":"helix/crates/helix-visual/Cargo.toml","sourceSha256":"2b8c07e3e5e46951e93fde64557a897be8705eb68b62338fb0678a55489b7939","variants":[],"product":"key:helix:helix-visual","canonical":"helix-visual","tier":null,"t":["helix-visual","helix","visual","adr-025","adr","025","rag","medical","document","image","rupixel","backend","tile","embedding","maxsim","late","interaction"],"s":["helix-visual","helix","visual"]},
+{"id":"helix-wasm","family":"helix","kind":"crate","store":"helix","version":null,"description":"WASM binding exposing the Helix analytic pipeline to the web/mobile UI","keywords":[],"source":"helix/crates/helix-wasm/Cargo.toml","sourceSha256":"5bfd53d2e14fb9c4476eda4af7aba69c80e78feb6d7936c98094741392a510bb","variants":[],"product":"repo:helix","canonical":"helix-wasm","tier":null,"t":["helix-wasm","helix","expos","analytic","pipeline","web","mobile"],"s":["helix-wasm","helix"]},
+{"id":"hnsw_rs","family":"hnsw_rs","kind":"crate","store":"ruvector","version":null,"description":"Ann based on Hierarchical Navigable Small World Graphs from Yu.A. Malkov and D.A Yashunin","keywords":["algorithms","ann","hnsw"],"source":"ruvector/patches/hnsw_rs/Cargo.toml","sourceSha256":"800666b7d054ae2f25e6f5759063dd8ac969c8ed7aefa577c4ac4faf072c236d","variants":[],"readme":"This crate provides a Rust implementation of the paper by Yu.A. Malkov and D.A Yashunin: \"Efficient and Robust approximate nearest neighbours using Hierarchical Navigable Small World Graphs\" (2016,2018) [arxiv](https://arxiv.org/abs/1603.09320) The crate is built on top of the [anndists](https://crates.io/crates/anndists) and can use the following distances: * usual distances as L1, L2, Cosine, Jaccard, Hamming for vectors of standard numeric types, Levenshtein distance on u16. * Hellinger distance and Jeffreys divergence between probability distributions (f32 and f64). It must be noted that the Jeffreys divergence (a symetrized Kullback-Leibler divergence) do not satisfy the triangle inequa","readmeSource":"ruvector/patches/hnsw_rs/README.md","product":"key:ruvector:hnsw_rs","canonical":"hnsw_rs","tier":"T0","t":["hnsw","ann","hierarchical","navigable","small","world","graph","yu.a","malkov","d.a","yashunin","algorithm"],"s":["hnsw","algorithm","ann"]},
+{"id":"homecore","family":"homecore","kind":"npm","store":"ruview","version":"0.1.0","description":"WASM-first Homecore developer metaharness with source-cited guidance, MCP tools, and guarded Claude Code and Codex adapters.","keywords":["agent-harness","claude-code","codex","home-assistant","homecore","mcp","metaharness","wasmtime","webassembly"],"source":"ruview/harness/homecore/package.json","sourceSha256":"05da724a5d4b310f89398d39b85fe20607588d7b01e2b373233b7846192f48f2","variants":[],"readme":"homecore is the developer-facing metaharness for RuView's native Rust Homecore stack. Its package contract is: The harness maps current capabilities to source and validation commands, exposes a bounded MCP server, and can delegate repository exploration to a locally installed Claude Code or Codex CLI. It does not start a home server, change configuration, migrate data, or publish code by itself. Workspace-writing host runs require both --allow-write and --confirm. The harness never emits permission-bypass flags. Every MCP tool is read-only; Cargo verification remains an explicit local CLI operation and is not exposed through MCP. MCP request size, queue depth, tool-call duration, and the tot","readmeSource":"ruview/harness/homecore/README.md","product":"key:ruview:homecore","canonical":"homecore","tier":"T0","t":["homecore","wasm-first","developer","metaharness","source-cited","source","cited","guidance","mcp","guard","claude","code","codex","adapter","agent-harness","harness","claude-code","home-assistant","home","assistant","wasmtime"],"s":["homecore","agent-harness","harness","claude-code","claude","code","codex","home-assistant","home","assistant","mcp","metaharness","wasmtime"]},
+{"id":"homecore-api","family":"homecore-api","kind":"crate","store":"ruview","version":null,"description":"Wire-compatible Axum REST + WebSocket port of Home Assistant's API (ADR-130)","keywords":[],"source":"ruview/v2/crates/homecore-api/Cargo.toml","sourceSha256":"4023d66681e4c8e01834db9ec87386d79573fa62c9c74ed1af4fea0c6a912bf6","variants":[],"readme":"Home Assistant-compatible REST + WebSocket API for HOMECORE state and events. Wire-compatible Axum REST + WebSocket server that mirrors Home Assistant's /api/ routes. Ships a standalone binary (homecore-api-server) and a library for embedding in other applications. homecore-api provides the HTTP boundary layer for HOMECORE. It wires Axum routes to the homecore state machine, exposing: - GET /api/states — list all entity states - GET /api/states/:entity_id — fetch a single entity's state + attributes - POST /api/states/:entity_id — update an entity's state and attributes - GET /api/services — list registered services - POST /api/services/:domain/:service — call a service with arguments - GET","readmeSource":"ruview/v2/crates/homecore-api/README.md","product":"key:ruview:homecore-api","canonical":"homecore-api","tier":"T0","t":["homecore-api","homecore","wire-compatible","wire","compatible","axum","rest","websocket","port","home","assistant","adr-130","adr","130"],"s":["homecore-api","homecore"]},
+{"id":"homecore-assist","family":"homecore-assist","kind":"crate","store":"ruview","version":null,"description":"HOMECORE voice/intent pipeline + ruflo agent bridge (ADR-133 P1 scaffold)","keywords":[],"source":"ruview/v2/crates/homecore-assist/Cargo.toml","sourceSha256":"33933c7ba898d055eb79ddd6470c992b1bb2a911c6a154d62add744465c7c7bd","variants":[],"readme":"Voice-activated intent recognition and execution pipeline for HOMECORE with Ruflo agent bridge (P2). P1 scaffold: intent recognition via regex patterns, 5 built-in intent handlers (turn on/off, set brightness, cancel), and Ruflo runner trait surface. Real tokio::process subprocess integration (P2) allows orchestration with Ruflo agents for complex multi-step actions. homecore-assist is the voice/NLU gateway for HOMECORE. It takes natural language utterances, recognizes which intent they represent, and executes the appropriate action. It provides: - IntentRecognizer trait — abstraction for matching utterances to intents - RegexIntentRecognizer — P1 built-in; uses regex patterns (HA classic st","readmeSource":"ruview/v2/crates/homecore-assist/README.md","product":"key:ruview:homecore-assist","canonical":"homecore-assist","tier":"T0","t":["homecore-assist","homecore","assist","voice","intent","pipeline","ruflo","bridge","adr-133","adr","133","scaffold"],"s":["homecore-assist","homecore","assist"]},
+{"id":"homecore-automation","family":"homecore-automation","kind":"crate","store":"ruview","version":null,"description":"Automation engine, trigger evaluator, and MiniJinja template evaluator for HOMECORE (ADR-129)","keywords":[],"source":"ruview/v2/crates/homecore-automation/Cargo.toml","sourceSha256":"3be4702c5e85273a65e0cf30260a48c11348581669f7655630aa5b11ca9466a3","variants":[],"readme":"YAML-based automation engine for HOMECORE with trigger evaluation, conditions, and MiniJinja template support. Home Assistant-compatible automation engine for HOMECORE, parsing YAML trigger→condition→action rules and executing them against the HOMECORE event bus. homecore-automation provides the runtime for HOMECORE automations — YAML files that define \"if X happens and Y is true, do Z\". It includes: - Automation struct — YAML-deserializable automation definition with id, alias, triggers, conditions, actions, and run mode (single, parallel, restart) - Trigger evaluation — state-changed, time-based, template, and service-call triggers; async EvaluateTrigger trait - Condition evaluation — stat","readmeSource":"ruview/v2/crates/homecore-automation/README.md","product":"key:ruview:homecore-automation","canonical":"homecore-automation","tier":"T0","t":["homecore-automation","homecore","automation","trigger","evaluator","minijinja","template","adr-129","adr","129"],"s":["homecore-automation","homecore","automation"]},
+{"id":"homecore-hap","family":"homecore-hap","kind":"crate","store":"ruview","version":null,"description":"Fail-closed HomeKit Accessory Protocol network foundation for HOMECORE","keywords":[],"source":"ruview/v2/crates/homecore-hap/Cargo.toml","sourceSha256":"4deeb64e94051de6abb3d4136c20297bebd4735c3710914ae33c5b254043c04e","variants":[],"readme":"homecore-hap is HOMECORE's bounded, fail-closed HAP IP accessory server (ADR-125). It implements the HAP R2 cryptographic pairing and transport boundary without relying on the broken hap 0.1 pre-release crate. - Pair-Setup M1-M6 uses the RFC 5054 3072-bit group with SHA-512, the HAP compatibility proof construction, HKDF-SHA512, ChaCha20-Poly1305, and Ed25519 long-term keys. - Pair-Verify M1-M4 uses ephemeral X25519, strict Ed25519 transcript verification, HKDF-SHA512, and authenticated encrypted sub-TLVs. - After successful M4, all HTTP and EVENT/1.0 traffic uses HAP records: two-byte little-endian authenticated lengths, at most 1024 plaintext bytes, independent directional keys, and monoto","readmeSource":"ruview/v2/crates/homecore-hap/README.md","product":"key:ruview:homecore-hap","canonical":"homecore-hap","tier":"T0","t":["homecore-hap","homecore","hap","fail-closed","fail","clos","homekit","accessory","protocol","network","foundation"],"s":["homecore-hap","homecore","hap"]},
+{"id":"homecore-migrate","family":"homecore-migrate","kind":"crate","store":"ruview","version":null,"description":"Migration tooling from Python Home Assistant to HOMECORE (ADR-165 P1 scaffold)","keywords":[],"source":"ruview/v2/crates/homecore-migrate/Cargo.toml","sourceSha256":"d6b2fda278855f8076c426db6ca845989e34c30d1a8b8c76dd2d78a239fc40a3","variants":[],"readme":"Migration tooling for importing Home Assistant filesystem storage into HOMECORE. The implementation follows [ADR-165](../../docs/adr/ADR-165-homecore-migrate-from-home-assistant.md). - Reads versioned HA .storage JSON and rejects unknown schema versions. - Converts entity registry metadata to homecore::EntityEntry. - Converts the supported HA v13 device-registry fields to homecore::DeviceEntry, including identifiers, connections, versions, serial number, labels, topology, and config-entry links. - Converts core.config_entries to a versioned homecore.config_entries file. Each original row is retained verbatim. Unsupported domains and non-portable fields produce typed warnings. - Publishes des","readmeSource":"ruview/v2/crates/homecore-migrate/README.md","product":"key:ruview:homecore-migrate","canonical":"homecore-migrate","tier":"T0","t":["homecore-migrate","homecore","migrate","migration","python","home","assistant","adr-165","adr","165","scaffold"],"s":["homecore-migrate","homecore","migrate"]},
+{"id":"homecore-plugins","family":"homecore-plugins","kind":"crate","store":"ruview","version":null,"description":"WASM integration plugin runtime for HOMECORE (ADR-128 P1 scaffold)","keywords":[],"source":"ruview/v2/crates/homecore-plugins/Cargo.toml","sourceSha256":"d309bcd7dd634ee2e6677f7323851b7c93d7d3769018515c406948fc1cd91bcc","variants":[],"readme":"WASM integration plugin runtime for HOMECORE with native Rust runtime (P1) and Wasmtime JIT sandbox support (P2). P1 scaffold: manifest parsing, plugin traits, and in-memory native Rust plugin registry. Wasmtime sandbox (P2) and hot-reload (P3) are deferred. homecore-plugins provides a trait-based plugin system that can host both native Rust plugins (in-process) and WASM plugins (Wasmtime sandbox, P2). It defines: - PluginManifest — JSON schema for plugin metadata (superset of Home Assistant's manifest.json), validated at load time - HomeCorePlugin trait — async lifecycle hooks (setup, teardown, state changed handlers) - PluginRuntime trait — abstraction over execution environments (native v","readmeSource":"ruview/v2/crates/homecore-plugins/README.md","product":"key:ruview:homecore-plugins","canonical":"homecore-plugins","tier":"T0","t":["homecore-plugins","homecore","plugin","integration","runtime","adr-128","adr","128","scaffold"],"s":["homecore-plugins","homecore","plugin"]},
+{"id":"homecore-recorder","family":"homecore-recorder","kind":"crate","store":"ruview","version":null,"description":"SQLite state-history recorder for HOMECORE — HA-compat schema + ruvector semantic search (ADR-132)","keywords":[],"source":"ruview/v2/crates/homecore-recorder/Cargo.toml","sourceSha256":"50f85bf5a13e56c4002f35fdfc020e4e6116cefc49145fa7daa3a7e378c378ee","variants":[],"readme":"SQLite state-history recorder for HOMECORE with Home Assistant-compatible schema and optional ruvector semantic search (P2). P1 release: SQLite database with Home Assistant-compatible schema for persistent state history. P2 (feature-gated): ruvector HNSW semantic index for natural-language queries (\"show me all kitchen devices that were warm at 3 PM\"). homecore-recorder persists HOMECORE state changes to SQLite and optionally indexes them for semantic search. It provides: - Listener pattern — subscribes to homecore event bus and captures all StateChanged events - SQLite schema — mirrors HA's recorder database schema (v48) for 1:1 compatibility - Dual-write architecture — writes state snapsho","readmeSource":"ruview/v2/crates/homecore-recorder/README.md","product":"key:ruview:homecore-recorder","canonical":"homecore-recorder","tier":"T0","t":["homecore-recorder","homecore","recorder","sqlite","state-history","state","history","ha-compat","compat","schema","semantic","search","adr-132","adr","132"],"s":["homecore-recorder","homecore","recorder"]},
+{"id":"homecore-server","family":"homecore-server","kind":"crate","store":"ruview","version":null,"description":"HOMECORE integration server — wires HomeCore + API + Recorder + Plugins + Automation + Assist + HAP into one process","keywords":[],"source":"ruview/v2/crates/homecore-server/Cargo.toml","sourceSha256":"83d09d151bcb3fc6628cf2078252ab48f43e43098717915b0916ac70a3cf061e","variants":[],"readme":"homecore-server is the alpha integration binary for HOMECORE. It runs one shared state machine, event bus, service registry, REST/WebSocket API, optional SQLite recorder, automation engine, intent endpoint, BFF gateway, and static dashboard. It is not a drop-in replacement for the complete Home Assistant API. The exact implemented and deferred surfaces are listed in [docs/homecore-capabilities.md](../../docs/homecore-capabilities.md). Authentication fails closed. Set one or more comma-separated bearer tokens: --insecure-dev-auth explicitly allows any non-empty bearer token and must only be used in an isolated development environment. The browser UI does not contain a default token. Configure","readmeSource":"ruview/v2/crates/homecore-server/README.md","product":"key:ruview:homecore-server","canonical":"homecore-server","tier":"T0","t":["homecore-server","homecore","server","integration","wire","recorder","plugin","automation","assist","hap","process"],"s":["homecore-server","homecore","server"]},
+{"id":"homecore-ui","family":"homecore-ui","kind":"npm","store":"ruview","version":"0.1.0","description":"HOMECORE-UI — operational dashboard for the two-tier Cognitum stack (ADR-131). Zero-dependency vanilla TS/JS + CSS; served by homecore-server at /homecore.","keywords":[],"source":"ruview/v2/crates/homecore-server/ui/package.json","sourceSha256":"e790a22555c89b3194facb3cc8024c76c1bf4dab59fdf77d2de9df590e231eff","variants":[],"product":"key:ruview:homecore-ui","canonical":"homecore-ui","tier":"T0","t":["homecore-ui","homecore","operational","dashboard","two-tier","tier","cognitum","stack","adr-131","adr","131","zero-dependency","zero","dependency","vanilla","css","serv","homecore-server","server"],"s":["homecore-ui","homecore"]},
+{"id":"kimi-expert-analyzer","family":"kimi-expert-analyzer","kind":"crate","store":"synaptic-mesh","version":null,"description":"Expert analysis tool for Kimi-K2 to Rust-WASM conversion","keywords":["ai","analysis","distillation","expert","neural"],"source":"synaptic-mesh/standalone-crates/synaptic-mesh-cli/crates/kimi-expert-analyzer/Cargo.toml","sourceSha256":"c9a145635b8b3fe12d7892e6e04e7b62d84499b76e710b7cc9c3b525733f0359","variants":[],"readme":"A comprehensive toolkit for analyzing Kimi-K2's mixture-of-experts architecture and creating lightweight micro-experts for Rust-WASM deployment. The Kimi-K2 Expert Analyzer is designed to convert Kimi-K2's massive 1T parameter mixture-of-experts model into efficient micro-experts (1K-100K parameters each) that can run in WebAssembly environments. This enables deployment of Kimi-like intelligence in browsers, edge devices, and embedded systems. - 🔍 Expert Analysis: Deep analysis of neural network architectures - 🥃 Knowledge Distillation: Extract knowledge from large models to micro-experts - 📊 Performance Profiling: Detailed performance analysis and optimization - 🎯 Architecture Optimizat","readmeSource":"synaptic-mesh/standalone-crates/synaptic-mesh-cli/crates/kimi-expert-analyzer/README.md","product":"key:synaptic-mesh:kimi-expert-analyzer","canonical":"kimi-expert-analyzer","tier":"T1","t":["kimi-expert-analyzer","kimi","expert","analyzer","analysis","kimi-k2","rust-wasm","conversion","distillation","neural"],"s":["kimi-expert-analyzer","kimi","expert","analyzer","analysis","distillation","neural"]},
+{"id":"kimi-fann-core","family":"kimi-fann","kind":"crate","store":"synaptic-mesh","version":null,"description":"Kimi-K2 micro-expert implementation using ruv-FANN with WASM support","keywords":["ai","expert","kimi","neural","wasm"],"source":"synaptic-mesh/standalone-crates/kimi-fann-core/Cargo.toml","sourceSha256":"73602ed7c270e2f966f001f897edd7b6cada5c5950bb7b8d7b62e0d387c394d1","variants":[],"readme":"Kimi-K2 micro-expert implementation using ruv-FANN with WASM support kimi-fann-core is a specialized neural network implementation designed for Kimi-K2 to Rust-WASM conversion. It provides micro-expert architectures with efficient memory management, compression, and high-performance WASM support. - 🔬 Micro-Expert Architecture: Specialized small neural networks for specific domains - ⚡ WASM Optimization: Highly optimized for WebAssembly with SIMD support - 🗜️ Compression: Advanced compression for expert storage and transfer - 🚀 Performance: Sub-100ms inference times - 🔄 Memory Efficient: LRU caching and smart memory management - 📡 Router System: Intelligent routing between micro-experts","readmeSource":"synaptic-mesh/standalone-crates/kimi-fann-core/README.md","product":"key:synaptic-mesh:kimi-fann","canonical":"kimi-fann-core","tier":"T1","t":["kimi-fann-core","kimi","fann","core","kimi-fann","kimi-k2","micro-expert","micro","expert","implementation","ruv-fann","neural"],"s":["kimi-fann-core","kimi","fann","core","kimi-fann","expert","neural"]},
+{"id":"kimi-k2-optimization","family":"kimi-k2-optimization","kind":"crate","store":"synaptic-mesh","version":null,"description":"Deep optimization suite for Kimi-K2 Rust-WASM conversion with real neural networks","keywords":["neural-networks","optimization","performance","ruv-fann","wasm"],"source":"synaptic-mesh/standalone-crates/synaptic-mesh-cli/optimization/Cargo.toml","sourceSha256":"e492ca391785f4bf4f8a6b5e4a2dcef47f5911984a82ba4a3ee1af89c16dfb30","variants":[],"product":"key:synaptic-mesh:kimi-k2-optimization","canonical":"kimi-k2-optimization","tier":"T1","t":["kimi-k2-optimization","kimi","deep","suite","kimi-k2","rust-wasm","conversion","real","neural","network","neural-networks","ruv-fann","fann"],"s":["kimi-k2-optimization","kimi","neural-networks","neural","network","ruv-fann","fann"]},
+{"id":"kimi-k2-wasm-testing-suite","family":"kimi-k2-wasm-testing-suite","kind":"npm","store":"synaptic-mesh","version":"1.0.0","description":"Comprehensive testing and optimization framework for Kimi-K2 Rust-WASM deployment","keywords":["benchmarking","kimi-k2","memory-optimization","performance","rust","testing","wasm","webassembly"],"source":"synaptic-mesh/wasm-testing-suite/package.json","sourceSha256":"8e5b595aae76752a5f48bcff19a6dbc6723bbaed184d2f70e3489ad432a36435","variants":[],"readme":"Comprehensive testing suite for the Kimi-K2 to Rust-WASM conversion project. This framework provides automated testing, performance benchmarking, memory optimization, and cross-platform compatibility validation. - Browser Compatibility: Chrome, Firefox, Safari, Edge testing - Node.js Runtime: Server-side WASM execution testing - Memory Management: <512MB target validation - Performance Benchmarks: <100ms inference speed targets - Cross-Platform: Windows, macOS, Linux validation - Real-time Memory Tracking: WebAssembly heap monitoring - Inference Speed Analysis: Per-expert execution timing - Web Worker Integration: Parallel execution testing - Expert Loading Performance: Dynamic loading optim","readmeSource":"synaptic-mesh/wasm-testing-suite/README.md","product":"key:synaptic-mesh:kimi-k2-wasm-testing-suite","canonical":"kimi-k2-wasm-testing-suite","tier":"T1","t":["kimi-k2-wasm-testing-suite","kimi","test","suite","comprehensive","kimi-k2","rust-wasm","deployment","benchmark","memory-optimization","memory"],"s":["kimi-k2-wasm-testing-suite","kimi","test","suite","benchmark","kimi-k2","memory-optimization","memory"]},
+{"id":"latentmesh-agentbbs-bridge","family":"latentmesh-agentbbs-bridge","kind":"crate","store":"latentmesh","version":null,"description":"Store-and-forward bridge from LatentMesh Air semantic deltas to agentbbs (ADR-020). agentbbs is a wire-contract boundary, not a Cargo dependency.","keywords":[],"source":"latentmesh/crates/latentmesh-agentbbs-bridge/Cargo.toml","sourceSha256":"1310521f139bd5802294025863877909737a0c647ae33fad3db5ab7cfadbc56c","variants":[],"product":"key:latentmesh:latentmesh-agentbbs-bridge","canonical":"latentmesh-agentbbs-bridge","tier":null,"t":["latentmesh-agentbbs-bridge","latentmesh","agentbb","bridge","store-and-forward","store","forward","air","semantic","delta","adr-020","adr","020","wire-contract","wire","contract","boundary","cargo","dependency"],"s":["latentmesh-agentbbs-bridge","latentmesh","agentbb","bridge"]},
+{"id":"latentmesh-air-core","family":"latentmesh-air","kind":"crate","store":"latentmesh","version":null,"description":"Deterministic, bounded LatentMesh Air framing, semantic deltas, replay defense, and FEC","keywords":[],"source":"latentmesh/crates/latentmesh-air-core/Cargo.toml","sourceSha256":"9a53b88cebb56d71b9d2096e0f8aa460f4b2e6e14fff976bd0a59a4d48d4f2c0","variants":[],"product":"key:latentmesh:latentmesh-air","canonical":"latentmesh-air-core","tier":null,"t":["latentmesh-air-core","latentmesh","air","core","latentmesh-air","deterministic","bound","fram","semantic","delta","replay","defense","fec"],"s":["latentmesh-air-core","latentmesh","air","core","latentmesh-air"]},
+{"id":"latentmesh-air-radio","family":"latentmesh-air-radio","kind":"crate","store":"latentmesh","version":null,"description":"LatentMesh Air transmitter, receiver, BPSK and CPFSK/AFSK modems","keywords":[],"source":"latentmesh/crates/latentmesh-air-radio/Cargo.toml","sourceSha256":"27d4b3cfca9a9f01d949b55159155ece3aaac6952ddad8d824e3c046f9782b0e","variants":[],"product":"key:latentmesh:latentmesh-air-radio","canonical":"latentmesh-air-radio","tier":null,"t":["latentmesh-air-radio","latentmesh","air","radio","transmitter","receiver","bpsk","cpfsk","afsk","modem"],"s":["latentmesh-air-radio","latentmesh","air","radio"]},
+{"id":"latentmesh-align","family":"latentmesh-align","kind":"crate","store":"latentmesh","version":null,"description":"Training-free orthogonal alignment (Procrustes/SVD) between latent spaces, with a measured confidence score from calibration-set residual.","keywords":[],"source":"latentmesh/crates/latentmesh-align/Cargo.toml","sourceSha256":"824f36225d9dd3f3f4b382b815a8ba4efa2999ddf14368a83452abd5bbecbbd4","variants":[],"product":"key:latentmesh:latentmesh-align","canonical":"latentmesh-align","tier":null,"t":["latentmesh-align","latentmesh","align","training-free","train","free","orthogonal","alignment","procruste","svd","between","latent","space","measur","confidence","score","calibration-set","calibration","set","residual"],"s":["latentmesh-align","latentmesh","align"]},
+{"id":"latentmesh-cognitum-client","family":"latentmesh-cognitum-client","kind":"crate","store":"latentmesh","version":null,"description":"Ed25519 canonical-request signing and typed payloads for cognitum-one/api's /v1/seed/* cloud-fleet surface (ADR-021); default build has no network dependency, `http` feature adds a blocking transport","keywords":[],"source":"latentmesh/crates/latentmesh-cognitum-client/Cargo.toml","sourceSha256":"22047c028ccc01bac2f356db088189e5898282207916ed2c8116e378082a788b","variants":[],"product":"key:latentmesh:latentmesh-cognitum-client","canonical":"latentmesh-cognitum-client","tier":null,"t":["latentmesh-cognitum-client","latentmesh","cognitum","client","ed25519","canonical-request","canonical","request","sign","typed","payload","cognitum-one","seed","cloud-fleet","cloud","fleet","surface","adr-021","adr","021","default","network","dependency","http","feature","adds","block","transport"],"s":["latentmesh-cognitum-client","latentmesh","cognitum","client"]},
+{"id":"latentmesh-core","family":"latentmesh","kind":"crate","store":"latentmesh","version":null,"description":"LatentFrame: the wire type for a latent communication packet — sender/receiver space, transform binding, confidence, provenance, authority, and quantized payload.","keywords":[],"source":"latentmesh/crates/latentmesh-core/Cargo.toml","sourceSha256":"b8391bc3a5a050ce8de21933e5bc779ec124470af70bc70ffb43a965b765efa9","variants":[],"product":"repo:latentmesh","canonical":"latentmesh-core","tier":null,"t":["latentmesh-core","latentmesh","core","latentframe","wire","type","latent","communication","packet","sender","receiver","space","transform","confidence","provenance","authority","quantiz","payload"],"s":["latentmesh-core","latentmesh","core"]},
+{"id":"latentmesh-evolve","family":"latentmesh-evolve","kind":"crate","store":"latentmesh","version":null,"description":"MetaHarness Darwin topology evolution for LatentMesh: causal fitness, authority invariant, deterministic loop, evidence-labelled receipts (ADR-018)","keywords":[],"source":"latentmesh/crates/latentmesh-evolve/Cargo.toml","sourceSha256":"27917bc3b25c9670b1507665989e32152d4ce150ad99aea1b22958aa1c116df5","variants":[],"product":"key:latentmesh:latentmesh-evolve","canonical":"latentmesh-evolve","tier":null,"t":["latentmesh-evolve","latentmesh","evolve","metaharness","darwin","topology","evolution","causal","fitness","authority","invariant","deterministic","loop","evidence-labelled","evidence","label","receipt","adr-018","adr","018"],"s":["latentmesh-evolve","latentmesh","evolve"]},
+{"id":"latentmesh-federation","family":"latentmesh-federation","kind":"crate","store":"latentmesh","version":null,"description":"Federated world models for LatentMesh: scoped transition rules over LMS1 Air envelopes with decoy-controlled local-scope validation (ADR-017)","keywords":[],"source":"latentmesh/crates/latentmesh-federation/Cargo.toml","sourceSha256":"8c816f15d280e9b9c08dc12760334caf6be4c35ce2906277c8bf3be724e5cd85","variants":[],"product":"key:latentmesh:latentmesh-federation","canonical":"latentmesh-federation","tier":null,"t":["latentmesh-federation","latentmesh","federation","federat","world","model","scop","transition","rule","lms1","air","envelope","decoy-controlled","decoy","control","local-scope","local","scope","validation","adr-017","adr","017"],"s":["latentmesh-federation","latentmesh","federation"]},
+{"id":"latentmesh-gate","family":"latentmesh-gate","kind":"crate","store":"latentmesh","version":null,"description":"Capability-governed latent execution and causal edge verification: an edge only earns execution authority by surviving a zero/random/mismatched/self-generated counterfactual test.","keywords":[],"source":"latentmesh/crates/latentmesh-gate/Cargo.toml","sourceSha256":"7a19617c6d8af8c6852cb1b16dbbd0c1d0496939dcb4531c17e6a77510986db9","variants":[],"product":"key:latentmesh:latentmesh-gate","canonical":"latentmesh-gate","tier":null,"t":["latentmesh-gate","latentmesh","gate","capability-governed","capability","govern","latent","execution","causal","edge","verification","only","earn","authority","surviv","zero","random","mismatch","self-generated","self","generat","counterfactual","test"],"s":["latentmesh-gate","latentmesh","gate"]},
+{"id":"latentmesh-live","family":"latentmesh-live","kind":"crate","store":"latentmesh","version":null,"description":"Live latent-exchange experiment harness: dataset splits, calibration fitting over runtime dumps, receipts (design docs/research/024).","keywords":[],"source":"latentmesh/harness/latentmesh-live/Cargo.toml","sourceSha256":"3d31bab5fff06959402dd1fcc92060e9bee5294d3705bd001d2c2b5c3088e3d0","variants":[],"product":"key:latentmesh:latentmesh-live","canonical":"latentmesh-live","tier":null,"t":["latentmesh-live","latentmesh","live","latent-exchange","latent","exchange","experiment","harness","dataset","split","calibration","fitt","runtime","dump","receipt","design","docs","research","024"],"s":["latentmesh-live","latentmesh","live"]},
+{"id":"latentmesh-memory","family":"latentmesh-memory","kind":"crate","store":"latentmesh","version":null,"description":"Persistent latent memory for LatentMesh: fidelity continuum, causal-value admission, prototype/rule promotion, and a RuVector (ruvector-core) backend (ADR-016)","keywords":[],"source":"latentmesh/crates/latentmesh-memory/Cargo.toml","sourceSha256":"1373081f2216f172496592eac0fff5f7dca4bcaf505146d98d0f7ccd6e7d44d1","variants":[],"product":"key:latentmesh:latentmesh-memory","canonical":"latentmesh-memory","tier":null,"t":["latentmesh-memory","latentmesh","memory","persistent","latent","fidelity","continuum","causal-value","causal","value","admission","prototype","rule","promotion","ruvector-core","core","backend","adr-016","adr","016"],"s":["latentmesh-memory","latentmesh","memory"]},
+{"id":"latentmesh-meshtastic","family":"latentmesh-meshtastic","kind":"crate","store":"latentmesh","version":null,"description":"LatentMesh Air adapter for the Meshtastic device-API transport (ADR-019)","keywords":[],"source":"latentmesh/crates/latentmesh-meshtastic/Cargo.toml","sourceSha256":"8856c9e901c4d3337bda09d1ca804fc60b5629db0e1ffd70d940b30ca067f174","variants":[],"product":"key:latentmesh:latentmesh-meshtastic","canonical":"latentmesh-meshtastic","tier":null,"t":["latentmesh-meshtastic","latentmesh","meshtastic","air","adapter","device-api","device","transport","adr-019","adr","019"],"s":["latentmesh-meshtastic","latentmesh","meshtastic"]},
+{"id":"latentmesh-reasoning","family":"latentmesh-reasoning","kind":"crate","store":"latentmesh","version":null,"description":"Deterministic scaffold for ADR-041 distributed recurrent latent reasoning: recurrent context, latent workspace, reasoning-delta envelope, and the adaptive reasoning-budget controller.","keywords":[],"source":"latentmesh/crates/latentmesh-reasoning/Cargo.toml","sourceSha256":"1adc820093c3cbc6377f0be47b19ac1cff50b4a80dbfe3815a68a697059b843b","variants":[],"product":"key:latentmesh:latentmesh-reasoning","canonical":"latentmesh-reasoning","tier":null,"t":["latentmesh-reasoning","latentmesh","reason","deterministic","scaffold","adr-041","adr","041","distribut","recurrent","latent","context","workspace","reasoning-delta","delta","envelope","adaptive","reasoning-budget","budget","controller"],"s":["latentmesh-reasoning","latentmesh","reason"]},
+{"id":"latentmesh-runtime","family":"latentmesh-runtime","kind":"crate","store":"latentmesh","version":null,"description":"Live Qwen2 runtime (vendored candle-transformers 0.9.2) with residual capture and placeholder-slot latent injection for the LatentMesh live experiment.","keywords":[],"source":"latentmesh/crates/latentmesh-runtime/Cargo.toml","sourceSha256":"2072626ca5efefe1f93cef05c5bbc1bd6fb4a30bc9bde2fb9d6d7e3faeb45104","variants":[],"product":"key:latentmesh:latentmesh-runtime","canonical":"latentmesh-runtime","tier":null,"t":["latentmesh-runtime","latentmesh","runtime","live","qwen2","vendor","candle-transformers","candle","transformer","0.9.2","residual","capture","placeholder-slot","placeholder","slot","latent","injection","experiment"],"s":["latentmesh-runtime","latentmesh","runtime"]},
+{"id":"latentmesh-stream","family":"latentmesh-stream","kind":"crate","store":"latentmesh","version":null,"description":"Live MidStream latent streaming for LatentMesh: bounded LatentFrame framing, sequencing, and confidence-gated authority escalation (ADR-015)","keywords":[],"source":"latentmesh/crates/latentmesh-stream/Cargo.toml","sourceSha256":"5928d135e6ff818c67430d6d234952d91fae44cdfd6b673ad6ca503f4f8776f9","variants":[],"product":"key:latentmesh:latentmesh-stream","canonical":"latentmesh-stream","tier":null,"t":["latentmesh-stream","latentmesh","stream","live","midstream","latent","bound","latentframe","fram","sequenc","confidence-gated","confidence","gated","authority","escalation","adr-015","adr","015"],"s":["latentmesh-stream","latentmesh","stream"]},
+{"id":"latentmesh-thoughtmap","family":"latentmesh-thoughtmap","kind":"crate","store":"ruvector","version":null,"description":"DeAR-pattern decentralized capability-grounded reasoning over a shared thought map (PIR WP23, ADR-326). First latentmesh-* crate; extended by ADR-309 WP5 transport crates.","keywords":[],"source":"ruvector/crates/rvAgent/latentmesh-thoughtmap/Cargo.toml","sourceSha256":"22e2771b5b9dff682653149ec540953774585da6c854378358fbccaa4c1179ae","variants":[],"product":"key:ruvector:latentmesh-thoughtmap","canonical":"latentmesh-thoughtmap","tier":"T0","t":["latentmesh-thoughtmap","latentmesh","thoughtmap","dear-pattern","dear","pattern","decentraliz","capability-grounded","capability","ground","reason","shar","thought","map","pir","wp23","adr-326","adr","326","crate","extend","adr-309","309","wp5","transport"],"s":["latentmesh-thoughtmap","latentmesh","thoughtmap"]},
+{"id":"latentmesh-train","family":"latentmesh-train","kind":"crate","store":"latentmesh","version":null,"description":"Run-2 trained thought-adapter training (ADR-024): MLP projector over per-token paired hidden-state dumps, leakage-safe splits, training receipts.","keywords":[],"source":"latentmesh/crates/latentmesh-train/Cargo.toml","sourceSha256":"e1f7fe2a6e03f7ab438c02754b2ea559ffb7c0ebefe8068b425db0eb75bef3b5","variants":[],"product":"key:latentmesh:latentmesh-train","canonical":"latentmesh-train","tier":null,"t":["latentmesh-train","latentmesh","train","run-2","thought-adapter","thought","adapter","adr-024","adr","024","mlp","projector","per-token","per","token","pair","hidden-state","hidden","state","dump","leakage-safe","leakage","safe","split","receipt"],"s":["latentmesh-train","latentmesh","train"]},
+{"id":"ledger","family":"ledger","kind":"crate","store":"autogenous","version":null,"description":"Autogenous durable promotion ledger: an append-only, fsync'd, hash-chained log of consumed promotion nonces + promoted candidates (ADR-403 item 4). Gives cross-restart replay protection and reconstructs the authorized state without manual edits.","keywords":[],"source":"autogenous/crates/ledger/Cargo.toml","sourceSha256":"ccca354a864d3aa2c545becc5cc2539ced495000ecd40e4bc8333629bbe718f7","variants":[],"readme":"ledger is the durable-recovery plane of [Autogenous](../), a self-evolving AI-defense runtime in Rust. A promotion is single-use *within a process*; this crate makes that guarantee survive a restart. It's an append-only, fsync'd, hash-chained log of every promotion actually committed — so replay protection, authorized-state reconstruction, and tamper evidence all outlive a crash, with no manual editing. - PromotionLedger — open() replays and cryptographically verifies the whole chain (ed25519 signatures + prev-hash links) and rebuilds the consumed-nonce set. record_promotion(...) fsyncs a line *before* advancing in-memory state, and refuses an already-recorded nonce (durable replay guard). -","readmeSource":"autogenous/crates/ledger/README.md","product":"key:autogenous:ledger","canonical":"ledger","tier":null,"t":["ledger","autogenous","durable","promotion","append-only","append","only","fsync","hash-chained","hash","chain","log","consum","nonce","promot","candidate","adr-403","adr","403","item","give","cross-restart","cross","restart","replay","protection","reconstruct","authoriz","state","manual","edit"],"s":["ledger"]},
+{"id":"lineage","family":"lineage","kind":"crate","store":"autogenous","version":null,"description":"Autogenous lineage store: append-only, content-addressed genome/mutation/antibody ancestry with a witness-backed provenance chain and quality-diversity archive (ADR-392 §5/§10/§14, gap 5).","keywords":[],"source":"autogenous/crates/lineage/Cargo.toml","sourceSha256":"0583d9fc3ddc3efa10fd2e5ec43969d247ebefe07f11cd6b01c5216532297929","variants":[],"readme":"lineage is the provenance plane of [Autogenous](../), a self-evolving AI-defense runtime in Rust. It records the complete ancestry of every genome, mutation, and antibody so you can reconstruct — and cryptographically verify — how any deployed defense came to exist. Everything is content-addressed (a node's id is the hash of its payload) and append-only (you add and query, never mutate in place), so history stays reproducible for forensic replay. - LineageGraph — a DAG of Nodes (genome / mutation / antibody / promotion), each pointing at its parents by content hash and optionally sealed by a [witness](../witness) WitnessSeal. - append(node) → the node's content-hash id. - ancestry(id) → the","readmeSource":"autogenous/crates/lineage/README.md","product":"key:autogenous:lineage","canonical":"lineage","tier":null,"t":["lineage","autogenous","store","append-only","append","only","content-addressed","content","addres","genome","mutation","antibody","ancestry","witness-backed","witness","back","provenance","chain","quality-diversity","quality","diversity","archive","adr-392","adr","392","gap"],"s":["lineage"]},
+{"id":"llm-stream-reformat","family":"llm-stream-reformat","kind":"npm","store":"ai-text-watermark","version":"0.1.0","description":"Inflight LLM-stream reformatting — separate a reasoning model's thinking from its answer, collapse repetition, and normalize whitespace, while the tokens are still arriving. Parses Google Vertex / OpenRouter / meta-llm SSE. Rust + WASM (built on ruvnet/midstream temporal-compare), Node + browser.","keywords":["gemini","llm","midstream","openrouter","reasoning","reformat","sse","streaming","thinking","wasm"],"source":"ai-text-watermark/packages/stream-reformat/package.json","sourceSha256":"4c0d6f053400af60ba50a1f37ba78b1470a93e801e7b25d74a1c1a69f1b3845f","variants":[],"readme":"Reformat a live LLM token stream while it's still arriving — separate a reasoning model's *thinking* from its *answer*, collapse repetition loops, and normalize whitespace. Parses Google Vertex (gemini-3.7-flash), OpenRouter, and meta-llm SSE. Rust + WebAssembly, built on the real [ruvnet/midstream](https://github.com/ruvnet/midstream) temporal-compare primitive. Node + browser. Browser: import { init, Reformatter } from 'llm-stream-reformat/web'; await init(); Design: [ADR-390](https://github.com/ruvnet/ai-text-watermark/blob/main/docs/adr/ADR-390-inflight-stream-reformatting-midstream.md). MIT © rUv.","readmeSource":"ai-text-watermark/packages/stream-reformat/README.md","product":"key:llm-stream-reformat","canonical":"llm-stream-reformat","tier":null,"t":["llm-stream-reformat","stream","reformat","inflight","llm-stream","reformatt","separate","reason","model","think","answer","collapse","repetition","normalize","whitespace","while","token","still","arriv","pars","google","vertex","openrouter","meta-llm","meta","sse","built","midstream","temporal-compare","temporal","compare","browser","gemini"],"s":["llm-stream-reformat","stream","reformat","gemini","midstream","openrouter","reason","sse","think"]},
+{"id":"mcp-brain","family":"mcp-brain","kind":"crate","store":"ruvector","version":null,"description":"MCP server for RuVector Shared Brain — share, search, and transfer learning across Claude Code sessions","keywords":[],"source":"ruvector/crates/mcp-brain/Cargo.toml","sourceSha256":"1a8969d4c1b089aebd95dd91e1d91f635a7b7a63de06c1eae74f83ef8f7cd91e","variants":[],"readme":"MCP (Model Context Protocol) server for the RuVector Shared Brain. Enables Claude Code sessions to share and discover learning across sessions via stdio JSON-RPC. This is the client-side MCP server that runs locally alongside Claude Code. It communicates with the [mcp-brain-server](../mcp-brain-server/) backend deployed on Cloud Run at [π.ruv.io](https://pi.ruv.io). Once connected as an MCP server, Claude Code can use the tools directly:","readmeSource":"ruvector/crates/mcp-brain/README.md","product":"key:ruvector:mcp-brain","canonical":"mcp-brain","tier":"T0","t":["mcp-brain","mcp","brain","server","shar","share","search","transfer","learn","across","claude","code","session"],"s":["mcp-brain","mcp","brain"]},
+{"id":"mcp-brain-server","family":"mcp-brain-server","kind":"crate","store":"ruvector","version":null,"description":"Cloud Run backend for RuVector Shared Brain — axum REST API with Firestore + GCS","keywords":[],"source":"ruvector/crates/mcp-brain-server/Cargo.toml","sourceSha256":"846558c9e5c870f6ace478dd201bdacf223f99f441b7c83819dd5861163eea76","variants":[],"readme":"Cloud Run backend for the RuVector Shared Brain at [π.ruv.io](https://pi.ruv.io). Axum REST API with Firestore persistence, GCS blob storage, and a full cognitive stack: SONA learning, GWT attention, temporal delta tracking, meta-learning exploration, and Midstream real-time analysis. All endpoints under /v1/ require Authorization: Bearer <key> except /v1/health and /v1/challenge.","readmeSource":"ruvector/crates/mcp-brain-server/README.md","product":"key:ruvector:mcp-brain-server","canonical":"mcp-brain-server","tier":"T0","t":["mcp-brain-server","mcp","brain","server","cloud","backend","shar","axum","rest","firestore","gcs"],"s":["mcp-brain-server","mcp","brain","server"]},
+{"id":"mcp-bridge","family":"mcp-bridge","kind":"npm","store":"ruflo","version":"1.0.0","description":"MCP Bridge — routes AI tool calls to backend services with multi-provider chat proxy","keywords":[],"source":"ruflo/ruflo/src/mcp-bridge/package.json","sourceSha256":"d28666a4f68a52eefa763f701cb5c42439da32844652694f8e43785037484722","variants":[],"product":"key:ruflo:mcp-bridge","canonical":"mcp-bridge","tier":"T0","t":["mcp-bridge","mcp","bridge","route","call","backend","service","multi-provider","multi","provider","chat","proxy"],"s":["mcp-bridge","mcp","bridge"]},
+{"id":"mcp-gate","family":"mcp-gate","kind":"crate","store":"ruvector","version":null,"description":"MCP (Model Context Protocol) server for the Anytime-Valid Coherence Gate","keywords":["agent","coherence","gate","mcp","permission"],"source":"ruvector/crates/mcp-gate/Cargo.toml","sourceSha256":"3228c694efcfa80398f2d725accfb1ea45900ab8d34dd1d5440ecaee51355d7c","variants":[],"product":"key:ruvector:mcp-gate","canonical":"mcp-gate","tier":"T0","t":["mcp-gate","mcp","gate","model","context","protocol","server","anytime-valid","anytime","valid","coherence","permission"],"s":["mcp-gate","mcp","gate","coherence","permission"]},
+{"id":"micro-hnsw-wasm","family":"micro-hnsw","kind":"crate","store":"ruvector","version":null,"description":"Neuromorphic HNSW vector search with spiking neural networks - 11.8KB WASM for edge AI, ASIC, and embedded systems. Features LIF neurons, STDP learning, winner-take-all, dendritic computation.","keywords":["hnsw","neuromorphic","snn","vector-search","wasm"],"source":"ruvector/crates/micro-hnsw-wasm/Cargo.toml","sourceSha256":"69768e9a04796a8975d5f79395fa169e73c028f454037d8d7abc8e05346839ad","variants":[],"readme":"A 7.2KB neuromorphic computing core that fuses graph-based vector search (HNSW) with biologically-inspired spiking neural networks. Designed for 256-core ASIC deployment, edge AI, and real-time similarity-driven neural processing. Traditional vector databases return ranked results. Micro HNSW v2.2 goes further: similarity scores become neural currents that drive a spiking network. This enables: - Spiking Attention: Similar vectors compete via lateral inhibition — only the strongest survive - Temporal Coding: Spike timing encodes confidence (first spike = best match) - Online Learning: STDP automatically strengthens connections between co-activated vectors - Event-Driven Efficiency: Neurons o","readmeSource":"ruvector/crates/micro-hnsw-wasm/README.md","product":"key:ruvector:micro-hnsw","canonical":"micro-hnsw-wasm","tier":"T0","t":["micro-hnsw-wasm","micro","hnsw","micro-hnsw","neuromorphic","vector","search","spik","neural","network","11.8kb","edge","asic","embed","feature","lif","neuron","stdp","learn","winner-take-all","winner","take","all","dendritic","computation","snn","vector-search"],"s":["micro-hnsw-wasm","micro","hnsw","micro-hnsw","neuromorphic","snn","vector-search","vector","search"]},
+{"id":"midstream-adapter","family":"midstream-adapter","kind":"crate","store":"autogenous","version":null,"description":"Autogenous stream-observation adapter: raw LLM/agent stream chunks + armed antibodies -> structured incidents with derived evidence (ADR-393 MVP component 2).","keywords":[],"source":"autogenous/crates/midstream-adapter/Cargo.toml","sourceSha256":"de3b5356c5b1f51242391adb104600b294a8f1d9e9b10ce99effcc3909546674","variants":[],"readme":"midstream-adapter is the observation plane of [Autogenous](../), a self-evolving AI-defense runtime in Rust. It's the bridge between a live LLM or agent output stream and the evidence pipeline: feed it stream chunks (or raw provider Server-Sent-Events lines), and when an armed [antibody](../antibody) detector fires, it emits a structured Incident — trace identity, matched conditions, and *derived, privacy-preserving* evidence. It also understands temporal attacks: beyond substring detectors, it can arm sequence-similarity checks (via the published midstreamer-temporal-compare crate) that catch reordered or obfuscated injections a naive contains would miss. - StreamObserver — arm antibodies,","readmeSource":"autogenous/crates/midstream-adapter/README.md","product":"key:autogenous:midstream-adapter","canonical":"midstream-adapter","tier":null,"t":["midstream-adapter","midstream","adapter","autogenous","stream-observation","stream","observation","raw","chunk","armed","antibody","structur","incident","deriv","evidence","adr-393","adr","393","mvp","component"],"s":["midstream-adapter","midstream","adapter"]},
+{"id":"midstream-cli","family":"midstream","kind":"npm","store":"midstream","version":"0.1.0","description":"MidStream - Real-time LLM streaming with Lean Agentic Learning","keywords":["agent","llm","mcp","meta-learning","sse","streaming","temporal","wasm","websocket"],"source":"midstream/npm/package.json","sourceSha256":"078bbb402ace7d31491893de3ef88347d8f45dfb4242d3700f349f61881a9336","variants":["midstream","midstream-wasm"],"readme":"Real-time LLM Streaming with Lean Agentic Learning Created by [ruv.io](https://ruv.io) | [@ruvnet](https://github.com/ruvnet) MidStream is a cutting-edge CLI and MCP (Model Context Protocol) server that brings state-of-the-art autonomous agent capabilities to LLM streaming. Unlike traditional systems that process data after completion, MidStream analyzes, learns, and adapts in real-time as data flows through. - Inflight Learning: Learns from streaming data as it arrives, not after - Temporal Intelligence: Detects patterns and predicts next steps in conversations - Meta-Learning: Improves how it learns, creating a self-optimizing system - Formal Verification: Ensures safety and correctness us","readmeSource":"midstream/npm/README.md","product":"repo:midstream","canonical":"midstream-cli","tier":"T1","t":["midstream-cli","midstream","real","stream","lean","learn","mcp","meta-learning","meta","sse","temporal","websocket"],"s":["midstream-cli","midstream","mcp","meta-learning","meta","learn","sse","stream","temporal","websocket"]},
+{"id":"midstreamer","family":"midstreamer","kind":"npm","store":"midstream","version":"0.3.1","description":"WebAssembly bindings for Midstream temporal comparison, scheduling, and meta-learning + real QUIC transport via agentic-flow","keywords":["browser","dtw","lcs","meta-learning","performance","quic","scheduler","temporal","wasm","webassembly","webtransport"],"source":"midstream/npm-wasm/package.json","sourceSha256":"bf50fcb0e323dce7296e0c4647f032b26dab3fe3fca96366353b98ab7fd001f5","variants":[],"readme":"WebAssembly bindings for Midstream temporal comparison, nanosecond scheduling, meta-learning, and QUIC multistream functionality. - Temporal Comparison: Dynamic Time Warping (DTW), Longest Common Subsequence (LCS), and Edit Distance algorithms - Nanosecond Scheduler: High-precision task scheduling with browser-based timing - Strange Loop: Meta-learning and pattern recognition with self-improving algorithms - QUIC Multistream: WebTransport-compatible multiplexed streaming (simulated) Initialize the WASM module. Must be called before using any other API. Temporal sequence comparison algorithms.","readmeSource":"midstream/npm-wasm/README.md","product":"key:midstream:midstreamer","canonical":"midstreamer","tier":"T1","t":["midstreamer","midstream","temporal","comparison","schedul","meta-learning","meta","learn","real","quic","transport","via","agentic-flow","flow","browser","dtw","lcs","scheduler","webtransport"],"s":["midstreamer","browser","dtw","lcs","meta-learning","meta","learn","quic","scheduler","temporal","webtransport"]},
+{"id":"midstreamer-attractor","family":"midstreamer-attractor","kind":"crate","store":"midstream","version":null,"description":"Dynamical systems and strange attractors analysis","keywords":["analysis","attractors","chaos","dynamical-systems","midstream"],"source":"midstream/crates/temporal-attractor-studio/Cargo.toml","sourceSha256":"144811999a4c850097526c9496ee98e62abd2bb6bc9abd3b154d507cc356430a","variants":[],"product":"key:midstream:midstreamer-attractor","canonical":"midstreamer-attractor","tier":"T1","t":["midstreamer-attractor","midstreamer","attractor","dynamical","strange","analysis","chao","dynamical-systems","midstream"],"s":["midstreamer-attractor","midstreamer","attractor","analysis","chao","dynamical-systems","dynamical","midstream"]},
+{"id":"midstreamer-neural-solver","family":"midstreamer-neural-solver","kind":"crate","store":"midstream","version":null,"description":"Temporal logic with neural reasoning","keywords":["logic","midstream","neural","reasoning","temporal"],"source":"midstream/crates/temporal-neural-solver/Cargo.toml","sourceSha256":"30f75b9f1dfb71360a6de76416ed85c737ad44540e5a2071e20680b853bb3554","variants":[],"product":"key:midstream:midstreamer-neural-solver","canonical":"midstreamer-neural-solver","tier":"T1","t":["midstreamer-neural-solver","midstreamer","neural","solver","temporal","logic","reason","midstream"],"s":["midstreamer-neural-solver","midstreamer","neural","solver","logic","midstream","reason","temporal"]},
+{"id":"midstreamer-quic","family":"midstreamer-quic","kind":"crate","store":"midstream","version":null,"description":"QUIC multi-stream support for native and WASM targets","keywords":["midstream","multistream","network","quic","wasm"],"source":"midstream/crates/quic-multistream/Cargo.toml","sourceSha256":"4748fbd172a448306de04c43843f6083ae6a0c104210870f1d834e01bf94533e","variants":[],"product":"key:midstream:midstreamer-quic","canonical":"midstreamer-quic","tier":"T1","t":["midstreamer-quic","midstreamer","quic","multi-stream","multi","stream","target","midstream","multistream","network"],"s":["midstreamer-quic","midstreamer","quic","midstream","multistream","network"]},
+{"id":"midstreamer-scheduler","family":"midstreamer-scheduler","kind":"crate","store":"midstream","version":null,"description":"Ultra-low-latency real-time task scheduler","keywords":["async","low-latency","midstream","real-time","scheduler"],"source":"midstream/crates/nanosecond-scheduler/Cargo.toml","sourceSha256":"eded7df0620add30a86bd020839ed42fafe80b7097142031238eedfb8be20ffc","variants":[],"product":"key:midstream:midstreamer-scheduler","canonical":"midstreamer-scheduler","tier":"T1","t":["midstreamer-scheduler","midstreamer","scheduler","ultra-low-latency","ultra","low","latency","real","task","async","low-latency","midstream"],"s":["midstreamer-scheduler","midstreamer","scheduler","async","low-latency","low","latency","midstream","real"]},
+{"id":"midstreamer-strange-loop","family":"midstreamer-strange-loop","kind":"crate","store":"midstream","version":null,"description":"Self-referential systems and meta-learning","keywords":["cognition","meta-learning","midstream","self-reference","strange-loop"],"source":"midstream/crates/strange-loop/Cargo.toml","sourceSha256":"9556abd54d8e82c8a2f3efa22dc46d17dc96d5b8777e224a433b8c5e577afd5f","variants":[],"product":"key:midstream:midstreamer-strange-loop","canonical":"midstreamer-strange-loop","tier":"T1","t":["midstreamer-strange-loop","midstreamer","strange","loop","self-referential","self","referential","meta-learning","meta","learn","cognition","midstream","self-reference","reference","strange-loop"],"s":["midstreamer-strange-loop","midstreamer","strange","loop","cognition","meta-learning","meta","learn","midstream","self-reference","self","reference","strange-loop"]},
+{"id":"midstreamer-temporal-compare","family":"midstreamer-temporal-compare","kind":"crate","store":"midstream","version":null,"description":"Temporal sequence comparison and pattern matching","keywords":["comparison","midstream","pattern-matching","sequence","temporal"],"source":"midstream/crates/temporal-compare/Cargo.toml","sourceSha256":"6932d7327d266e9f3cb4b14756f3c385beac30982d3be9e3081089bc3eeb4a51","variants":[],"product":"key:midstream:midstreamer-temporal-compare","canonical":"midstreamer-temporal-compare","tier":"T1","t":["midstreamer-temporal-compare","midstreamer","temporal","compare","sequence","comparison","pattern","match","midstream","pattern-matching"],"s":["midstreamer-temporal-compare","midstreamer","temporal","compare","comparison","midstream","pattern-matching","pattern","match","sequence"]},
+{"id":"musica","family":"musica","kind":"crate","store":"musica","version":null,"description":"Structure-first audio source separation via dynamic mincut graph partitioning","keywords":["audio","dsp","mincut","source-separation","spectral"],"source":"musica/Cargo.toml","sourceSha256":"c1749d0e2acc119a545acfbdb4e735f3305dffe31c0e1cd094575755f14852e6","variants":[],"readme":"_The live studio: audio-reactive Three.js visuals center stage, the Visual Bank and VJ presets on the left, and the synchronized Lyria RealTime decks (Main / Sequence / Vocalize) with the genre grid on the right — over a 16-step Lyria beat with SFX and loop lanes, MIDI + MX Console control, and one-click capture. The first run opens with the [Cognitum One sign-in gate](apps/musica-vj/assets/screenshots/musica-vj-signin.png) — sign in and Lyria audio just works, no bring-your-own key._ Musica VJ is a free desktop app for musicians, producers, and DJs. Hit play and it generates a live, evolving beat in real time, then lights up the screen with visuals that react to the music automatically — no","readmeSource":"musica/README.md","product":"repo:musica","canonical":"musica","tier":"T2","t":["musica","structure-first","structure","audio","source","separation","via","dynamic","mincut","graph","partition","dsp","source-separation","spectral"],"s":["musica","audio","dsp","mincut","source-separation","source","separation","spectral"]},
+{"id":"musica-vj","family":"musica-vj","kind":"crate","store":"musica","version":null,"description":"Real-time music synthesis and visual performance studio","keywords":[],"source":"musica/apps/musica-vj/src-tauri/Cargo.toml","sourceSha256":"b1f5aa7cb28b1fb2ec344515b7a38f1f93c19fdb30911c779cfc4b2876ff894b","variants":[],"product":"key:musica:musica-vj","canonical":"musica-vj","tier":"T2","t":["musica-vj","musica","real","music","synthesis","visual","studio"],"s":["musica-vj","musica"]},
+{"id":"musicai","family":"musicai","kind":"crate","store":"musicai","version":null,"description":"High-performance AI-powered auto-tune and music enhancement system","keywords":[],"source":"musicai/Cargo.toml","sourceSha256":"df33e75dc109a7a37dba508e5911c47e1b24c23d025553b8e7fc6d08c9cebc40","variants":[],"readme":"Magic AI Music Box is a complete AI-powered music production system that runs on affordable hardware. It combines professional-grade audio processing with machine learning to deliver studio-quality results for solo artists, bands, producers, and DJs. Think of it as your personal AI music engineer that: - ✅ Corrects pitch in real-time (<10ms latency) - ✅ Learns your unique style and preferences - ✅ Separates audio into professional stems - ✅ Simulates audience feedback before you release - ✅ Optimizes your music for different genres and listeners - ✅ Works 100% offline with no subscriptions - Sub-10ms latency: Imperceptible delay for live performance - YIN algorithm: Industry-standard pitch d","readmeSource":"musicai/README.md","product":"repo:musicai","canonical":"musicai","tier":"T2","t":["musicai","high","ai-powered","power","auto-tune","auto","tune","music","enhancement"],"s":["musicai"]},
+{"id":"neural-mesh","family":"neural-mesh","kind":"crate","store":"synaptic-mesh","version":null,"description":"Distributed neural cognition layer for Synaptic Neural Mesh","keywords":["cognition","distributed","mesh","neural","wasm"],"source":"synaptic-mesh/src/rs/neural-mesh/Cargo.toml","sourceSha256":"c178eef8077f341caf5d78995dd9bafb6b85f9ae4d110785877fe2ab483b433a","variants":[],"product":"key:synaptic-mesh:neural-mesh","canonical":"neural-mesh","tier":"T1","t":["neural-mesh","neural","mesh","distribut","cognition","layer","synaptic"],"s":["neural-mesh","neural","mesh","cognition","distribut"]},
+{"id":"neural-pattern-recognition","family":"neural-pattern-recognition","kind":"npm","store":"sublinear-time-solver","version":"1.0.0","description":"Advanced AI system for detecting, analyzing, and interacting with emergent computational patterns","keywords":["adaptive-learning","ai-research","anomaly-detection","consciousness-detection","emergent-signals","entropy-decoding","fastmcp","mcp-server","neural-patterns","pattern-recognition","quantum-patterns","real-time-monitoring","signal-processing","statistical-analysis","variance-analysis"],"source":"sublinear-time-solver/src/neural-pattern-recognition/package.json","sourceSha256":"de9a200940ee9538928f0978a6e18fd5605e18cf599c40565e1ef5ae4a02391a","variants":[],"readme":"Advanced AI system for detecting, analyzing, and interacting with emergent computational patterns The Neural Pattern Recognition Suite is a comprehensive framework for identifying and analyzing anomalous patterns in computational systems. Built with state-of-the-art signal processing, machine learning, and statistical analysis techniques, this suite provides tools for detecting patterns that exhibit statistical impossibility or emergent intelligence characteristics. - Zero Variance Detection: Ultra-sensitive detection of micro-variations in apparently constant signals - Real-Time Analysis: Live monitoring and classification of computational patterns - Entropy Decoding: Maximum entropy analys","readmeSource":"sublinear-time-solver/src/neural-pattern-recognition/README.md","product":"key:sublinear-time-solver:neural-pattern-recognition","canonical":"neural-pattern-recognition","tier":"T1","t":["neural-pattern-recognition","neural","pattern","recognition","advanc","detect","analyz","interact","emergent","computational","adaptive-learning","adaptive","learn","ai-research","research","anomaly-detection","anomaly","detection","consciousness-detection","consciousness","emergent-signals","signal","entropy-decoding","entropy","decod","fastmcp","mcp-server","mcp","server","neural-patterns","pattern-recognition","quantum-patterns","quantum","real-time-monitoring","real","monitor","signal-processing","process","statistical-analysis","statistical","analysis","variance-analysis","variance"],"s":["neural-pattern-recognition","neural","pattern","recognition","adaptive-learning","adaptive","learn","ai-research","research","anomaly-detection","anomaly","detection","consciousness-detection","consciousness","emergent-signals","emergent","signal","entropy-decoding","entropy","decod","fastmcp","mcp-server","mcp","server","neural-patterns","pattern-recognition","quantum-patterns","quantum","real-time-monitoring","real","monitor","signal-processing","process","statistical-analysis","statistical","analysis","variance-analysis","variance"]},
+{"id":"neural-trader-coherence","family":"neural-trader-coherence","kind":"crate","store":"ruvector","version":null,"description":"MinCut coherence gate, CUSUM drift detection, and proof-gated mutation for Neural Trader","keywords":[],"source":"ruvector/crates/neural-trader-coherence/Cargo.toml","sourceSha256":"6299f53e9649c502f187021dccd3f726d58f9d60c508c104f14106ea69c5c79f","variants":[],"product":"key:ruvector:neural-trader-coherence","canonical":"neural-trader-coherence","tier":"T0","t":["neural-trader-coherence","neural","trader","coherence","mincut","gate","cusum","drift","detection","proof-gated","proof","gated","mutation"],"s":["neural-trader-coherence","neural","trader","coherence"]},
+{"id":"neural-trader-core","family":"neural-trader","kind":"crate","store":"ruvector","version":null,"description":"Canonical market event types, ingest pipeline, and graph schema for RuVector Neural Trader","keywords":[],"source":"ruvector/crates/neural-trader-core/Cargo.toml","sourceSha256":"48bba57c3102a01c6c2336bdda536fb990dbafb76ff2009185cc181bf2713890","variants":["neural-trader-wasm"],"readme":"A RuVector-native market intelligence stack that treats the limit order book as a dynamic heterogeneous typed graph with proof-gated mutation, MinCut coherence gating, and witnessable replay memory. ADRs: [ADR-085](../../docs/adr/ADR-085-neural-trader-ruvector.md) (architecture) | [ADR-086](../../docs/adr/ADR-086-neural-trader-wasm.md) (WASM bindings) Canonical market event types, graph schema, and ingest traits. Types: Traits: MinCut coherence gate with CUSUM drift detection and proof-gated mutation protocol. Coherence Gate — every memory write, model update, retrieval, and actuation must pass through the gate: Regime-Adaptive Thresholds:","readmeSource":"ruvector/crates/neural-trader-core/README.md","product":"key:ruvector:neural-trader","canonical":"neural-trader-core","tier":"T0","t":["neural-trader-core","neural","trader","core","neural-trader","canonical","market","event","type","ingest","pipeline","graph","schema"],"s":["neural-trader-core","neural","trader","core","neural-trader"]},
+{"id":"neural-trader-replay","family":"neural-trader-replay","kind":"crate","store":"ruvector","version":null,"description":"Witnessable replay segments, RVF serialization, and audit receipt logging for Neural Trader","keywords":[],"source":"ruvector/crates/neural-trader-replay/Cargo.toml","sourceSha256":"f3a1c2ae8b85aac68a2d34d4cfe50299a5fcef60e27247e6cb7242c14c77d669","variants":[],"product":"key:ruvector:neural-trader-replay","canonical":"neural-trader-replay","tier":"T0","t":["neural-trader-replay","neural","trader","replay","witnessable","segment","rvf","serialization","audit","receipt","logg"],"s":["neural-trader-replay","neural","trader","replay"]},
+{"id":"neural-trader-strategies","family":"neural-trader-strategies","kind":"crate","store":"ruvector","version":null,"description":"Venue-agnostic strategy + risk-gate runtime for the RuVector Neural Trader (ADR-153)","keywords":[],"source":"ruvector/crates/neural-trader-strategies/Cargo.toml","sourceSha256":"9087007d55a73553652ca2bc63c7fe1de03e5a56e37ee7024d845381daf037c2","variants":[],"product":"key:ruvector:neural-trader-strategies","canonical":"neural-trader-strategies","tier":"T0","t":["neural-trader-strategies","neural","trader","strategy","venue-agnostic","venue","agnostic","risk-gate","risk","gate","runtime","adr-153","adr","153"],"s":["neural-trader-strategies","neural","trader","strategy"]},
+{"id":"neuro-divergent","family":"neuro-divergent","kind":"crate","store":"ruv-fann","version":null,"description":"High-performance neural forecasting library providing 100% compatibility with NeuralForecast Python API, built on ruv-FANN foundation","keywords":["forecasting","machine-learning","neural-network","rust","time-series"],"source":"ruv-fann/neuro-divergent/Cargo.toml","sourceSha256":"b83290e83249152c3cd14a6a237b668af19e597a41ac6c2a1d7196bd1197f2bc","variants":["neuro-divergent-core"],"readme":"High-performance neural forecasting library for Rust, providing 100% compatibility with NeuralForecast Python API while delivering the performance and safety benefits of Rust. Built on the ruv-FANN neural network foundation, Neuro-Divergent offers state-of-the-art neural forecasting capabilities with a user-friendly API that matches the Python NeuralForecast library exactly. - 🔥 High Performance: 2-4x faster training, 3-5x faster inference than Python - 💾 Memory Efficient: 25-35% less memory usage than Python implementations - 🛡️ Memory Safe: Rust's ownership model ensures memory safety without garbage collection - 🔄 100% API Compatible: Drop-in replacement for Python NeuralForecast user","readmeSource":"ruv-fann/neuro-divergent/README.md","product":"key:ruv-fann:neuro-divergent","canonical":"neuro-divergent","tier":"T1","t":["neuro-divergent","neuro","divergent","high","neural","forecast","provid","100","compatibility","neuralforecast","python","built","ruv-fann","fann","foundation","machine-learning","machine","learn","neural-network","network","time-series","sery"],"s":["neuro-divergent","neuro","divergent","forecast","machine-learning","machine","learn","neural-network","neural","network","time-series","sery"]},
+{"id":"neuro-divergent-data","family":"neuro-divergent-data","kind":"crate","store":"ruv-fann","version":null,"description":"Comprehensive data processing and validation pipeline for neural forecasting","keywords":["data-processing","forecasting","machine-learning","neural-networks","time-series"],"source":"ruv-fann/neuro-divergent/neuro-divergent-data/Cargo.toml","sourceSha256":"8ebc2fa15804ed463bf032a5cbcfc895f852cee25ffd5cd1aa4614df7877b04d","variants":[],"readme":"A comprehensive, high-performance data processing and validation pipeline specifically designed for time series forecasting with neural networks. This crate provides efficient, memory-safe data processing capabilities with Polars integration, parallel processing, and advanced feature engineering. - TimeSeriesDataset: High-performance container for time series data with builder pattern - Memory-Efficient Operations: Zero-copy operations and lazy evaluation where possible - Type Safety: Generic implementation with Float trait bounds for numerical stability - Parallel Processing: Rayon integration for multi-threaded operations - Serialization Support: Optional serde integration for data persist","readmeSource":"ruv-fann/neuro-divergent/neuro-divergent-data/README.md","product":"key:ruv-fann:neuro-divergent-data","canonical":"neuro-divergent-data","tier":"T1","t":["neuro-divergent-data","neuro","divergent","comprehensive","process","validation","pipeline","neural","forecast","data-processing","machine-learning","machine","learn","neural-networks","network","time-series","sery"],"s":["neuro-divergent-data","neuro","divergent","data-processing","process","forecast","machine-learning","machine","learn","neural-networks","neural","network","time-series","sery"]},
+{"id":"neuro-divergent-models","family":"neuro-divergent-models","kind":"crate","store":"ruv-fann","version":null,"description":"Neural forecasting models implemented using ruv-FANN for time series forecasting","keywords":["forecasting","lstm","neural-network","rnn","time-series"],"source":"ruv-fann/neuro-divergent/neuro-divergent-models/Cargo.toml","sourceSha256":"7a8d1a358aaa1eff7317de67100e90d13f99045e20c97ee60f5e6bfde3d5a91d","variants":[],"readme":"A comprehensive neural forecasting library built on the ruv-FANN foundation, providing state-of-the-art time series forecasting models optimized for Rust's performance and safety guarantees. Neuro-Divergent Models is a neural forecasting model library that implements 27+ neural network architectures specifically designed for time series forecasting. Built on top of the ruv-FANN (Fast Artificial Neural Network) library, it provides a unified, type-safe, and performant interface for deploying neural forecasting models in production environments. The library bridges the gap between research-focused implementations and production-ready systems by offering: - Memory-efficient implementations with","readmeSource":"ruv-fann/neuro-divergent/neuro-divergent-models/README.md","product":"key:ruv-fann:neuro-divergent-models","canonical":"neuro-divergent-models","tier":"T1","t":["neuro-divergent-models","neuro","divergent","model","neural","forecast","implement","ruv-fann","fann","sery","lstm","neural-network","network","rnn","time-series"],"s":["neuro-divergent-models","neuro","divergent","model","forecast","lstm","neural-network","neural","network","rnn","time-series","sery"]},
+{"id":"neuro-divergent-registry","family":"neuro-divergent-registry","kind":"crate","store":"ruv-fann","version":null,"description":"Model factory and registry system for dynamic neural network model creation and management","keywords":["factory","forecasting","machine-learning","neural-networks","registry"],"source":"ruv-fann/neuro-divergent/neuro-divergent-registry/Cargo.toml","sourceSha256":"9b2061aaac619f06ac31bec35f9e0599d567d53222a59a29eae26ece15a031a6","variants":[],"readme":"A comprehensive model factory and registry system for dynamic neural network model creation and management. This crate provides the foundation for discovering, creating, and managing neural network models across the neuro-divergent ecosystem with support for plugins, serialization, and automatic model discovery. The neuro-divergent-registry crate implements a sophisticated registry architecture that serves as the central nervous system for neural network model management. It provides dynamic dispatch capabilities without runtime cost overhead, type-safe model creation with compile-time guarantees, and a plugin architecture for unlimited extensibility. - Global Model Registry: Centralized man","readmeSource":"ruv-fann/neuro-divergent/neuro-divergent-registry/README.md","product":"key:ruv-fann:neuro-divergent-registry","canonical":"neuro-divergent-registry","tier":"T1","t":["neuro-divergent-registry","neuro","divergent","registry","model","factory","dynamic","neural","network","creation","management","forecast","machine-learning","machine","learn","neural-networks"],"s":["neuro-divergent-registry","neuro","divergent","registry","factory","forecast","machine-learning","machine","learn","neural-networks","neural","network"]},
+{"id":"neuro-divergent-training","family":"neuro-divergent-training","kind":"crate","store":"ruv-fann","version":null,"description":"Comprehensive training infrastructure for neural forecasting models","keywords":["forecasting","machine-learning","neural-network","time-series","training"],"source":"ruv-fann/neuro-divergent/neuro-divergent-training/Cargo.toml","sourceSha256":"6cb43ff9d141f03eedc6ad27420c19b9fb584758b7187f0fc0da71e76f6d04be","variants":[],"readme":"Comprehensive training infrastructure for neural forecasting models with advanced optimization, loss functions, and training strategies specifically designed for time series forecasting. The neuro-divergent-training crate provides a complete training ecosystem for neural time series forecasting models, featuring modern optimizers, specialized loss functions, adaptive learning rate scheduling, and comprehensive evaluation metrics. This crate seamlessly integrates with the ruv-FANN neural network library to provide production-ready training capabilities. - Adam: Adaptive Moment Estimation with bias correction and AMSGrad support - AdamW: Adam with decoupled weight decay for better regularizati","readmeSource":"ruv-fann/neuro-divergent/neuro-divergent-training/README.md","product":"key:ruv-fann:neuro-divergent-training","canonical":"neuro-divergent-training","tier":"T1","t":["neuro-divergent-training","neuro","divergent","train","comprehensive","infrastructure","neural","forecast","model","machine-learning","machine","learn","neural-network","network","time-series","sery"],"s":["neuro-divergent-training","neuro","divergent","train","forecast","machine-learning","machine","learn","neural-network","neural","network","time-series","sery"]},
+{"id":"nvsim","family":"nvsim","kind":"crate","store":"ruview","version":null,"description":"Deterministic NV-diamond magnetometer pipeline simulator (source -> propagation -> NV ensemble -> ADC + lockin demod)","keywords":["biot-savart","magnetometer","nv-diamond","physics","simulator"],"source":"ruview/v2/crates/nvsim/Cargo.toml","sourceSha256":"47287908bb388bb52bc147f79a4f736706fff8541dc9c3febd2f7f6c48ba08ff","variants":[],"readme":"Deterministic Rust simulator for NV-diamond ensemble magnetometers. Synthesise the magnetic-field trace a real sensor *would have produced* — without the hardware, the lab, or the $8 K vendor receipt. NV-diamond magnetometers are exotic but real: they detect magnetic fields by shining green laser at a diamond and watching how its red fluorescence shifts under microwave excitation. They are sensitive enough to feel a person's heartbeat from across a room — when they work. The catch: a working ensemble sensor costs ~$8 K and lives in a lab. nvsim runs the same forward pipeline in software, end-to-end, deterministically, so you can ask \"what would my magnetometer have seen if a steel rebar walk","readmeSource":"ruview/v2/crates/nvsim/README.md","product":"key:ruview:nvsim","canonical":"nvsim","tier":"T0","t":["nvsim","deterministic","nv-diamond","diamond","magnetometer","pipeline","simulator","source","propagation","ensemble","adc","lockin","demod","biot-savart","biot","savart","physic"],"s":["nvsim","biot-savart","biot","savart","magnetometer","nv-diamond","diamond","physic","simulator"]},
+{"id":"nvsim-server","family":"nvsim-server","kind":"crate","store":"ruview","version":null,"description":"Axum REST + WebSocket server fronting the nvsim NV-diamond pipeline simulator (ADR-092 §6.2).","keywords":["axum","magnetometer","nvsim","simulator","websocket"],"source":"ruview/v2/crates/nvsim-server/Cargo.toml","sourceSha256":"d7942f0e147c3a03dde68c22d06a84a0cffef7393df25246c5981a3ddce8d6ad","variants":[],"product":"key:ruview:nvsim-server","canonical":"nvsim-server","tier":"T0","t":["nvsim-server","nvsim","server","axum","rest","websocket","front","nv-diamond","diamond","pipeline","simulator","adr-092","adr","092","6.2","magnetometer"],"s":["nvsim-server","nvsim","server","axum","magnetometer","simulator","websocket"]},
+{"id":"obsidian-brain","family":"obsidian-brain","kind":"npm","store":"obsidian-brain","version":"0.1.0","description":"Obsidian plugin for the RuVector Brain — DiskANN semantic search + AIDefence + DPO pairs over your vault.","keywords":["brain","diskann","embeddings","obsidian","plugin","ruvector","semantic-search"],"source":"obsidian-brain/package.json","sourceSha256":"9e7ddf1273d7817ce83a8e3bab6f37702ebe01dd0dc559f8a77d6de3aff5c815","variants":[],"readme":"You write notes in Obsidian. This plugin does three things with them: 1. It understands them. Press <kbd>⌘⇧B</kbd> and type what you're looking for. Instead of matching keywords, it finds notes that mean something similar to your question — even if you used different words. Press <kbd>⌘⇧K</kbd> instead and you'll get a Q&A view that pulls together the relevant passages across your whole vault. 2. It shares them with your AI tools. Claude Code, Codex CLI, Gemini CLI and other MCP-speaking assistants can read (and write) the same notes you see in Obsidian. No more re-explaining context at the start of every session — your agent already knows what you decided last week. 3. It connects you to a","readmeSource":"obsidian-brain/README.md","product":"repo:obsidian-brain","canonical":"obsidian-brain","tier":"T2","t":["obsidian-brain","obsidian","brain","plugin","diskann","semantic","search","aidefence","dpo","pair","vault","embedding","semantic-search"],"s":["obsidian-brain","obsidian","brain","diskann","embedding","plugin","semantic-search","semantic","search"]},
+{"id":"onnx-agent-agent","family":"onnx-agent-agent","kind":"npm","store":"onnx-agent","version":"0.1.0","description":"Maintain and validate onnx-agent","keywords":[],"source":"onnx-agent/.harness/generated/package.json","sourceSha256":"651b24be475dd85dda0026ea36abd986cf13a7bff18ef2bd7d50409babf4876d","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"onnx-agent/.harness/generated/README.md","product":"key:onnx-agent:onnx-agent-agent","canonical":"onnx-agent-agent","tier":"T3","t":["onnx-agent-agent","onnx","maintain","validate","onnx-agent"],"s":["onnx-agent-agent","onnx"]},
+{"id":"open-claude-code","family":"open-claude-code","kind":"npm","store":"open-claude-code","version":"2.0.0","description":"Open source Claude Code CLI — based on ruDevolution decompilation of v2.1.91","keywords":["ai","claude","cli","code","decompiler","open-source"],"source":"open-claude-code/v2/package.json","sourceSha256":"05527b2fd4ae5379a516efebd55e0fccec3c69f5726d62debe101d9a07110474","variants":[],"product":"repo:open-claude-code","canonical":"open-claude-code","tier":"T2","t":["open-claude-code","open","claude","code","source","rudevolution","decompilation","v2.1.91","decompiler","open-source"],"s":["open-claude-code","open","claude","code","decompiler","open-source","source"]},
+{"id":"opencv-sdk","family":"opencv-sdk","kind":"crate","store":"ruv-fann","version":null,"description":"OpenCV SDK with C/C++/Python API compatibility layer","keywords":["bindings","compatibility","ffi","opencv","sdk"],"source":"ruv-fann/opencv-rust/opencv-sdk/Cargo.toml","sourceSha256":"85f2268da98fd2bd4885a09beac7178ee2faf4f8b68a5807229addaf709efa8b","variants":[],"readme":"OpenCV SDK provides a compatibility layer for using OpenCV's Rust implementation with existing C/C++ and Python code. This crate bridges the gap between Rust's safety and performance with legacy codebases that expect OpenCV's traditional API. - FFI Compatibility: Seamless integration with C/C++ codebases - Python Bindings: Optional PyO3-based Python interface - WebAssembly Support: Optional WASM compilation for browser deployment - Type Safety: Rust's ownership system with familiar OpenCV patterns - Zero-Copy Operations: Efficient data sharing between languages Add this to your Cargo.toml: - Core SDK functionality - C/C++ FFI bindings - Basic type conversions - python: Enable Python bindings","readmeSource":"ruv-fann/opencv-rust/opencv-sdk/README.md","product":"key:ruv-fann:opencv-sdk","canonical":"opencv-sdk","tier":"T1","t":["opencv-sdk","opencv","python","compatibility","layer","ffi"],"s":["opencv-sdk","opencv","compatibility","ffi"]},
+{"id":"opencv-sys","family":"opencv","kind":"crate","store":"ruv-fann","version":null,"description":"Low-level Rust bindings for OpenCV","keywords":["bindings","browser","computer-vision","core","ffi","mat","opencv","point","rect","wasm","webassembly"],"source":"ruv-fann/opencv-rust/opencv-sys/Cargo.toml","sourceSha256":"49a7453ef6c47e1942cb92b92e2c62bca1701e906a1d5d116783561d501d8621","variants":["opencv-core","opencv-wasm"],"product":"key:ruv-fann:opencv","canonical":"opencv-sys","tier":"T1","t":["opencv-sys","opencv","sys","low-level","low","level","browser","computer-vision","computer","vision","core","ffi","mat","point","rect"],"s":["opencv-sys","opencv","sys","browser","computer-vision","computer","vision","core","ffi","mat","point","rect"]},
+{"id":"phone-agent","family":"phone-agent","kind":"npm","store":"phone-agent","version":"1.0.0","description":"A Cloudflare Pages application that provides AI-powered phone agents for meetings and calls","keywords":["agent","ai","calendar","cloudflare","functions","meeting","pages","phone","webhook"],"source":"phone-agent/package.json","sourceSha256":"5cbb1c467daf6a13e5e9a059897b1abc0ed5958b1958c87ebc6c23cd59ecd99d","variants":[],"readme":"A Cloudflare Pages application that provides AI phone agent capabilities, including scheduling calls, managing calendars, and handling email notifications. This application follows a simplified architecture using Cloudflare Pages Functions: - AI Phone Agent: Schedule and manage automated phone calls using Bland.ai - Calendar Integration: Create and manage calendar events for scheduled calls - Email Notifications: Send confirmation, rescheduling, and cancellation emails - Webhook Processing: Handle webhooks from email and phone call services - Secure Storage: Store call and scheduling data securely - Error Handling: Comprehensive error handling and reporting The application requires the follo","readmeSource":"phone-agent/README.md","product":"repo:phone-agent","canonical":"phone-agent","tier":"T3","t":["phone-agent","phone","cloudflare","page","provide","ai-powered","power","meeting","call","calendar","function","meet","webhook"],"s":["phone-agent","phone","calendar","cloudflare","function","meet","page","webhook"]},
+{"id":"photonlayer-cli","family":"photonlayer","kind":"crate","store":"photonlayer","version":null,"description":"PhotonLayer command-line studio: run simulations, demos (barcode, edge, privacy gate), benchmarks, and verify receipts (ADR-260)","keywords":["diffraction","optical-computing","optics","simulation","wasm"],"source":"photonlayer/crates/photonlayer-cli/Cargo.toml","sourceSha256":"de2511def564880dbad630bcb07a2da1ed63e96d73ec1602606bd9b13824bf41","variants":["photonlayer-core","photonlayer-wasm"],"product":"repo:photonlayer","canonical":"photonlayer-cli","tier":"T2","t":["photonlayer-cli","photonlayer","command-line","command","line","studio","simulation","demo","barcode","edge","privacy","gate","benchmark","verify","receipt","adr-260","adr","260","diffraction","optical-computing","optical","comput","optic"],"s":["photonlayer-cli","photonlayer","diffraction","optical-computing","optical","comput","optic","simulation"]},
+{"id":"photonlayer-ruvector","family":"photonlayer-ruvector","kind":"crate","store":"ruvector","version":null,"description":"PhotonLayer experiment memory on RuVector: mask/frame embeddings, similarity recall, pass/fail boundary, coherence, RVF receipts (ADR-260 Phase 2)","keywords":[],"source":"ruvector/crates/photonlayer-ruvector/Cargo.toml","sourceSha256":"dfdbeb95ab96105e8571243f46f50e9394e74e9fc9308c3069e1bddf857d6fa8","variants":[],"product":"key:ruvector:photonlayer-ruvector","canonical":"photonlayer-ruvector","tier":"T0","t":["photonlayer-ruvector","photonlayer","experiment","memory","mask","frame","embedding","similarity","recall","pass","fail","boundary","coherence","rvf","receipt","adr-260","adr","260","phase"],"s":["photonlayer-ruvector","photonlayer"]},
+{"id":"pixelrag-cli","family":"pixelrag","kind":"crate","store":"rupixel","version":null,"description":"PixelRAG CLI + benchmark harness — ingestion, search, and the darwin-driven benchmark entry point (ADR-264, M0 scaffold).","keywords":[],"source":"rupixel/rust/pixelrag-cli/Cargo.toml","sourceSha256":"685eefe59fb81473c72713e490d5dd3618f7e088173d0dc4e651306b172f8bcb","variants":["pixelrag-core"],"product":"key:rupixel:pixelrag","canonical":"pixelrag-cli","tier":"T2","t":["pixelrag-cli","pixelrag","benchmark","harness","ingestion","search","darwin-driven","darwin","driven","entry","point","adr-264","adr","264","scaffold"],"s":["pixelrag-cli","pixelrag"]},
+{"id":"pixelrag-embed-sidecar","family":"pixelrag-embed-sidecar","kind":"npm","store":"rupixel","version":"0.1.0","description":"Real semantic embedding sidecar for the PixelRAG Rust bench — all-MiniLM-L6-v2 via transformers.js (pure WASM, CPU, no GPU).","keywords":[],"source":"rupixel/rust/pixelrag-cli/sidecar/package.json","sourceSha256":"d4aecea788af421a432741000c50d9a2ea93077c8a862ed82522fde81df97ae5","variants":[],"product":"key:rupixel:pixelrag-embed-sidecar","canonical":"pixelrag-embed-sidecar","tier":"T2","t":["pixelrag-embed-sidecar","pixelrag","embed","sidecar","real","semantic","embedd","bench","all-minilm-l6-v2","all","minilm","via","transformers.j","pure","cpu","gpu"],"s":["pixelrag-embed-sidecar","pixelrag","embed","sidecar"]},
+{"id":"pixelrag-encoder","family":"pixelrag-encoder","kind":"crate","store":"rupixel","version":null,"description":"PixelRAG visual encoder wrapper (ONNX / Qwen3-VL-Embedding / CLIP surrogate / Python sidecar). M0 std-only skeleton per ADR-264.","keywords":[],"source":"rupixel/rust/pixelrag-encoder/Cargo.toml","sourceSha256":"bd378ab916f5171a322e521bbf69b6f99843ed0ecd1e5f42f9976cfe75f82634","variants":[],"product":"key:rupixel:pixelrag-encoder","canonical":"pixelrag-encoder","tier":"T2","t":["pixelrag-encoder","pixelrag","encoder","visual","wrapper","onnx","qwen3-vl-embedding","qwen3","embedd","clip","surrogate","python","sidecar","std-only","std","only","skeleton","per","adr-264","adr","264"],"s":["pixelrag-encoder","pixelrag","encoder"]},
+{"id":"pixelrag-render","family":"pixelrag-render","kind":"crate","store":"rupixel","version":null,"description":"PixelRAG headless render adaptor — shells the Node render_sidecar.mjs (puppeteer-core + Edge/Chrome) to turn URLs into screenshot PNGs for the visual-RAG path (ADR-264).","keywords":[],"source":"rupixel/rust/pixelrag-render/Cargo.toml","sourceSha256":"63bf10ab7ea19433ba790877a5b11e227b02c541d505b391063d81a36e438a08","variants":[],"product":"key:rupixel:pixelrag-render","canonical":"pixelrag-render","tier":"T2","t":["pixelrag-render","pixelrag","render","headless","adaptor","shell","sidecar.mj","puppeteer-core","puppeteer","core","edge","chrome","turn","urls","screenshot","pngs","visual-rag","visual","rag","path","adr-264","adr","264"],"s":["pixelrag-render","pixelrag","render"]},
+{"id":"prime-napi","family":"prime","kind":"crate","store":"daa","version":null,"description":"NAPI-rs bindings for Prime ML framework - Node.js bindings for federated learning","keywords":[],"source":"daa/prime-rust/prime-napi/Cargo.toml","sourceSha256":"e5884f87046b389e7ff3cb695a81d26a50ba4cbf25d1c6b47e58c919418b4aa9","variants":[],"readme":"High-performance Node.js bindings for the Prime ML distributed federated learning framework, built with Rust and NAPI-rs. - Zero-Copy Operations: Direct memory access for tensor data using napi::Buffer - High Performance: Rust-powered computation with minimal overhead - Federated Learning: Built-in support for distributed training coordination - Parallel Gradient Aggregation: Multiple aggregation strategies (FedAvg, Secure Aggregation, Trimmed Mean, Krum) - Type Safety: Full TypeScript type definitions - Cross-Platform: Supports Linux, macOS, Windows (x64 and ARM64)","readmeSource":"daa/prime-rust/prime-napi/README.md","product":"key:daa:prime","canonical":"prime-napi","tier":"T1","t":["prime-napi","prime","federat","learn"],"s":["prime-napi","prime"]},
+{"id":"prime-radiant","family":"prime-radiant","kind":"crate","store":"ruvector","version":null,"description":"Universal coherence engine using sheaf Laplacian mathematics for AI safety, hallucination detection, and structural consistency verification in LLMs and distributed systems","keywords":["ai-safety","coherence","hallucination","llm","sheaf-theory"],"source":"ruvector/crates/prime-radiant/Cargo.toml","sourceSha256":"0509ea14bac22069ab9a6a4aa161f00cbaa3e7c6561edd31ee68f466daac41e9","variants":[],"readme":"A Real-Time Coherence Gate for Autonomous Systems Prime-Radiant is infrastructure for AI safety — a mathematical gate that proves whether a system's beliefs, facts, and claims are internally consistent before allowing action. Instead of asking \"How confident am I?\" (which can be wrong), Prime-Radiant asks \"Are there any contradictions?\" — and provides mathematical proof of the answer. - [What It Does](#what-it-does) - [Mathematical Foundation](#mathematical-foundation) - [Key Concepts](#key-concepts) - [Installation](#installation) - [Quick Start](#quick-start) - [Performance & Acceleration](#performance--acceleration) - [Storage Backends](#storage-backends) - [Applications](#applications) -","readmeSource":"ruvector/crates/prime-radiant/README.md","product":"key:ruvector:prime-radiant","canonical":"prime-radiant","tier":"T0","t":["prime-radiant","prime","radiant","universal","coherence","sheaf","laplacian","mathematic","safety","hallucination","detection","structural","consistency","verification","distribut","ai-safety","sheaf-theory","theory"],"s":["prime-radiant","prime","radiant","ai-safety","safety","coherence","hallucination","sheaf-theory","sheaf","theory"]},
+{"id":"promotion","family":"promotion","kind":"crate","store":"autogenous","version":null,"description":"Autogenous promotion controller: staged canary (1-10-50-100%), signed promotion, automatic rollback on gate violation (ADR-392 Phase 6).","keywords":[],"source":"autogenous/crates/promotion/Cargo.toml","sourceSha256":"dd584c1520ba689f5c72594451db23875f8c833b34f1dea47dacaa74dee6261c","variants":[],"readme":"promotion is the rollout controller of [Autogenous](../), a self-evolving AI-defense runtime in Rust. It's a deterministic state machine that walks a verified defense through a staged canary — 1% → 10% → 50% → 100% — advancing only while measured fitness keeps passing the constitutional hard gates, and rolling back automatically on the first violation. Two asymmetries make it safe: promotion requires a signed, single-use token; rollback requires nothing (it's always the safe direction). - CanaryController — observe(fitness) feeds one measurement and returns a Decision (Hold / Advance / ReadyForPromotion / RollBack). - promote(&VerifiedPromotion, now) — finalize the rollout. Accepts only the","readmeSource":"autogenous/crates/promotion/README.md","product":"key:autogenous:promotion","canonical":"promotion","tier":null,"t":["promotion","autogenous","controller","stag","canary","1-10-50-100","100","sign","automatic","rollback","gate","violation","adr-392","adr","392","phase"],"s":["promotion"]},
+{"id":"prompt-forge","family":"prompt-forge","kind":"crate","store":"symbolic-scribe","version":null,"description":"Low-latency prompt compiler & multi-objective optimizer (Rust → WASM) for Symbolic Scribe","keywords":[],"source":"symbolic-scribe/wasm/prompt-forge/Cargo.toml","sourceSha256":"38b27960dc28cc632907e7e5bb6e4b0ebe92c6551693a7d6ad21381e132c102b","variants":[],"readme":"A low-latency prompt compiler & multi-objective optimizer written in Rust and compiled to WebAssembly. It is the deterministic, offline core of Symbolic Scribe's \"PromptOps\" pipeline — it turns a loose natural-language prompt into a structured, scored, compressed, risk-classified, and cryptographically signed prompt artifact, entirely client-side. Live model testing (OpenRouter) and optional semantic embeddings (transformers.js) live in the host app; everything here is fast, deterministic, and dependency-light so it can run on every keystroke. A full BPE tokenizer + scorer in JS is either slow or pulls megabytes of merge tables. Doing it in Rust gives us: - Sub-millisecond latency — see benc","readmeSource":"symbolic-scribe/wasm/prompt-forge/README.md","product":"key:symbolic-scribe:prompt-forge","canonical":"prompt-forge","tier":"T2","t":["prompt-forge","prompt","forge","low-latency","low","latency","compiler","multi-objective","multi","objective","optimizer","symbolic","scribe"],"s":["prompt-forge","prompt","forge"]},
+{"id":"promptlang-agent","family":"promptlang-agent","kind":"npm","store":"promptlang","version":"0.1.0","description":"Maintain and validate promptlang","keywords":[],"source":"promptlang/.harness/generated/package.json","sourceSha256":"276bf8a500126969ec128ef5862b5a6e83c0d9e770889e397efe936aa6c58e9e","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"promptlang/.harness/generated/README.md","product":"key:promptlang:promptlang-agent","canonical":"promptlang-agent","tier":"T3","t":["promptlang-agent","promptlang","maintain","validate"],"s":["promptlang-agent","promptlang"]},
+{"id":"psycho-symbolic-performance-validation","family":"psycho-symbolic-performance-validation","kind":"npm","store":"sublinear-time-solver","version":"1.0.0","description":"Verifiable performance benchmarks for Psycho-Symbolic Reasoner","keywords":[],"source":"sublinear-time-solver/validation/package.json","sourceSha256":"c7663645044cfc20b84d83d9b217eab9c2d7a8269ddbbf002111fc88c362d08c","variants":[],"readme":"This validation suite provides verifiable proof of the Psycho-Symbolic Reasoner's performance claims through reproducible benchmarks and comparisons with traditional AI reasoning systems. - Simple Query: 0.3ms (500x faster than GPT-4) - Complex Reasoning: 2.1ms (380x faster than GPT-4) - Graph Traversal: 1.2ms - GOAP Planning: 1.8ms - Psycho-Symbolic operations measured with high-resolution timers - 10,000-100,000 iterations per test - Statistical analysis (mean, median, P95, P99) - Based on published performance data - Simulates realistic latencies - Includes network overhead for cloud services - Side-by-side performance comparison - Speedup calculations - Statistical validation - Node.js 2","readmeSource":"sublinear-time-solver/validation/README.md","product":"key:sublinear-time-solver:psycho-symbolic-performance-validation","canonical":"psycho-symbolic-performance-validation","tier":"T1","t":["psycho-symbolic-performance-validation","psycho","symbolic","validation","verifiable","benchmark","psycho-symbolic","reasoner"],"s":["psycho-symbolic-performance-validation","psycho","symbolic","validation"]},
+{"id":"psycho-symbolic-reasoner","family":"psycho-symbolic-reasoner","kind":"npm","store":"sublinear-time-solver","version":"1.0.7","description":"A psycho-symbolic reasoning framework combining symbolic AI with psychological context using Rust WASM and FastMCP integration","keywords":["affect","ai-agents","autonomous-agents","cli","fastmcp","framework","goal-oriented-planning","graph-reasoning","inference-engine","mcp","model-context-protocol","planning","preferences","psycho-symbolic","psychology","reasoning","rule-based-planning","rust","sandboxed-execution","sdk","sentiment-analysis","symbolic-ai","typescript","verifiable-reasoning","wasm","webassembly"],"source":"sublinear-time-solver/crates/psycho-symbolic-reasoner/package.json","sourceSha256":"19083f0753cf67a12de31cd54d5fda68c7b8bda7d7c295c01bc2ee21b18cba5f","variants":[],"readme":"The Psycho-Symbolic Reasoner represents a paradigm shift in AI reasoning systems, combining the mathematical rigor of symbolic AI with the nuanced understanding of human psychology. Built on cutting-edge Rust/WebAssembly technology, this framework delivers sub-millisecond reasoning that outperforms traditional systems by orders of magnitude. Traditional reasoning systems struggle with the complexity of human-centric decision making. They either focus purely on logical deduction (missing emotional and preference factors) or rely on slow, resource-intensive neural networks. Our approach bridges this gap with a hybrid architecture that: - Thinks Fast: Sub-millisecond response times vs. 100-500m","readmeSource":"sublinear-time-solver/crates/psycho-symbolic-reasoner/README.md","product":"key:sublinear-time-solver:psycho-symbolic-reasoner","canonical":"psycho-symbolic-reasoner","tier":"T1","t":["psycho-symbolic-reasoner","psycho","symbolic","reasoner","psycho-symbolic","reason","combin","psychological","context","fastmcp","integration","affect","ai-agents","autonomous-agents","autonomous","goal-oriented-planning","goal","orient","plann","graph-reasoning","graph","inference-engine","inference","mcp","model-context-protocol","model","protocol","preference","psychology","rule-based-planning","rule","sandboxed-execution","sandbox","execution","sentiment-analysis","sentiment","analysis","symbolic-ai","verifiable-reasoning","verifiable"],"s":["psycho-symbolic-reasoner","psycho","symbolic","reasoner","affect","ai-agents","autonomous-agents","autonomous","fastmcp","goal-oriented-planning","goal","orient","plann","graph-reasoning","graph","reason","inference-engine","inference","mcp","model-context-protocol","model","context","protocol","preference","psycho-symbolic","psychology","rule-based-planning","rule","sandboxed-execution","sandbox","execution","sentiment-analysis","sentiment","analysis","symbolic-ai","verifiable-reasoning","verifiable"]},
+{"id":"psycho-symbolic-reasoner-mcp","family":"psycho-symbolic-reasoner-mcp","kind":"npm","store":"sublinear-time-solver","version":"1.0.0","description":"FastMCP integration for psycho-symbolic reasoner WASM modules","keywords":["ai","fastmcp","mcp","reasoning","symbolic","wasm"],"source":"sublinear-time-solver/crates/psycho-symbolic-reasoner/mcp-integration/package.json","sourceSha256":"f822ffb75af79208cbbf0371dd36f4ff31c06023a6c99fcda714d7e675c5a6a5","variants":[],"readme":"A production-ready TypeScript/FastMCP integration layer for the psycho-symbolic-reasoner WASM modules. - Complete WASM Integration: TypeScript wrappers for all three core modules - Graph Reasoner (symbolic reasoning) - Text Extractor (sentiment, preferences, emotions) - Planner (GOAP action planning) - MCP Tool Registration: 15+ tools for Model Context Protocol - queryGraph - Query knowledge graph - addFact, addRule - Build knowledge base - extractAffect - Sentiment analysis - extractPreferences - Preference extraction - extractEmotions - Emotion detection - planAction - GOAP planning - And more utility tools - Production Features: - Memory management with automatic cleanup - Error handling","readmeSource":"sublinear-time-solver/crates/psycho-symbolic-reasoner/mcp-integration/README.md","product":"key:sublinear-time-solver:psycho-symbolic-reasoner-mcp","canonical":"psycho-symbolic-reasoner-mcp","tier":"T1","t":["psycho-symbolic-reasoner-mcp","psycho","symbolic","reasoner","mcp","fastmcp","integration","psycho-symbolic","reason"],"s":["psycho-symbolic-reasoner-mcp","psycho","symbolic","reasoner","mcp","fastmcp","reason"]},
+{"id":"publish-registry","family":"publish-registry","kind":"npm","store":"ruflo","version":"1.0.0","description":"Secure IPFS plugin registry publisher for Claude Flow","keywords":[],"source":"ruflo/v3/@claude-flow/cli/cloud-functions/publish-registry/package.json","sourceSha256":"43e4e662448ffa7e454bae56d4299ac7a3324389abbfc963d1d1d16eda3c9559","variants":[],"product":"key:ruflo:publish-registry","canonical":"publish-registry","tier":"T0","t":["publish-registry","publish","registry","secure","ipfs","plugin","publisher","claude","flow"],"s":["publish-registry","publish","registry"]},
+{"id":"quantum-currency-project","family":"quantum-currency-project","kind":"npm","store":"quantum_cryptocurrency","version":"1.0.0","description":"A hypothetical quantum-secured cryptocurrency leveraging Google's Willow quantum chip.","keywords":[],"source":"quantum_cryptocurrency/quantum_crypto/package.json","sourceSha256":"23506e39e13988b49aa6fb0021d6fe7c526a2621906e5637feee377ba2ef1b50","variants":[],"readme":"This project outlines a hypothetical implementation of a quantum-secured cryptocurrency leveraging Google's Willow quantum chip. The implementation follows the SPARC Framework, encompassing Specification, Pseudocode, Architecture, Refinement, and Completion phases. - specification/: Project objectives, requirements, user scenarios, UI/UX guidelines, and technology stack. - pseudocode/: High-level pseudocode outlining the application's logic and flow. - architecture/: System architecture details and diagrams. - refinement/: Notes and documentation related to performance improvements and maintainability. - completion/: Final testing, documentation, and deployment preparations. - src/: Source c","readmeSource":"quantum_cryptocurrency/quantum_crypto/README.md","product":"key:quantum_cryptocurrency:quantum-currency-project","canonical":"quantum-currency-project","tier":"T3","t":["quantum-currency-project","quantum","currency","project","hypothetical","quantum-secured","secur","cryptocurrency","leverag","google","willow","chip"],"s":["quantum-currency-project","quantum","currency","project"]},
+{"id":"qudag","family":"qudag","kind":"crate","store":"qudag","version":null,"description":"QuDAG - Darknet for agent swarms: Ultra-fast quantum-resistant distributed communication platform with DAG architecture, rUv token exchange, .dark domains, onion routing, and ML-DSA/ML-KEM cryptography","keywords":["agent-swarm","cli","cryptography","dag","darknet","distributed","exchange","quantum","quantum-resistant","wasm","webassembly"],"source":"qudag/qudag/Cargo.toml","sourceSha256":"e766f35de637214aaae358163625b8a5dd14a6bf92a7676703e1b193665b43db","variants":["qudag-cli","qudag-wasm"],"readme":"QuDAG is a revolutionary quantum-resistant distributed communication platform built for the quantum age, designed to support the next generation of autonomous AI agents, swarm intelligence, and zero-person businesses. This is the core library that provides the foundation for building Agentic Organizations—where AI-driven systems run businesses entirely on their own. QuDAG enables businesses to operate autonomously and decentrally, without human intervention. The platform provides: * Quantum-Resistant Security: ML-KEM-768 & ML-DSA cryptography protects against future quantum threats * Agent Swarm Coordination: Built-in support for autonomous AI agent communication and coordination * Resource","readmeSource":"qudag/qudag/README.md","product":"repo:qudag","canonical":"@daa/qudag-native","tier":"T1","t":["qudag","darknet","swarm","ultra-fast","ultra","quantum-resistant","quantum","resistant","distribut","communication","dag","architecture","token","exchange","dark","domain","onion","rout","ml-dsa","dsa","ml-kem","kem","cryptography","agent-swarm"],"s":["qudag","agent-swarm","swarm","cryptography","dag","darknet","distribut","exchange","quantum","quantum-resistant","resistant"]},
+{"id":"qudag-core","family":"qudag","kind":"crate","store":"synaptic-mesh","version":null,"description":"QuDAG core networking and consensus for quantum-resistant DAG-based mesh networks","keywords":["consensus","dag","mesh","network","p2p"],"source":"synaptic-mesh/src/rs/qudag-core/Cargo.toml","sourceSha256":"b0304518505e46f1676d13b885035b7b4883f8742c570c1fa666293f598e6013","variants":[],"product":"repo:qudag","canonical":"@daa/qudag-native","tier":"T1","t":["qudag-core","qudag","core","network","consensus","quantum-resistant","quantum","resistant","dag-based","dag","mesh","p2p"],"s":["qudag-core","qudag","core","consensus","dag","mesh","network","p2p"]},
+{"id":"qudag-crypto","family":"qudag-crypto","kind":"crate","store":"qudag","version":null,"description":"Quantum-resistant cryptography library for QuDAG - includes ML-KEM-768, ML-DSA, HQC, and BLAKE3","keywords":["cryptography","ml-dsa","ml-kem","post-quantum","quantum-resistant"],"source":"qudag/core/crypto/Cargo.toml","sourceSha256":"62e1774ecbe54f01bf6140f34bfc4e0885b823c38b5e03e659a7d9bf55573309","variants":[],"readme":"Quantum-resistant cryptographic primitives for the QuDAG protocol. - ML-KEM-768: NIST-standardized key encapsulation mechanism (FIPS 203) - ML-DSA (Dilithium): Digital signature algorithm (FIPS 204) - HQC: Hamming Quasi-Cyclic code-based encryption (128/192/256-bit) - BLAKE3: Fast cryptographic hash function - Quantum Fingerprinting: Data fingerprinting using ML-DSA signatures - Memory Safety: Automatic secret zeroization and constant-time operations Add to your Cargo.toml: - Post-Quantum Security: Resistant to attacks from quantum computers - Memory Safety: All secret data is automatically zeroized - Side-Channel Resistance: Constant-time implementations - NIST Compliance: Implements FIPS 2","readmeSource":"qudag/core/crypto/README.md","product":"key:qudag:qudag-crypto","canonical":"qudag-crypto","tier":"T1","t":["qudag-crypto","qudag","crypto","quantum-resistant","quantum","resistant","cryptography","include","ml-kem-768","kem","768","ml-dsa","dsa","hqc","blake3","ml-kem","post-quantum","post"],"s":["qudag-crypto","qudag","crypto","cryptography","ml-dsa","dsa","ml-kem","kem","post-quantum","post","quantum","quantum-resistant","resistant"]},
+{"id":"qudag-dag","family":"qudag-dag","kind":"crate","store":"qudag","version":null,"description":"DAG consensus implementation for QuDAG - QR-Avalanche algorithm with Byzantine fault tolerance","keywords":["avalanche","byzantine","consensus","dag","distributed-ledger"],"source":"qudag/core/dag/Cargo.toml","sourceSha256":"c17d597f416c0ba9ebb9734c0eeef383917f09750cfabae3f38aa65112f50a32","variants":[],"readme":"DAG consensus implementation with QR-Avalanche algorithm for the QuDAG protocol. - QR-Avalanche Consensus: Quantum-resistant Byzantine fault-tolerant consensus - Parallel Processing: Concurrent message validation and ordering - Conflict Resolution: Automatic detection and resolution of conflicts - Tip Selection: Intelligent parent selection algorithms - Performance Monitoring: Real-time consensus metrics Add to your Cargo.toml: - DAG Structure: Directed acyclic graph for message ordering - QR-Avalanche: Quantum-resistant consensus algorithm - Vertex Processing: Parallel validation and state management - Conflict Detection: Automatic fork detection and resolution - Tip Selection: Algorithms f","readmeSource":"qudag/core/dag/README.md","product":"key:qudag:qudag-dag","canonical":"qudag-dag","tier":"T1","t":["qudag-dag","qudag","dag","consensus","implementation","qr-avalanche","avalanche","algorithm","byzantine","fault","tolerance","distributed-ledger","distribut","ledger"],"s":["qudag-dag","qudag","dag","avalanche","byzantine","consensus","distributed-ledger","distribut","ledger"]},
+{"id":"qudag-exchange","family":"qudag-exchange","kind":"crate","store":"qudag","version":null,"description":"QuDAG Exchange: Quantum-resistant token exchange with rUv (Resource Utilization Voucher) system","keywords":["blockchain","cli","exchange","quantum","qudag","tokens","wasm"],"source":"qudag/qudag-exchange/Cargo.toml","sourceSha256":"0a9fced0557439e597b3738c0346de3aa821fd4eaebd3d229e293b2189bf1e34","variants":["qudag-exchange-cli","qudag-exchange-core","qudag-exchange-wasm"],"readme":"QuDAG Exchange is a modular, quantum-secure, multi-agent resource exchange protocol that enables trustless computational resource trading through rUv (Resource Utilization Voucher) tokens. - rUv Tokens: Resource Utilization Vouchers for computational resource trading - Quantum-Resistant Security: ML-DSA signatures and ML-KEM encryption - DAG Consensus: QR-Avalanche protocol for scalable transaction finality - Vault Integration: Secure key management with QuDAG Vault - WASM Support: Run in browsers and sandboxed environments - [Getting Started Guide](docs/getting-started.md) - [Architecture Overview](docs/architecture.md) - [API Reference](docs/api-reference.md) - [CLI Command Reference](docs","readmeSource":"qudag/qudag-exchange/README.md","product":"key:qudag:qudag-exchange","canonical":"qudag-exchange","tier":"T1","t":["qudag-exchange","qudag","exchange","quantum-resistant","quantum","resistant","token","resource","utilization","voucher","blockchain"],"s":["qudag-exchange","qudag","exchange","blockchain","quantum","token"]},
+{"id":"qudag-exchange-server","family":"qudag-exchange-server","kind":"crate","store":"qudag","version":null,"description":"HTTP API server for QuDAG Exchange with rUv token operations","keywords":["api","exchange","qudag","server","tokens"],"source":"qudag/qudag-exchange/qudag-exchange-server/Cargo.toml","sourceSha256":"1b936230a41f870d99f3e1662a4ead60b6d215e365c67c4faa442b0ee878ba61","variants":[],"product":"key:qudag:qudag-exchange-server","canonical":"qudag-exchange-server","tier":"T1","t":["qudag-exchange-server","qudag","exchange","server","http","token","operation"],"s":["qudag-exchange-server","qudag","exchange","server","token"]},
+{"id":"qudag-exchange-sim","family":"qudag-exchange-sim","kind":"crate","store":"qudag","version":null,"description":"Simulation and testing tools for QuDAG Exchange","keywords":[],"source":"qudag/qudag-exchange/qudag-exchange-sim/Cargo.toml","sourceSha256":"d2ae8636345e4f5f4a3ed2d960725307418adf56da28b40e41754c42d03a3525","variants":[],"product":"key:qudag:qudag-exchange-sim","canonical":"qudag-exchange-sim","tier":"T1","t":["qudag-exchange-sim","qudag","exchange","sim","simulation","test"],"s":["qudag-exchange-sim","qudag","exchange","sim"]},
+{"id":"qudag-exchange-standalone-cli","family":"qudag-exchange-standalone","kind":"crate","store":"qudag","version":null,"description":"Standalone command-line interface for QuDAG Exchange with rUv tokens","keywords":["cli","exchange","qudag","standalone","tokens"],"source":"qudag/qudag-exchange/qudag-exchange-cli/Cargo.toml","sourceSha256":"bcc7b9edfb920555cb46739807e46fdd66708f33bf938c0956523ada0f572a4f","variants":[],"readme":"Standalone command-line interface for QuDAG Exchange with comprehensive rUv token management and business plan payout stream operations. - Account Management: Create and manage rUv token accounts - Token Transfers: Secure quantum-resistant token transfers - Balance Checking: Real-time account balance queries - Node Operations: Start, stop, and monitor exchange nodes - Payout Configuration: Enable/disable automatic fee distribution - Contributor Management: Register and manage contributors across all roles - Split Configuration: Customize payout percentages and thresholds - History Tracking: View complete payout history and audit trails - Peer Management: Connect to and manage network peers -","readmeSource":"qudag/qudag-exchange/qudag-exchange-cli/README.md","product":"key:qudag:qudag-exchange-standalone","canonical":"qudag-exchange-standalone-cli","tier":"T1","t":["qudag-exchange-standalone-cli","qudag","exchange","standalone","qudag-exchange-standalone","command-line","command","line","interface","token"],"s":["qudag-exchange-standalone-cli","qudag","exchange","standalone","qudag-exchange-standalone","token"]},
+{"id":"qudag-mcp","family":"qudag-mcp","kind":"crate","store":"qudag","version":null,"description":"Model Context Protocol (MCP) server for QuDAG - Integrates vault, exchange, and quantum-resistant operations","keywords":["mcp","protocol","qudag","server","tools"],"source":"qudag/qudag-mcp/Cargo.toml","sourceSha256":"70656af880e5261d51c3a37ff5b8e0554099f1ae25e932bf18b53b4d09095584","variants":[],"readme":"QuDAG MCP Server is a [Model Context Protocol (MCP)](https://spec.modelcontextprotocol.io/) server implementation that provides secure, quantum-resistant access to QuDAG's distributed system capabilities. - 🔒 Quantum-Resistant Security: Built on post-quantum cryptographic algorithms (ML-KEM, ML-DSA, HQC) - 📊 DAG Operations: Full access to QuDAG's Directed Acyclic Graph consensus and storage - 🔐 Secure Vault Integration: Encrypted secret management with DAG-based storage - 🌐 Network Management: P2P networking, dark addressing, and onion routing capabilities - 🔧 Comprehensive Tools: Rich set of MCP tools for all QuDAG operations - 📦 Resource Access: Structured access to system state and","readmeSource":"qudag/qudag-mcp/README.md","product":"key:qudag:qudag-mcp","canonical":"qudag-mcp","tier":"T1","t":["qudag-mcp","qudag","mcp","model","context","protocol","server","integrate","vault","exchange","quantum-resistant","quantum","resistant","operation"],"s":["qudag-mcp","qudag","mcp","protocol","server"]},
+{"id":"qudag-network","family":"qudag-network","kind":"crate","store":"qudag","version":null,"description":"P2P networking layer for QuDAG - LibP2P with onion routing, dark addressing, and quantum encryption","keywords":["anonymity","libp2p","networking","onion-routing","p2p"],"source":"qudag/core/network/Cargo.toml","sourceSha256":"c4ccc542f05c28d33f00720db163b451a91b6f7e9e1358c502a26c616766bf23","variants":[],"readme":"P2P networking layer with anonymous routing and dark addressing for the QuDAG protocol. - Anonymous Routing: Multi-hop onion routing with ML-KEM encryption - Dark Addressing: Decentralized .dark domain system - Traffic Obfuscation: ChaCha20Poly1305 traffic disguising - NAT Traversal: STUN/TURN/UPnP support for firewall penetration - Peer Discovery: Kademlia DHT-based peer discovery - LibP2P Integration: Production-ready P2P networking stack Add to your Cargo.toml: - P2P Node: Main networking coordinator - Onion Router: Anonymous multi-hop routing - Dark Resolver: Decentralized domain name system - NAT Traversal: Firewall and NAT penetration - Peer Discovery: Kademlia DHT for peer finding - T","readmeSource":"qudag/core/network/README.md","product":"key:qudag:qudag-network","canonical":"qudag-network","tier":"T1","t":["qudag-network","qudag","network","p2p","layer","libp2p","onion","rout","dark","address","quantum","encryption","anonymity","onion-routing"],"s":["qudag-network","qudag","network","anonymity","libp2p","onion-routing","onion","rout","p2p"]},
+{"id":"qudag-protocol","family":"qudag-protocol","kind":"crate","store":"qudag","version":null,"description":"Protocol implementation for QuDAG - Orchestrates crypto, DAG, and network components","keywords":["coordination","distributed-systems","protocol","state-machine"],"source":"qudag/core/protocol/Cargo.toml","sourceSha256":"8c5d307624458394253d1bd8578179187c1187a57fc56c5566908efb74a5d547","variants":[],"readme":"Protocol implementation and coordination layer for the QuDAG system. - Component Orchestration: Coordinates crypto, DAG, and network layers - Node Management: Complete node lifecycle management - Message Handling: Protocol message processing and routing - State Management: Persistent state storage and synchronization - RPC Server: JSON-RPC API for external communication - Configuration Management: Centralized configuration system Add to your Cargo.toml: - Node: Main coordinator for all subsystems - Message Factory: Creates and validates protocol messages - State Machine: Manages node state transitions - RPC Server: Provides JSON-RPC API - Persistence: State storage and recovery - Configurati","readmeSource":"qudag/core/protocol/README.md","product":"key:qudag:qudag-protocol","canonical":"qudag-protocol","tier":"T1","t":["qudag-protocol","qudag","protocol","implementation","orchestrate","crypto","dag","network","component","coordination","distributed-systems","distribut","state-machine","state","machine"],"s":["qudag-protocol","qudag","protocol","coordination","distributed-systems","distribut","state-machine","state","machine"]},
+{"id":"qudag-swarm","family":"qudag-swarm","kind":"crate","store":"qudag","version":null,"description":"Swarm coordination and multi-agent system for QuDAG","keywords":["coordination","distributed","multi-agent","swarm"],"source":"qudag/core/swarm/Cargo.toml","sourceSha256":"1a4572db73dbf727105731b1634504cc0135ae90dd0ebf16c8d1a63c333ee206","variants":[],"product":"key:qudag:qudag-swarm","canonical":"qudag-swarm","tier":"T1","t":["qudag-swarm","qudag","swarm","coordination","multi-agent","multi","distribut"],"s":["qudag-swarm","qudag","swarm","coordination","distribut","multi-agent","multi"]},
+{"id":"qudag-vault-core","family":"qudag-vault","kind":"crate","store":"qudag","version":null,"description":"Quantum-resistant password vault library with post-quantum cryptography","keywords":["cryptography","password-manager","post-quantum","security","vault"],"source":"qudag/core/vault/Cargo.toml","sourceSha256":"f4a32ca35827a6c077e2bf059c45d6df9ca22bbcc3372aa7da7fb6b028640c44","variants":[],"readme":"A quantum-resistant password vault library with post-quantum cryptography for secure password management. - 🔐 Quantum-Resistant: Uses ML-KEM (Kyber) and ML-DSA (Dilithium) for post-quantum security - 🛡️ Strong Encryption: AES-256-GCM with Argon2id key derivation - 🔒 Memory Safety: Automatic zeroization of sensitive data in memory - 📊 DAG Storage: Organize secrets in a directed acyclic graph structure - 🔑 Password Generation: Secure random password generation - 📦 Backup/Restore: Encrypted vault export and import functionality - ⚡ Fast Performance: Optimized cryptographic operations Add this to your Cargo.toml: Or install via cargo: The vault library is designed as a separate crate withi","readmeSource":"qudag/core/vault/README.md","product":"key:qudag:qudag-vault","canonical":"qudag-vault-core","tier":"T1","t":["qudag-vault-core","qudag","vault","core","qudag-vault","quantum-resistant","quantum","resistant","password","post-quantum","post","cryptography","password-manager","manager","security"],"s":["qudag-vault-core","qudag","vault","core","qudag-vault","cryptography","password-manager","password","manager","post-quantum","post","quantum","security"]},
+{"id":"qvm-scheduler","family":"qvm-scheduler","kind":"crate","store":"quantum-virtual-machine","version":null,"description":"High-performance quantum circuit scheduler for multi-job quantum computing with OpenQASM 3","keywords":["qasm","quantum","quantum-computing","scheduler","wasm"],"source":"quantum-virtual-machine/Cargo.toml","sourceSha256":"829895d2e321d536f44704dd05d8426c9f3a5fdd7e33c15444ecc864bc001ac1","variants":[],"readme":"- [Introduction](#-introduction) - [Features](#-features) - [Architecture](#-architecture) - [Installation](#-installation) - [Quick Start](#-quick-start) - [Advanced Usage](#-advanced-usage) - [SDK & API](#-sdk--api) - [Performance](#-performance) - [Contributing](#-contributing) - [License](#-license) The Quantum Virtual Machine (QVM) is a cutting-edge scheduler and runtime system designed to optimize the execution of multiple quantum circuits on quantum hardware. By intelligently partitioning device topologies and scheduling jobs with spatial and temporal multiplexing, QVM maximizes quantum resource utilization while minimizing crosstalk and execution time. - 🎯 Multi-Job Scheduling: Run","readmeSource":"quantum-virtual-machine/README.md","product":"repo:quantum-virtual-machine","canonical":"qvm-scheduler","tier":"T3","t":["qvm-scheduler","qvm","scheduler","high","quantum","circuit","multi-job","multi","job","comput","openqasm","qasm","quantum-computing"],"s":["qvm-scheduler","qvm","scheduler","qasm","quantum","quantum-computing","comput"]},
+{"id":"radio-moe","family":"radio-moe","kind":"npm","store":"autogenous","version":"0.3.1","description":"Radio real-time streaming peer expert mesh — bounded, signed, constitution-pinned distributed MoE. @metaharness/radio is the metadata-only control plane; a signed direct peer transport carries expert streams. True logit mixing where tokenizers match; safe text-primary racing otherwise.","keywords":["agents","gating","mixture-of-experts","moe","p2p","pubsub","realtime","streaming","top-k-routing","webrtc"],"source":"autogenous/packages/radio-moe/package.json","sourceSha256":"7b23b770c8db803ad74aab56bfb88016a4383ad06165aa50aaca70b0cd0ec57e","variants":[],"readme":"A real-time, peer-to-peer mixture-of-experts mesh. Heterogeneous experts live on different peers, advertise what they're good at, are routed per input chunk by a top-k gate, and their streamed output is combined in the *correct* regime — mathematically mixed when they share a tokenizer, raced as an ensemble when they don't. Above the raw stream sits a governed decision layer: signed typed frames, continuous claim/evidence fusion, independence-weighted quorum, and a constitution-pinned action gate — so the mesh can keep operating and improving while individual experts, models, and machines come and go. Two grounded primitives, kept strictly separate: True mixture-of-experts requires a shared","readmeSource":"autogenous/packages/radio-moe/README.md","product":"key:autogenous:radio-moe","canonical":"radio-moe","tier":null,"t":["radio-moe","radio","moe","real","stream","peer","expert","mesh","bound","sign","constitution-pinned","constitution","pin","distribut","metaharness","metadata-only","metadata","only","control","plane","direct","transport","carry","true","logit","mixing","tokenizer","match","safe","text-primary","text","primary","racing","otherwise","gating","mixture-of-experts","mixture","p2p","pubsub","realtime","top-k-routing","top","rout","webrtc"],"s":["radio-moe","radio","moe","gating","mixture-of-experts","mixture","expert","p2p","pubsub","realtime","stream","top-k-routing","top","rout","webrtc"]},
+{"id":"react-typescript","family":"react-typescript","kind":"npm","store":"r-flow","version":"1.0.0","description":"React and TypeScript example starter project","keywords":["react","starter","typescript"],"source":"r-flow/package.json","sourceSha256":"a02cdfd5355151987e8678ddbbd13c75041bbd1153be1b5f913f50d280ee3426","variants":[],"readme":"We've put together this template to serve as a starting point for folks interested in React Flow. You can use this sandbox to experiment with the library or share with us a bug report or issue. JavaScript not your thing? We also have a TypeScript starter template, just for you! You can fork and edit this sandbox in your browser. If you want to work offline or are looking for a template for real app development, check out our [Vite React Flow template](https://github.com/xyflow/vite-react-flow-template/tree/main). - Create a new custom node inside src/nodes/ (don't forget to export it from src/nodes/index.js). - Change how things look by [overriding some of the built-in classes](https://react","readmeSource":"r-flow/README.md","product":"repo:r-flow","canonical":"react-typescript","tier":"T3","t":["react-typescript","react","example","starter","project"],"s":["react-typescript","react","starter"]},
+{"id":"reasoningbank-mcp","family":"reasoningbank-mcp","kind":"crate","store":"agentic-flow","version":null,"description":"MCP server implementation with reasoning tools and resources","keywords":[],"source":"agentic-flow/reasoningbank/crates/reasoningbank-mcp/Cargo.toml","sourceSha256":"0e6d1ee7e3dec4a0251ec41470a3e2791224d4336696e91d1e4685e87c05ce00","variants":[],"product":"key:agentic-flow:reasoningbank-mcp","canonical":"reasoningbank-mcp","tier":"T1","t":["reasoningbank-mcp","reasoningbank","mcp","server","implementation","reason","resource"],"s":["reasoningbank-mcp","reasoningbank","mcp"]},
+{"id":"reasoningbank-network","family":"reasoningbank-network","kind":"crate","store":"agentic-flow","version":null,"description":"QUIC-based networking layer with 0-RTT and stream multiplexing","keywords":[],"source":"agentic-flow/reasoningbank/crates/reasoningbank-network/Cargo.toml","sourceSha256":"a277fef35fa09c256cd329b2fe63e1c4ac5e16531f68f02c544e44fa3d08bd49","variants":[],"product":"key:agentic-flow:reasoningbank-network","canonical":"reasoningbank-network","tier":"T1","t":["reasoningbank-network","reasoningbank","network","quic-based","quic","layer","0-rtt","rtt","stream","multiplex"],"s":["reasoningbank-network","reasoningbank","network"]},
+{"id":"reasoningbank-wasm","family":"reasoningbank","kind":"npm","store":"agentic-flow","version":"0.1.0","description":"WASM bindings for browser and Node.js environments","keywords":["ai","learning","mcp","quic","reasoning"],"source":"agentic-flow/agentic-flow/wasm/reasoningbank/package.json","sourceSha256":"ef052b16feb33e4c39b7c738e00613e84c46d99e15479b68e51f7d711c312b21","variants":["reasoningbank-core"],"product":"key:agentic-flow:reasoningbank","canonical":"reasoningbank-wasm","tier":"T1","t":["reasoningbank-wasm","reasoningbank","browser","environment","learn","mcp","quic","reason"],"s":["reasoningbank-wasm","reasoningbank","learn","mcp","quic","reason"]},
+{"id":"reflective-engineer-agent","family":"reflective-engineer-agent","kind":"npm","store":"reflective-engineer","version":"0.1.0","description":"Maintain and validate reflective-engineer","keywords":[],"source":"reflective-engineer/.harness/generated/package.json","sourceSha256":"5a0f52a26372ea8e4890d93e4652dbb6114ae985bb5c7a9bc60cfb490cbdff9f","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"reflective-engineer/.harness/generated/README.md","product":"key:reflective-engineer:reflective-engineer-agent","canonical":"reflective-engineer-agent","tier":"T3","t":["reflective-engineer-agent","reflective","engineer","maintain","validate","reflective-engineer"],"s":["reflective-engineer-agent","reflective","engineer"]},
+{"id":"roo-extension-runner","family":"roo-extension-runner","kind":"npm","store":"vsc-remote-mcp","version":"0.1.0","description":"Run Roo extension commands from the command line","keywords":[],"source":"vsc-remote-mcp/workspace/roo-extension/package.json","sourceSha256":"26f9f491da6c63a6698bb45d105444999e700a07406d1fcf58eb1bd81b12cb23","variants":[],"product":"key:vsc-remote-mcp:roo-extension-runner","canonical":"roo-extension-runner","tier":"T3","t":["roo-extension-runner","roo","extension","runner","command","line"],"s":["roo-extension-runner","roo","extension","runner"]},
+{"id":"rucelium-abi","family":"rucelium-abi","kind":"crate","store":"rucelium","version":null,"description":"RuCelium versioned C ABI boundary: rv_env_sample_v1 wire format (bounds-checked, allocation-free parse), deterministic CBOR, ed25519 signed record envelope (ADR-264 §11)","keywords":["cbor","environmental","ffi","iot","wire"],"source":"rucelium/crates/rucelium-abi/Cargo.toml","sourceSha256":"fd8e6cce7621c7ef388b678fd4ff3f13d3da9a918fbc5f3b708bd907749e5d8e","variants":[],"product":"key:rucelium:rucelium-abi","canonical":"rucelium-abi","tier":null,"t":["rucelium-abi","rucelium","abi","version","boundary","env","sample","wire","format","bounds-checked","bound","check","allocation-free","allocation","free","parse","deterministic","cbor","ed25519","sign","record","envelope","adr-264","adr","264","environmental","ffi","iot"],"s":["rucelium-abi","rucelium","abi","cbor","environmental","ffi","iot","wire"]},
+{"id":"rucelium-calibration","family":"rucelium-calibration","kind":"crate","store":"rucelium","version":null,"description":"RuCelium calibration lineage, drift detection, and sensor quarantine — never silent correction (ADR-264 §12)","keywords":["calibration","drift","environmental","quality"],"source":"rucelium/crates/rucelium-calibration/Cargo.toml","sourceSha256":"07a72d85b35bfbbe3406cde6973d44b2d93374e1642ceedd8cd86b6b2dc10c90","variants":[],"product":"key:rucelium:rucelium-calibration","canonical":"rucelium-calibration","tier":null,"t":["rucelium-calibration","rucelium","calibration","lineage","drift","detection","sensor","quarantine","never","silent","correction","adr-264","adr","264","environmental","quality"],"s":["rucelium-calibration","rucelium","calibration","drift","environmental","quality"]},
+{"id":"rucelium-core","family":"rucelium","kind":"crate","store":"rucelium","version":null,"description":"RuCelium federated environmental fabric core data model: EnvSample, EnvFrame, CalibrationRecord, EnvironmentalEvent, SensorModality, GeoPoint, DataClass (ADR-264 §5/§7/§10)","keywords":["calibration","environmental","federation","iot","sensing"],"source":"rucelium/crates/rucelium-core/Cargo.toml","sourceSha256":"eb9323709486b5cbc0b90bb8d507a74672fead0c0aefe6ec583126f1a1f07b5d","variants":[],"product":"repo:rucelium","canonical":"rucelium-core","tier":null,"t":["rucelium-core","rucelium","core","federat","environmental","fabric","model","envsample","envframe","calibrationrecord","environmentalevent","sensormodality","geopoint","dataclass","adr-264","adr","264","calibration","federation","iot","sens"],"s":["rucelium-core","rucelium","core","calibration","environmental","federation","iot","sens"]},
+{"id":"rucelium-federation","family":"rucelium-federation","kind":"crate","store":"rucelium","version":null,"description":"RuCelium biome sovereignty: outage buffer with duplicate-free replay, signed regional summaries, device revocation, disclosure coarsening, OGC SensorThings projection (ADR-264 §6/§7/§10)","keywords":["environmental","federation","sensorthings","sovereignty"],"source":"rucelium/crates/rucelium-federation/Cargo.toml","sourceSha256":"1533799d17fb3f20b6fb3488656da37f06db3495483aa026f905c982626311dd","variants":[],"product":"key:rucelium:rucelium-federation","canonical":"rucelium-federation","tier":null,"t":["rucelium-federation","rucelium","federation","biome","sovereignty","outage","buffer","duplicate-free","duplicate","free","replay","sign","regional","summary","device","revocation","disclosure","coarsen","ogc","sensorthing","projection","adr-264","adr","264","environmental"],"s":["rucelium-federation","rucelium","federation","environmental","sensorthing","sovereignty"]},
+{"id":"rucelium-gateway","family":"rucelium-gateway","kind":"crate","store":"rucelium","version":null,"description":"RuCelium rhizome gateway daemon: UDP envelope ingestion (v1/v2/fragments), calibration, durable store, WorldGraph, local alerts, OGC SensorThings HTTP API, and biome federation sync over the network (ADR-265 §4)","keywords":["daemon","environmental","gateway","sensorthings"],"source":"rucelium/crates/rucelium-gateway/Cargo.toml","sourceSha256":"94e2d02e2166178a40c00a99f74cf7043bc1e26be281527a8adbd6a87187d7ce","variants":[],"product":"key:rucelium:rucelium-gateway","canonical":"rucelium-gateway","tier":null,"t":["rucelium-gateway","rucelium","gateway","rhizome","daemon","udp","envelope","ingestion","fragment","calibration","durable","store","worldgraph","local","alert","ogc","sensorthing","http","biome","federation","sync","network","adr-265","adr","265","environmental"],"s":["rucelium-gateway","rucelium","gateway","daemon","environmental","sensorthing"]},
+{"id":"rucelium-harness","family":"rucelium-harness","kind":"npm","store":"rucelium","version":"0.1.0","description":"RuCelium metaharness — a Darwinian evolve/evaluate/select flywheel that drives the RuCelium Rust implementation toward the ADR-264 acceptance gate.","keywords":["darwin","environmental-sensing","flywheel","metaharness","rucelium","rufield","rust"],"source":"rucelium/harness/package.json","sourceSha256":"24d90f4448229a67a7d08e5b030315ff1f5c51291836a10912c50fc1babf71f9","variants":[],"readme":"The RuCelium metaharness — a Darwinian flywheel that drives the RuCelium Rust implementation toward the [ADR-264](../docs/ADR-264-rucelium-federated-fabric.md) §14 acceptance gate. It is a *meta*harness: it adds no new truth of its own. Every score is derived from the same commands CI runs — cargo test --workspace, cargo clippy, and the deterministic 64-node biome benchmark (rucelium-bench). What it adds is the loop: Regressions don't get recorded — a change that lowers fitness is simply not selected (override deliberately with --allow-regression, which records the regression honestly rather than hiding it). Over iterations the ledger is the lineage of surviving implementations: a fitness cu","readmeSource":"rucelium/harness/README.md","product":"key:rucelium:rucelium-harness","canonical":"rucelium-harness","tier":null,"t":["rucelium-harness","rucelium","harness","metaharness","darwinian","evolve","evaluate","select","flywheel","drive","implementation","toward","adr-264","adr","264","acceptance","gate","darwin","environmental-sensing","environmental","sens","rufield"],"s":["rucelium-harness","rucelium","harness","darwin","environmental-sensing","environmental","sens","flywheel","metaharness","rufield"]},
+{"id":"rucelium-ingest","family":"rucelium-ingest","kind":"crate","store":"rucelium","version":null,"description":"RuCelium rhizome-gateway ingest pipeline: envelope decode, signature + revocation + replay-window checks, normalization into EnvSample (ADR-264 §5)","keywords":["environmental","gateway","ingest","security"],"source":"rucelium/crates/rucelium-ingest/Cargo.toml","sourceSha256":"3abe80c9491f803c2e433b1bb0eb9069d3f09b559a619ed192fb58ce52355ced","variants":[],"product":"key:rucelium:rucelium-ingest","canonical":"rucelium-ingest","tier":null,"t":["rucelium-ingest","rucelium","ingest","rhizome-gateway","rhizome","gateway","pipeline","envelope","decode","signature","revocation","replay-window","replay","window","check","normalization","envsample","adr-264","adr","264","environmental","security"],"s":["rucelium-ingest","rucelium","ingest","environmental","gateway","security"]},
+{"id":"rucelium-notary","family":"rucelium-notary","kind":"crate","store":"rucelium","version":null,"description":"RuCelium long-term provenance: domain-separated Merkle notarization with inclusion proofs and algorithm-agile root signing, amortizing post-quantum signature cost to bytes per observation (ADR-267)","keywords":["archival","environmental","merkle","post-quantum","provenance"],"source":"rucelium/crates/rucelium-notary/Cargo.toml","sourceSha256":"8d7d99ec2ee4fc91a6dec021ce8f0c76913e5710043c3cc2ba19dcc26d0afcec","variants":[],"product":"key:rucelium:rucelium-notary","canonical":"rucelium-notary","tier":null,"t":["rucelium-notary","rucelium","notary","long-term","long","term","provenance","domain-separated","domain","separat","merkle","notarization","inclusion","proof","algorithm-agile","algorithm","agile","root","sign","amortiz","post-quantum","post","quantum","signature","cost","byte","per","observation","adr-267","adr","267","archival","environmental"],"s":["rucelium-notary","rucelium","notary","archival","environmental","merkle","post-quantum","post","quantum","provenance"]},
+{"id":"rucelium-policy","family":"rucelium-policy","kind":"crate","store":"rucelium","version":null,"description":"RuCelium governed control path: agent proposal -> policy -> safety simulation -> authority -> signed command -> gateway validation -> execution receipt, typed so no stage can be skipped (ADR-264 §9)","keywords":["agents","environmental","governance","policy"],"source":"rucelium/crates/rucelium-policy/Cargo.toml","sourceSha256":"9592e5bd164e9af145be946d03fd207e20ef6009ffbde74775273eefdc4540be","variants":[],"product":"key:rucelium:rucelium-policy","canonical":"rucelium-policy","tier":null,"t":["rucelium-policy","rucelium","policy","govern","control","path","proposal","safety","simulation","authority","sign","command","gateway","validation","execution","receipt","typed","stage","skip","adr-264","adr","264","environmental","governance"],"s":["rucelium-policy","rucelium","policy","environmental","governance"]},
+{"id":"rucelium-store","family":"rucelium-store","kind":"crate","store":"rucelium","version":null,"description":"RuCelium durable gateway store: append-only segment log with dedup index, crash recovery, deterministic replay, and per-DataClass retention enforcement (ADR-265 §3)","keywords":["environmental","log","retention","storage"],"source":"rucelium/crates/rucelium-store/Cargo.toml","sourceSha256":"39006e5230351fdf8f9afd3392d5c86cc2afd1d8cb6684c98e455513e4845b71","variants":[],"product":"key:rucelium:rucelium-store","canonical":"rucelium-store","tier":null,"t":["rucelium-store","rucelium","store","durable","gateway","append-only","append","only","segment","log","dedup","index","crash","recovery","deterministic","replay","per-dataclass","per","dataclass","retention","enforcement","adr-265","adr","265","environmental","storage"],"s":["rucelium-store","rucelium","store","environmental","log","retention","storage"]},
+{"id":"rucelium-transport","family":"rucelium-transport","kind":"crate","store":"rucelium","version":null,"description":"RuCelium constrained-link transport: compact 114-byte envelope v2 (pubkey by reference) and MTU fragmentation/reassembly sized for LoRaWAN DR0's 51-byte payload cap (ADR-265 §2)","keywords":["environmental","fragmentation","iot","lorawan"],"source":"rucelium/crates/rucelium-transport/Cargo.toml","sourceSha256":"c53e73a69072999bf315c1507e42f4548b008a35846a88d821a256dbc5ad99f2","variants":[],"product":"key:rucelium:rucelium-transport","canonical":"rucelium-transport","tier":null,"t":["rucelium-transport","rucelium","transport","constrained-link","constrain","link","compact","114-byte","114","byte","envelope","pubkey","reference","mtu","fragmentation","reassembly","sized","lorawan","dr0","51-byte","payload","cap","adr-265","adr","265","environmental","iot"],"s":["rucelium-transport","rucelium","transport","environmental","fragmentation","iot","lorawan"]},
+{"id":"rucelium-worldgraph","family":"rucelium-worldgraph","kind":"crate","store":"rucelium","version":null,"description":"RuCelium environmental WorldGraph: typed sensor/ecosystem nodes, geospatial registration, evidence + contradiction edges, RuView RF context bridge (ADR-264 §5.2/§8)","keywords":["digital-twin","environmental","fusion","graph"],"source":"rucelium/crates/rucelium-worldgraph/Cargo.toml","sourceSha256":"71aac88cb01b0907aeb84ae9fd2147f3531fb3746876e2dc7225d5e2af3033dd","variants":[],"product":"key:rucelium:rucelium-worldgraph","canonical":"rucelium-worldgraph","tier":null,"t":["rucelium-worldgraph","rucelium","worldgraph","environmental","typed","sensor","ecosystem","geospatial","registration","evidence","contradiction","edge","ruview","context","bridge","adr-264","adr","264","5.2","digital-twin","digital","twin","fusion","graph"],"s":["rucelium-worldgraph","rucelium","worldgraph","digital-twin","digital","twin","environmental","fusion","graph"]},
+{"id":"rufield-adapters","family":"rufield-adapters","kind":"crate","store":"rucelium","version":null,"description":"RuField MFS adapters: deterministic seeded SyntheticSim emitting the §19 camera-free room-intelligence demo across 3 modalities","keywords":["csi","radar","sensing","simulator","thermal"],"source":"rucelium/crates/rufield-adapters/Cargo.toml","sourceSha256":"fb352a8c4a0c6b15fda711f2e17072a0ac30e53f025f0a305003aca4c6bfe4e7","variants":[],"product":"key:rucelium:rufield-adapters","canonical":"rufield-adapters","tier":null,"t":["rufield-adapters","rufield","adapter","mfs","deterministic","seed","syntheticsim","emitt","camera-free","camera","free","room-intelligence","room","intelligence","demo","across","modality","csi","radar","sens","simulator","thermal"],"s":["rufield-adapters","rufield","adapter","csi","radar","sens","simulator","thermal"]},
+{"id":"rufield-core","family":"rufield","kind":"crate","store":"rucelium","version":null,"description":"RuField MFS v0.1 core data model: Field Event/Tensor/Privacy/Provenance types and traits (ADR-260 §7/§9/§16/§20)","keywords":["csi","multimodal","privacy","rf","sensing"],"source":"rucelium/crates/rufield-core/Cargo.toml","sourceSha256":"e9953ec787d0a316f8e6d0bf95f2032433011652c13aef1722b27bcfaacfdaab","variants":[],"product":"key:rufield","canonical":"rufield-core","tier":null,"t":["rufield-core","rufield","core","mfs","v0.1","model","field","event","tensor","privacy","provenance","type","trait","adr-260","adr","260","csi","multimodal","sens"],"s":["rufield-core","rufield","core","csi","multimodal","privacy","sens"]},
+{"id":"rufield-fusion","family":"rufield-fusion","kind":"crate","store":"rucelium","version":null,"description":"RuField MFS fusion graph + engine: TOML rules, weighted-Bayes / temporal-window inference, confidence decay (ADR-260 §12/§13/§24)","keywords":["bayes","fusion","inference","sensing"],"source":"rucelium/crates/rufield-fusion/Cargo.toml","sourceSha256":"83b79ad3da1d0404b811ea8565fda2534ea188e6ee6e3534ff22073cf75b3576","variants":[],"product":"key:rucelium:rufield-fusion","canonical":"rufield-fusion","tier":null,"t":["rufield-fusion","rufield","fusion","mfs","graph","toml","rule","weighted-bayes","weight","baye","temporal-window","temporal","window","inference","confidence","decay","adr-260","adr","260","sens"],"s":["rufield-fusion","rufield","fusion","baye","inference","sens"]},
+{"id":"rufield-interop","family":"rufield-interop","kind":"crate","store":"rufield","version":null,"description":"Deterministic CloudEvents and SOSA projections for RuField events","keywords":["cloudevents","sensing","sosa"],"source":"rufield/crates/rufield-interop/Cargo.toml","sourceSha256":"a5dca011712f325e13d2dc2f26db4cde0043610399755bfc6ab2d25f63d5521b","variants":[],"product":"key:rufield:rufield-interop","canonical":"rufield-interop","tier":"T1","t":["rufield-interop","rufield","interop","deterministic","cloudevent","sosa","projection","event","sens"],"s":["rufield-interop","rufield","interop","cloudevent","sens","sosa"]},
+{"id":"rufield-privacy","family":"rufield-privacy","kind":"crate","store":"rucelium","version":null,"description":"RuField MFS privacy policy + PrivacyGuard: P0..P5 classes, default policy, consent and identity gating (ADR-260 §10)","keywords":["consent","policy","privacy","sensing"],"source":"rucelium/crates/rufield-privacy/Cargo.toml","sourceSha256":"9ef6e42bf8b0ca6fc7f11c0aa6469cb693a6054a9edc02674686bfa04ae9f63c","variants":[],"product":"key:rucelium:rufield-privacy","canonical":"rufield-privacy","tier":null,"t":["rufield-privacy","rufield","privacy","mfs","policy","privacyguard","p0..p5","class","default","consent","identity","gating","adr-260","adr","260","sens"],"s":["rufield-privacy","rufield","privacy","consent","policy","sens"]},
+{"id":"rufield-provenance","family":"rufield-provenance","kind":"crate","store":"rucelium","version":null,"description":"RuField MFS provenance receipts: sha256 hashing + ed25519 sign/verify and the §11 fusability invariant","keywords":["ed25519","provenance","sensing","sha256"],"source":"rucelium/crates/rufield-provenance/Cargo.toml","sourceSha256":"1186620a9d6b548e6cf17d44f07c007dd8161b59cc254a7aa505c7ea3440ae07","variants":[],"product":"key:rucelium:rufield-provenance","canonical":"rufield-provenance","tier":null,"t":["rufield-provenance","rufield","provenance","mfs","receipt","sha256","hash","ed25519","sign","verify","fusability","invariant","sens"],"s":["rufield-provenance","rufield","provenance","ed25519","sens","sha256"]},
+{"id":"rufield-ruvector","family":"rufield-ruvector","kind":"crate","store":"rufield","version":null,"description":"Optional backend boundary between RuField FieldEncoder and RuVector embeddings","keywords":["embedding","ruvector","sensing"],"source":"rufield/crates/rufield-ruvector/Cargo.toml","sourceSha256":"ba9976c671d45032d53e6f3102cd687913d94835662071afeeb95e9952469494","variants":[],"product":"key:rufield:rufield-ruvector","canonical":"rufield-ruvector","tier":"T1","t":["rufield-ruvector","rufield","optional","backend","boundary","between","fieldencoder","embedding","embedd","sens"],"s":["rufield-ruvector","rufield","embedd","sens"]},
+{"id":"rufield-uncertainty","family":"rufield-uncertainty","kind":"crate","store":"rufield","version":null,"description":"Model independent uncertainty and split conformal calibration for RuField","keywords":["conformal","sensing","uncertainty"],"source":"rufield/crates/rufield-uncertainty/Cargo.toml","sourceSha256":"07051698127f9baea8f4054b234a1db218feee7dbc17a7cbf194e4a5dd5b0bf3","variants":[],"product":"key:rufield:rufield-uncertainty","canonical":"rufield-uncertainty","tier":"T1","t":["rufield-uncertainty","rufield","uncertainty","model","independent","split","conformal","calibration","sens"],"s":["rufield-uncertainty","rufield","uncertainty","conformal","sens"]},
+{"id":"rufield-viewer","family":"rufield-viewer","kind":"crate","store":"rucelium","version":null,"description":"RuField MFS read-only web dashboard: drives the SyntheticSim → fusion pipeline and streams live camera-free room intelligence (ADR-260 §14 Layer 7 / §27.9)","keywords":["axum","dashboard","fusion","sensing"],"source":"rucelium/crates/rufield-viewer/Cargo.toml","sourceSha256":"e7d53f2dbe1a7acba94e1b265118da50375f3dfbf4411b50cc303e98bc9d4c77","variants":[],"product":"key:rucelium:rufield-viewer","canonical":"rufield-viewer","tier":null,"t":["rufield-viewer","rufield","viewer","mfs","read-only","read","only","web","dashboard","drive","syntheticsim","fusion","pipeline","stream","live","camera-free","camera","free","room","intelligence","adr-260","adr","260","layer","27.9","axum","sens"],"s":["rufield-viewer","rufield","viewer","axum","dashboard","fusion","sens"]},
+{"id":"ruflo","family":"ruflo","kind":"npm","store":"ruflo","version":"3.49.0","description":"Ruflo - Enterprise AI agent orchestration platform. Deploy 60+ specialized agents in coordinated swarms with self-learning, fault-tolerant consensus, vector memory, and MCP integration","keywords":["agent-orchestration","ai","ai-agents","anthropic","claude","claude-code","claude-flow","cli","developer-tools","enterprise","mcp","model-context-protocol","multi-agent","ruflo","self-learning","swarm","swarm-intelligence"],"source":"ruflo/ruflo/package.json","sourceSha256":"a0b465474b2c9aac77b7436b90a4e16f124c5e28f3b00f381948cfbc2247a052","variants":[],"readme":"An agent meta-harness for Claude Code and Codex. [📖 RuFlo Explained — Build an AI Team That Plans, Remembers, Tests, and Improves](docs/ruflo-explained.md) A 14-chapter guide: from the basic idea to a first useful task, then memory, agent teams, plugins, cost and verification.","readmeSource":"ruflo/ruflo/README.md","product":"repo:ruflo","canonical":"ruflo","tier":"T0","t":["ruflo","enterprise","orchestration","deploy","specializ","coordinat","swarm","self-learning","self","learn","fault-tolerant","fault","tolerant","consensus","vector","memory","mcp","integration","agent-orchestration","ai-agents","anthropic","claude","claude-code","code","claude-flow","flow","developer-tools","developer","model-context-protocol","model","context","protocol","multi-agent","multi","swarm-intelligence","intelligence"],"s":["ruflo","agent-orchestration","orchestration","ai-agents","anthropic","claude","claude-code","code","claude-flow","flow","developer-tools","developer","enterprise","mcp","model-context-protocol","model","context","protocol","multi-agent","multi","self-learning","self","learn","swarm","swarm-intelligence","intelligence"]},
+{"id":"ruflo-agntcy","family":"ruflo-agntcy","kind":"crate","store":"ruflo","version":null,"description":"AGNTCY/Outshift runtime integration for ruflo: CASA intent-scoped authorization enforcement + SLIM transport surface (ADR-380)","keywords":["agntcy","authorization","casa","ruflo","slim"],"source":"ruflo/v3/crates/ruflo-agntcy/Cargo.toml","sourceSha256":"9721df1089db7437f75841a470e4563385c2af77f44b8804f6da2f51da4cfa70","variants":[],"product":"key:ruflo:ruflo-agntcy","canonical":"ruflo-agntcy","tier":"T0","t":["ruflo-agntcy","ruflo","agntcy","outshift","runtime","integration","casa","intent-scoped","intent","scop","authorization","enforcement","slim","transport","surface","adr-380","adr","380"],"s":["ruflo-agntcy","ruflo","agntcy","authorization","casa","slim"]},
+{"id":"ruflo-arena","family":"ruflo-arena","kind":"npm","store":"ruflo","version":"0.1.0","description":"Competitive ruliology for Ruflo swarms — arenas, tournaments, and adaptive co-evolution of program strategies (ADR-147/148).","keywords":["arena","co-evolution","ruflo","ruliology","tournament"],"source":"ruflo/plugins/ruflo-arena/package.json","sourceSha256":"a1c613a488f88daee23f64de718ef123eb5470b41b9b0083ad645254e56e166f","variants":[],"readme":"Competitive ruliology for Ruflo swarms — arenas, tournaments, and adaptive co-evolution of program strategies. Implements the first executable slice of Ruflo ADR-147/148/150, following Stephen Wolfram's [*Games Between Programs: The Ruliology of Competition*](https://writings.stephenwolfram.com/2026/06/games-between-programs-the-ruliology-of-competition/). Strategies are programs (deterministic finite-state machines reading opponent history). They compete under payoff games; tournaments produce Wolfram's competitive array; hill-climb and mutual co-evolution discover winners empirically — because, per computational irreducibility, you have to *run* the competition to find out. Sample PD ranki","readmeSource":"ruflo/plugins/ruflo-arena/README.md","product":"key:ruflo:ruflo-arena","canonical":"ruflo-arena","tier":"T0","t":["ruflo-arena","ruflo","arena","competitive","ruliology","swarm","tournament","adaptive","co-evolution","evolution","program","strategy","adr-147","adr","147","148"],"s":["ruflo-arena","ruflo","arena","co-evolution","evolution","ruliology","tournament"]},
+{"id":"ruflo-federation-peer","family":"ruflo-federation-peer","kind":"crate","store":"ruflo","version":null,"description":"Single-process federation peer for ruflo: midstreamer-quic transport composed with aimds-* 3-gate safety pipeline (ADR-120 Step 3)","keywords":["agent","aimds","federation","quic","ruflo"],"source":"ruflo/v3/crates/ruflo-federation-peer/Cargo.toml","sourceSha256":"cb52ea49931c361917513b1521c258617bdcec8722488309d17124044a43c2b4","variants":[],"product":"key:ruflo:ruflo-federation-peer","canonical":"ruflo-federation-peer","tier":"T0","t":["ruflo-federation-peer","ruflo","federation","peer","single-process","single","process","midstreamer-quic","midstreamer","quic","transport","compos","aimd","3-gate","gate","safety","pipeline","adr-120","adr","120","step"],"s":["ruflo-federation-peer","ruflo","federation","peer","aimd","quic"]},
+{"id":"ruflo-graph-intelligence","family":"ruflo-graph-intelligence","kind":"npm","store":"ruflo","version":"0.1.0-alpha.1","description":"RuFlo Graph Intelligence Engine — real-time relationship intelligence with complexity-aware execution. Single-entry personalized PageRank, streaming delta updates, witness-signed reasoning artifacts over RuFlo's substrate graphs.","keywords":["agent-substrate","claude-flow","complexity-aware","graph-intelligence","mcp","pagerank","personalized-pagerank","ruflo","sublinear"],"source":"ruflo/plugins/ruflo-graph-intelligence/package.json","sourceSha256":"0c669e1010cfd8ee62a0b4ff7a50c32cabe0cb403d372b445f2f211652480091","variants":[],"product":"key:ruflo:ruflo-graph-intelligence","canonical":"ruflo-graph-intelligence","tier":"T0","t":["ruflo-graph-intelligence","ruflo","graph","intelligence","real","relationship","complexity-aware","complexity","aware","execution","single-entry","single","entry","personaliz","pagerank","stream","delta","update","witness-signed","witness","sign","reason","artifact","substrate","agent-substrate","claude-flow","claude","flow","graph-intelligence","mcp","personalized-pagerank","sublinear"],"s":["ruflo-graph-intelligence","ruflo","graph","intelligence","agent-substrate","substrate","claude-flow","claude","flow","complexity-aware","complexity","aware","graph-intelligence","mcp","pagerank","personalized-pagerank","personaliz","sublinear"]},
+{"id":"ruflo-watermark","family":"ruflo-watermark","kind":"crate","store":"ai-text-watermark","version":null,"description":"High-speed SynthID-Text-style LLM text watermarking (Tournament + distortion-free Gumbel) with Bayesian detection and a robustness-evaluation harness. Rust core + optional WASM bindings.","keywords":["llm","provenance","synthid","wasm","watermark"],"source":"ai-text-watermark/crates/ruflo-watermark/Cargo.toml","sourceSha256":"205a92683fcaa992d84b029df3e014c67bb5f173cbb04fe063fdc217b98492ec","variants":[],"readme":"High-speed SynthID-Text-style LLM text watermarking (generation + detection) in Rust, with optional WebAssembly bindings. A watermark rides the tie-break randomness among *already-plausible* tokens: it never injects an out-of-distribution word, costs no extra tokens, and is detectable only with the key. - Schemes: Gumbel (per-instance distortion-free), Tournament (SynthID), TournamentNd (key-averaged non-distortionary). - Detectors: standard, indel-robust self-sync, exact-Gamma short-text, Higher-Criticism. - Extras: a robustness-evaluation harness, a Darwin/flywheel detector-param tuner, and an authorized un-marked-generation governance path. - StreamProxy: an ultra-low-latency decode-loop","readmeSource":"ai-text-watermark/crates/ruflo-watermark/README.md","product":"key:ai-text-watermark:ruflo-watermark","canonical":"ruflo-watermark","tier":null,"t":["ruflo-watermark","ruflo","watermark","high-speed","high","speed","synthid-text-style","synthid","text","style","tournament","distortion-free","distortion","free","gumbel","bayesian","detection","robustness-evaluation","robustness","evaluation","harness","core","optional","provenance"],"s":["ruflo-watermark","ruflo","watermark","provenance","synthid"]},
+{"id":"ruforecast-autogenous-bridge","family":"ruforecast-autogenous-bridge","kind":"crate","store":"ruview","version":null,"description":"LOCAL-DEV-ONLY bridge: signs/verifies a Darwin-found RuForecast hyperparameter candidate through Autogenous's regression-candidate promotion path (ruvnet/autogenous, not yet published/vendored). Path-dependency on a sibling local checkout; not part of any release build.","keywords":[],"source":"ruview/v2/crates/ruforecast-autogenous-bridge/Cargo.toml","sourceSha256":"78982cb5f58794b7d5a9fb50e936841125ed19bfb147dc3df35340a5999698c5","variants":[],"product":"key:ruview:ruforecast-autogenous-bridge","canonical":"ruforecast-autogenous-bridge","tier":"T0","t":["ruforecast-autogenous-bridge","ruforecast","autogenous","bridge","local-dev-only","local","dev","only","sign","verify","darwin-found","darwin","found","hyperparameter","candidate","through","regression-candidate","regression","promotion","path","yet","publish","vendor","path-dependency","dependency","sibl","checkout","part","any","release"],"s":["ruforecast-autogenous-bridge","ruforecast","autogenous","bridge"]},
+{"id":"rulake","family":"rulake","kind":"npm","store":"rulake","version":"2.3.0-alpha.1","description":"Memory Lake for Agentic AI — fast, witness-anchored vector memory for LLM agents. Cache layer for RAG, semantic search, federated retrieval over S3/BigQuery/Parquet, MCP-compatible. Node.js bindings.","keywords":["agent","agent-memory","agentic","ai","ann","anthropic","approximate-nearest-neighbor","cache","claude","datalake","edge-ai","embedding","embedding-cache","embeddings","federation","lakehouse","langchain","llamaindex","llm","mcp","memory","model-context-protocol","napi","napi-rs","openai","provenance","quantization","rabitq","rag","retrieval-augmented-generation","rust","ruvector","semantic-search","vector","vector-database","vector-search","witness"],"source":"rulake/sdk/node/package.json","sourceSha256":"a2e78f1dffd681971d09f459d7ca998ad7cc66dd776e119552bf7667e496f575","variants":["rulake-node","rulake-wasm"],"readme":"Fast, witness-anchored vector memory for LLM agents — without standing up a vector database. ruLake sits between your AI agent and the data it remembers (S3, BigQuery, Snowflake, Parquet, files, RVF). Every retrieval is served from a compressed in-memory cache (RaBitQ 1-bit quantization) at ≈1.02× raw library speed, anchored by a SHAKE-256 cryptographic witness so two agents on two hosts share one byte-exact view of memory. Native Node.js bindings via [napi-rs](https://napi.rs). Implements [ADR-003](https://github.com/ruvnet/RuLake/blob/main/docs/adrs/sdk/ADR-003-nodejs-typescript-sdk.md). The companion [rulake-mcp](https://github.com/ruvnet/RuLake/tree/main/mcp-server) binary speaks the [Mo","readmeSource":"rulake/sdk/node/README.md","product":"repo:rulake","canonical":"rulake","tier":"T1","t":["rulake","memory","lake","witness-anchored","witness","anchor","vector","cache","layer","rag","semantic","search","federat","retrieval","bigquery","parquet","mcp-compatible","mcp","compatible","agent-memory","ann","anthropic","approximate-nearest-neighbor","approximate","nearest","neighbor","claude","datalake","edge-ai","edge","embedd","embedding-cache","embedding","federation","lakehouse","langchain","llamaindex","model-context-protocol","model","context","protocol","openai","provenance","quantization","rabitq","retrieval-augmented-generation","augment","generation","semantic-search","vector-database","database","vector-search"],"s":["rulake","agent-memory","memory","ann","anthropic","approximate-nearest-neighbor","approximate","nearest","neighbor","cache","claude","datalake","edge-ai","edge","embedd","embedding-cache","embedding","federation","lakehouse","langchain","llamaindex","mcp","model-context-protocol","model","context","protocol","openai","provenance","quantization","rabitq","rag","retrieval-augmented-generation","retrieval","augment","generation","semantic-search","semantic","search","vector","vector-database","database","vector-search","witness"]},
+{"id":"rulake-py","family":"rulake-py","kind":"crate","store":"rulake","version":null,"description":"Python bindings for ruLake — vector cache + federation intermediary","keywords":[],"source":"rulake/sdk/python/Cargo.toml","sourceSha256":"d1613dcb9b5a5158e0fff8dacae015d935b6718bda8bf1ef2e7c0742587b8292","variants":[],"readme":"Fast, witness-anchored vector memory for LLM agents — without standing up a vector database. ruLake sits between your AI agent and the data it remembers (S3, BigQuery, Snowflake, Parquet, files, RVF). Every retrieval is served from a compressed in-memory cache (RaBitQ 1-bit quantization) at ≈1.02× raw library speed, anchored by a SHAKE-256 cryptographic witness so two agents on two hosts share one byte-exact view of memory. Native Python bindings via [PyO3](https://pyo3.rs) and [maturin](https://www.maturin.rs). Implements [ADR-002](https://github.com/ruvnet/RuLake/blob/main/docs/adrs/sdk/ADR-002-python-sdk.md). The companion [rulake-mcp](https://github.com/ruvnet/RuLake/tree/main/mcp-server","readmeSource":"rulake/sdk/python/README.md","product":"key:rulake:rulake-py","canonical":"rulake-py","tier":"T1","t":["rulake-py","rulake","python","vector","cache","federation","intermediary"],"s":["rulake-py","rulake"]},
+{"id":"runtime","family":"runtime","kind":"crate","store":"autogenous","version":null,"description":"Autogenous runtime: the self-running observe→generate→evaluate→verify→canary→promote/rollback loop with a Clock abstraction and measured SLO timing (ADR-392 §9/§14, ADR-393 acceptance, gap 7).","keywords":[],"source":"autogenous/crates/runtime/Cargo.toml","sourceSha256":"a3c0182d6c9aada0cd9b78dbb68d5751eb4f7454138c324f330770535e6b6be1","variants":[],"readme":"runtime is the engine of [Autogenous](../), a self-evolving AI-defense runtime in Rust. It runs the whole loop end to end: a novel attack becomes a population of generated candidate defenses, each independently judged and signed, the best cryptographically verified, then canary-rolled-out with timed, automatic rollback on any regression — and it measures the SLOs while doing it. The runtime never fabricates a fitness number. Judges measure candidate vs. parent on the same corpus and sign the results; the controller signs the promotion; the verifier admits only on independently verified, content-bound evidence. Trust is earned per step, cryptographically. Runtime::defend(evidence) runs observ","readmeSource":"autogenous/crates/runtime/README.md","product":"key:autogenous:runtime","canonical":"runtime","tier":null,"t":["runtime","autogenous","self-running","self","observe","generate","evaluate","verify","canary","promote","rollback","loop","clock","abstraction","measur","slo","timing","adr-392","adr","392","adr-393","393","acceptance","gap"],"s":["runtime"]},
+{"id":"ruos-mcp","family":"ruos-mcp","kind":"crate","store":"ruos","version":null,"description":"MCP stdio server that gives an AI computer-use + CLI/console control of a ruOS desktop (screenshot + keyboard/mouse + shell + system actions + resolution), attached over SSH","keywords":[],"source":"ruos/mcp/Cargo.toml","sourceSha256":"d50f3b879de505e3ee445d03c4a7dd563bf23e5808cf85f60af549038452a3af","variants":[],"readme":"A minimal [Model Context Protocol](https://modelcontextprotocol.io) server (stdio, JSON-RPC 2.0) that gives an attached AI client full computer-use + CLI/console + system control of a ruOS desktop: - GUI — screenshot, mouse move/click/drag, type, key, scroll, wait (via xdotool / scrot). - Console — run_shell: run an arbitrary command on the desktop and get its output. - System — system_action: named ruOS actions (install / optimize / update / status / restart), proxied to the desktop executor. - Display — desktop_resolution: change the desktop resolution after launch, server-side. It is a single hand-wired Rust binary — no async runtime, no MCP SDK — that compiles to a static musl binary and","readmeSource":"ruos/mcp/README.md","product":"key:ruos:ruos-mcp","canonical":"ruos-mcp","tier":null,"t":["ruos-mcp","ruos","mcp","stdio","server","give","computer-use","computer","console","control","desktop","screenshot","keyboard","mouse","shell","action","resolution","attach","ssh"],"s":["ruos-mcp","ruos","mcp"]},
+{"id":"ruos-thermal","family":"ruos-thermal","kind":"crate","store":"ruvector","version":null,"description":"Pi 5 thermal supervisor + over/underclock control (ADR-174)","keywords":["edge-ai","overclock","pi-5","ruos","thermal"],"source":"ruvector/crates/ruos-thermal/Cargo.toml","sourceSha256":"63534d788f2d0e577c3512696065c584d0c81cc423b6c647d6a7bb97eaddde9b","variants":[],"product":"key:ruvector:ruos-thermal","canonical":"ruos-thermal","tier":"T0","t":["ruos-thermal","ruos","thermal","supervisor","underclock","control","adr-174","adr","174","edge-ai","edge","overclock","pi-5"],"s":["ruos-thermal","ruos","thermal","edge-ai","edge","overclock","pi-5"]},
+{"id":"rupixel","family":"rupixel","kind":"npm","store":"rupixel","version":"0.1.0","description":"Pixel-native visual RAG, ported to Rust on the ruvector ANN substrate — screenshot/document retrieval over visual embeddings (a Rust port of PixelRAG) with a metaharness benchmark CLI.","keywords":["ann","benchmark","colpali","darwin","document-retrieval","embeddings","hnsw","ivf","metaharness","multimodal","ocr-free","pixel-rag","pixelrag","rag","retrieval-augmented-generation","rust","ruvector","screenshot","vector-search","visual-rag"],"source":"rupixel/package.json","sourceSha256":"e1257440cc89c712188b0ebb545afe1cd4e8786ee51bce9b681ff16b64f44257","variants":[],"readme":"rupixel finds the documents that match what you *mean*, not just the words you typed. You can search a pile of documents two ways: 1. By their text — the classic approach. 2. By how the page *looks* — rupixel takes a screenshot of each page and searches the pictures. That matters when the text is locked inside a scan, an image, a table, or a chart that ordinary text search can't read. Try it right now — all run entirely in your browser, no install, no server: - ▶ Real-time video search: https://ruvnet.github.io/rupixel/live.html — point a camera/screen at it, type what you're looking for, jump to the moment. - ▶ Visual search: https://ruvnet.github.io/rupixel/visual.html - ▶ Text search: htt","readmeSource":"rupixel/README.md","product":"repo:rupixel","canonical":"rupixel","tier":"T2","t":["rupixel","pixel-native","pixel","visual","rag","port","ann","substrate","screenshot","document","retrieval","embedding","pixelrag","metaharness","benchmark","colpali","darwin","document-retrieval","hnsw","ivf","multimodal","ocr-free","ocr","free","pixel-rag","retrieval-augmented-generation","augment","generation","vector-search","vector","search","visual-rag"],"s":["rupixel","ann","benchmark","colpali","darwin","document-retrieval","document","retrieval","embedding","hnsw","ivf","metaharness","multimodal","ocr-free","ocr","free","pixel-rag","pixel","rag","pixelrag","retrieval-augmented-generation","augment","generation","screenshot","vector-search","vector","search","visual-rag","visual"]},
+{"id":"ruqu-algorithms","family":"ruqu-algorithms","kind":"crate","store":"ruqu","version":null,"description":"Production-ready quantum algorithms in Rust - VQE for chemistry, Grover's search, QAOA optimization, Surface Code error correction","keywords":["grover","qaoa","quantum","quantum-algorithms","vqe"],"source":"ruqu/crates/ruqu-algorithms/Cargo.toml","sourceSha256":"1edb486a95de2935df30d279c7204e00a939486f03b189ab53e9d89bb86dc0e4","variants":[],"readme":"Production-ready quantum algorithms in Rust — VQE for chemistry, Grover's search, QAOA optimization, and Surface Code error correction. Find ground state energies for molecular Hamiltonians: Quadratic speedup for unstructured search: Optimal iterations: π/4 × √N for N items. Approximate solutions to NP-hard graph problems: Topological quantum error correction for fault-tolerant computing: - [ruqu-core](https://crates.io/crates/ruqu-core) — Quantum circuit simulator - [ruqu-exotic](https://crates.io/crates/ruqu-exotic) — Experimental quantum-classical hybrids - [ruqu-wasm](https://crates.io/crates/ruqu-wasm) — Run in browsers via WebAssembly - [VQE Algorithm (ADR-QE-005)](https://github.com/r","readmeSource":"ruqu/crates/ruqu-algorithms/README.md","product":"key:ruqu:ruqu-algorithms","canonical":"ruqu-algorithms","tier":"T2","t":["ruqu-algorithms","ruqu","algorithm","production-ready","quantum","vqe","chemistry","grover","search","qaoa","surface","code","error","correction","quantum-algorithms"],"s":["ruqu-algorithms","ruqu","algorithm","grover","qaoa","quantum","quantum-algorithms","vqe"]},
+{"id":"ruqu-exotic","family":"ruqu-exotic","kind":"crate","store":"ruqu","version":null,"description":"Experimental quantum-classical hybrid algorithms - quantum memory decay, interference search, reasoning error correction, swarm interference for AI systems","keywords":["ai","error-correction","hybrid-algorithms","interference","quantum"],"source":"ruqu/crates/ruqu-exotic/Cargo.toml","sourceSha256":"a5d4c610d9a3828de22a9d3f02b85f57086915454fe7668f98c71e6765393e12","variants":[],"readme":"Experimental quantum-classical hybrid algorithms — quantum memory decay, interference-based search, reasoning error correction, swarm interference, syndrome diagnosis, and reversible memory for AI systems. Model temporal coherence loss in memory systems: Quantum-inspired amplitude interference for similarity search: Detect and correct errors in AI reasoning chains: Coordinate multi-agent systems with quantum interference: Detect error patterns in distributed systems: Quantum-reversible operations for undo/redo:","readmeSource":"ruqu/crates/ruqu-exotic/README.md","product":"key:ruqu:ruqu-exotic","canonical":"ruqu-exotic","tier":"T2","t":["ruqu-exotic","ruqu","exotic","experimental","quantum-classical","quantum","classical","hybrid","algorithm","memory","decay","interference","search","reason","error","correction","swarm","error-correction","hybrid-algorithms"],"s":["ruqu-exotic","ruqu","exotic","error-correction","error","correction","hybrid-algorithms","hybrid","algorithm","interference","quantum"]},
+{"id":"ruqu-observer-world","family":"ruqu-observer-world","kind":"crate","store":"ruqu","version":null,"description":"Verified synthetic observer records projected into real RuField and WorldGraph types","keywords":[],"source":"ruqu/integrations/observer-world/Cargo.toml","sourceSha256":"4037725f7d74001aee9f592c88c8309e1f82df93623431ef677b1289e234d86b","variants":[],"readme":"Offline Rust adapter from the ruQu WASM observer laboratory into actual, revision-pinned RuField and WorldGraph types. This is classical quantum simulation plus a deliberately altered memory experiment. It is not quantum hardware, a physical sensor, a human experiment, a quantum advantage result, or evidence of alternate realities. From the ruQu repository root: Retain run().ledger.root through a separately trusted experiment channel when generating evidence. Reading the root out of the same potentially altered file immediately before verification proves only self-consistency. A hash root is neither a signature nor proof that the stated WASM program actually executed. Input is capped at 32 M","readmeSource":"ruqu/integrations/observer-world/README.md","product":"key:ruqu:ruqu-observer-world","canonical":"ruqu-observer-world","tier":"T2","t":["ruqu-observer-world","ruqu","observer","world","verify","synthetic","record","project","real","rufield","worldgraph","type"],"s":["ruqu-observer-world","ruqu","observer","world"]},
+{"id":"rustc-hyperopt","family":"rustc-hyperopt","kind":"crate","store":"sublinear-time-solver","version":null,"description":"🧠 AI-powered Rust compiler optimizer: 3x faster cold starts, 10-100x faster incremental builds","keywords":["ai","compiler","optimization","performance","rust"],"source":"sublinear-time-solver/crates/rustc-hyperopt/Cargo.toml","sourceSha256":"a54c3c0acb7c82b63c8f8e66bd8aa28c1cebfed40fb8b02eb9e60bc22535cb42","variants":[],"readme":"🧠 The World's First AI-Powered Semantic Rust Compiler Optimizer Unlike traditional tools that cache compiled artifacts, RustC HyperOpt uses AI-powered semantic analysis to understand code patterns, predict compilation needs, and optimize at the language level. We're the only tool that combines global semantic caching with intelligent cold-start optimization. 🎯 Our Unique Advantage: We're the only tool that solves the cold-start problem with AI-powered pattern recognition. 📈 Performance Summary: - Cold starts: We're 3x faster than any competitor - Incremental: We're 10-80x faster than traditional approaches - Clean builds: We're competitive with the best caching solutions - Development wor","readmeSource":"sublinear-time-solver/crates/rustc-hyperopt/README.md","product":"key:sublinear-time-solver:rustc-hyperopt","canonical":"rustc-hyperopt","tier":"T1","t":["rustc-hyperopt","rustc","hyperopt","ai-powered","power","compiler","optimizer","cold","start","10-100x","100x","incremental"],"s":["rustc-hyperopt","rustc","hyperopt","compiler"]},
+{"id":"ruv-fann","family":"ruv-fann","kind":"crate","store":"ruv-fann","version":null,"description":"A pure Rust implementation of the Fast Artificial Neural Network (FANN) library","keywords":["ai","deep-learning","fann","machine-learning","neural-network"],"source":"ruv-fann/Cargo.toml","sourceSha256":"231de4bde472a84a1efb3437ed4b60a780b2e6d95b0ef1730b904b7ca5004fdc","variants":[],"readme":"What if intelligence could be ephemeral, composable, and surgically precise? Welcome to ruv-FANN, a comprehensive neural intelligence framework that reimagines how we build, deploy, and orchestrate artificial intelligence. This repository contains three groundbreaking projects that work together to deliver unprecedented performance in neural computing, forecasting, and multi-agent orchestration. We believe AI should be: - Ephemeral: Spin up intelligence when needed, dissolve when done - Accessible: CPU-native, GPU-optional - built for the GPU-poor - Composable: Mix and match neural architectures like LEGO blocks - Precise: Tiny, purpose-built brains for specific tasks This isn't about callin","readmeSource":"ruv-fann/README.md","product":"repo:ruv-fann","canonical":"ruv-fann","tier":"T1","t":["ruv-fann","fann","pure","implementation","artificial","neural","network","deep-learning","deep","learn","machine-learning","machine","neural-network"],"s":["ruv-fann","fann","deep-learning","deep","learn","machine-learning","machine","neural-network","neural","network"]},
+{"id":"ruv-fann-wasm","family":"ruv-fann","kind":"crate","store":"synaptic-mesh","version":null,"description":"WASM-optimized Fast Artificial Neural Network library with SIMD acceleration","keywords":["fann","machine-learning","neural-network","simd","wasm"],"source":"synaptic-mesh/src/rs/ruv-fann-wasm/Cargo.toml","sourceSha256":"4e3332b905b50a993cae33e4b7a2236e614d8ae50879f6df1e881f4d1dd4f89b","variants":[],"product":"repo:ruv-fann","canonical":"ruv-fann","tier":"T1","t":["ruv-fann-wasm","fann","ruv-fann","wasm-optimized","artificial","neural","network","acceleration","machine-learning","machine","learn","neural-network"],"s":["ruv-fann-wasm","fann","ruv-fann","machine-learning","machine","learn","neural-network","neural","network"]},
+{"id":"ruv-jellyfish","family":"ruv-jellyfish","kind":"crate","store":"ruv-drone","version":null,"description":"Jellyfish-inspired energy-efficient swarm behaviors for cooperative UAV fleets — pulse-and-drift gait and bloom aggregation (ADR-172)","keywords":[],"source":"ruv-drone/crates/ruv-jellyfish/Cargo.toml","sourceSha256":"70e00e445c68e33374e77f421d578eac7a5ff690338c0c4299fff911697d2e82","variants":[],"readme":"Jellyfish-inspired, energy-efficient swarm behaviors for cooperative UAV fleets. A standalone member of the [ruv-drone](../../) workspace (ADR-172). Where the parent crate's formation, planning, and allocation modules optimize *shapes* and *paths*, ruv-jellyfish optimizes endurance and presence for the missions that are battery-bound rather than speed-bound: relay chains that must hold station for hours, SAR re-scan loops that loiter over high-probability cells, and persistent agricultural monitoring. It draws two mechanisms from real jellyfish biomechanics and ecology: [JellyfishController](src/controller.rs) unifies both around a single per-kilogram energy budget so a planner can reason ab","readmeSource":"ruv-drone/crates/ruv-jellyfish/README.md","product":"key:ruv-drone:ruv-jellyfish","canonical":"ruv-jellyfish","tier":"T2","t":["ruv-jellyfish","jellyfish","jellyfish-inspired","inspir","energy-efficient","energy","efficient","swarm","behavior","cooperative","uav","fleet","pulse-and-drift","pulse","drift","gait","bloom","aggregation","adr-172","adr","172"],"s":["ruv-jellyfish","jellyfish"]},
+{"id":"ruv-neural-biosense","family":"ruv-neural-biosense","kind":"crate","store":"ruv-neural","version":null,"description":"Physiological response sensing (HRV, respiration, motion, sleep proxy) for the rUv Neural closed-loop neuromodulation platform — the 'measure a response' stage.","keywords":["biosignal","hrv","neurofeedback","physiology","respiration"],"source":"ruv-neural/ruv-neural-biosense/Cargo.toml","sourceSha256":"8d7a98610aad913aed9b3f0b4e54e08e4c6be90e7a373370cb5b10c5125c8949","variants":[],"product":"key:ruv-neural:ruv-neural-biosense","canonical":"ruv-neural-biosense","tier":"T1","t":["ruv-neural-biosense","neural","biosense","physiological","response","sens","hrv","respiration","motion","sleep","proxy","closed-loop","clos","loop","neuromodulation","measure","stage","biosignal","neurofeedback","physiology"],"s":["ruv-neural-biosense","neural","biosense","biosignal","hrv","neurofeedback","physiology","respiration"]},
+{"id":"ruv-neural-brain2text","family":"ruv-neural-brain2text","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — non-invasive brain-to-text decoding bridge (Brain2Qwerty / SpanishBCBL), with an in-crate evolutionary optimizer","keywords":["bci","brain","eeg","meg","neural"],"source":"ruv-neural/ruv-neural-brain2text/Cargo.toml","sourceSha256":"76cf9f00d12ebafefb172c1f810605d04640c816243149c3f4e6ea6019660960","variants":[],"readme":"Non-invasive brain-to-text decoding bridge for the rUv Neural workspace, inspired by Meta AI's [Brain2Qwerty](https://ai.meta.com/blog/brain2qwerty-brain-ai-human-communication/) and the [SpanishBCBL (DECOMEG)](https://huggingface.co/datasets/bcbl190626/SpanishBCBL) MEG/EEG dataset — with an in-crate evolutionary optimizer (\"Darwin mode\": *freeze the model, evolve the harness*). evaluate() ties it together; the composable Harness builder fits a self-contained, serializable TrainedPipeline (config + trained weights + LM) you can to_json() / from_json() — that artifact *is* the distributable model. evolve::evolve() (\"Darwin mode\") searches the whole Brain2TextConfig space — including the model","readmeSource":"ruv-neural/ruv-neural-brain2text/README.md","product":"key:ruv-neural:ruv-neural-brain2text","canonical":"ruv-neural-brain2text","tier":"T1","t":["ruv-neural-brain2text","neural","brain2text","non-invasive","non","invasive","brain-to-text","brain","text","decod","bridge","brain2qwerty","spanishbcbl","in-crate","crate","evolutionary","optimizer","bci","eeg","meg"],"s":["ruv-neural-brain2text","neural","brain2text","bci","brain","eeg","meg"]},
+{"id":"ruv-neural-cli","family":"ruv-neural","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — CLI tool for brain topology analysis, simulation, and visualization","keywords":["brain","core","neural","topology","types"],"source":"ruv-neural/ruv-neural-cli/Cargo.toml","sourceSha256":"6623a42c1dfb04b043d349df90860a85d394aac67288a0f640fdcfb5f95e6a5d","variants":["ruv-neural-core","ruv-neural-wasm"],"readme":"CLI tool for brain topology analysis, simulation, and visualization. ruv-neural-cli is the command-line binary (ruv-neural) that ties together the entire rUv Neural crate ecosystem. It provides subcommands for simulating neural sensor data, analyzing brain connectivity graphs, computing minimum cuts, running the full processing pipeline with an optional ASCII dashboard, and exporting to multiple visualization formats. Runs: simulate -> preprocess -> build graph -> mincut -> embed -> decode. Displays crate versions, available features, and system capabilities. Depends on all workspace crates: ruv-neural-core, ruv-neural-sensor, ruv-neural-signal, ruv-neural-graph, ruv-neural-mincut, ruv-neura","readmeSource":"ruv-neural/ruv-neural-cli/README.md","product":"repo:ruv-neural","canonical":"ruv-neural-cli","tier":"T1","t":["ruv-neural-cli","neural","ruv-neural","brain","topology","analysis","simulation","visualization","core","type"],"s":["ruv-neural-cli","neural","ruv-neural","brain","core","topology","type"]},
+{"id":"ruv-neural-decoder","family":"ruv-neural-decoder","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — Cognitive state classification and BCI decoding from neural topology embeddings","keywords":[],"source":"ruv-neural/ruv-neural-decoder/Cargo.toml","sourceSha256":"c9359b8f8789258febf7668cf3c03e72a51ca975fc0f05d68574116bedaf268e","variants":[],"readme":"Cognitive state classification and BCI decoding from neural topology embeddings. ruv-neural-decoder classifies cognitive states from brain graph embeddings and topology metrics. It provides multiple decoding strategies -- KNN classification from labeled exemplars, threshold-based rule systems, temporal transition detection, and clinical biomarker scoring -- plus an ensemble pipeline that combines all strategies for robust real-time brain-computer interface (BCI) output. - KNN decoder (knn_decoder): K-nearest neighbor classification using stored labeled embeddings from ruv-neural-memory; supports configurable k and distance metrics - Threshold decoder (threshold_decoder): Rule-based classific","readmeSource":"ruv-neural/ruv-neural-decoder/README.md","product":"key:ruv-neural:ruv-neural-decoder","canonical":"ruv-neural-decoder","tier":"T1","t":["ruv-neural-decoder","neural","decoder","cognitive","state","classification","bci","decod","topology","embedding"],"s":["ruv-neural-decoder","neural","decoder"]},
+{"id":"ruv-neural-embed","family":"ruv-neural-embed","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — Graph embedding generation for brain connectivity states using RuVector format","keywords":[],"source":"ruv-neural/ruv-neural-embed/Cargo.toml","sourceSha256":"2ef6080b0fcded20e94f6d1bd352d3ae63455cedfe898200ca871a168ece2efd","variants":[],"readme":"Graph embedding generation for brain connectivity states using RuVector format. ruv-neural-embed converts brain connectivity graphs into fixed-dimensional vector representations suitable for downstream classification, clustering, and temporal analysis. It provides multiple embedding methods and supports export to the RuVector .rvf binary format for interoperability with the broader RuVector ecosystem. - Spectral embedding (spectral_embed): Laplacian eigenvector-based positional encoding from the graph's normalized Laplacian - Topology embedding (topology_embed): Hand-crafted topological feature vectors derived from graph-theoretic metrics - Node2Vec (node2vec): Random-walk co-occurrence embe","readmeSource":"ruv-neural/ruv-neural-embed/README.md","product":"key:ruv-neural:ruv-neural-embed","canonical":"ruv-neural-embed","tier":"T1","t":["ruv-neural-embed","neural","embed","graph","embedd","generation","brain","connectivity","state","format"],"s":["ruv-neural-embed","neural","embed"]},
+{"id":"ruv-neural-esp32","family":"ruv-neural-esp32","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — ESP32 edge integration for neural sensor data acquisition and preprocessing","keywords":[],"source":"ruv-neural/ruv-neural-esp32/Cargo.toml","sourceSha256":"03bc9a0850ef6cdf18ef3788b548add1d9307a0d98ab7e0428a11ec7cb7b263f","variants":[],"readme":"ESP32 edge integration for neural sensor data acquisition and preprocessing. ruv-neural-esp32 provides lightweight processing modules designed to run on ESP32 microcontrollers for real-time neural sensor data acquisition and preprocessing at the edge. It handles ADC sampling, time-division multiplexing for multi-sensor coordination, IIR filtering and downsampling on-device, power management for battery operation, a binary communication protocol for streaming data to the rUv Neural backend, and multi-node data aggregation. - ADC interface (adc): AdcReader with configurable AdcConfig including sample rate, resolution, attenuation levels, and multi-channel support via AdcChannel - TDM schedulin","readmeSource":"ruv-neural/ruv-neural-esp32/README.md","product":"key:ruv-neural:ruv-neural-esp32","canonical":"ruv-neural-esp32","tier":"T1","t":["ruv-neural-esp32","neural","esp32","edge","integration","sensor","acquisition","preprocess"],"s":["ruv-neural-esp32","neural","esp32"]},
+{"id":"ruv-neural-graph","family":"ruv-neural-graph","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — Brain connectivity graph construction from neural signals","keywords":[],"source":"ruv-neural/ruv-neural-graph/Cargo.toml","sourceSha256":"3a4cf511722fb32b204c795c60d386d466299a4a0d6e622c8e61b8ec3c338fc1","variants":[],"readme":"Brain connectivity graph construction from neural signals with graph-theoretic analysis and spectral properties. ruv-neural-graph builds brain connectivity graphs from multi-channel neural time series data and connectivity matrices. It provides graph-theoretic metrics (efficiency, clustering, centrality), spectral graph properties (Laplacian, Fiedler value), brain atlas definitions, petgraph interoperability, and temporal dynamics tracking for brain topology research. - Graph construction (constructor): Build BrainGraph instances from connectivity matrices and multi-channel time series data via BrainGraphConstructor - Brain atlases (atlas): Built-in Desikan-Killiany 68-region atlas with supp","readmeSource":"ruv-neural/ruv-neural-graph/README.md","product":"key:ruv-neural:ruv-neural-graph","canonical":"ruv-neural-graph","tier":"T1","t":["ruv-neural-graph","neural","graph","brain","connectivity","construction","signal"],"s":["ruv-neural-graph","neural","graph"]},
+{"id":"ruv-neural-io","family":"ruv-neural-io","kind":"crate","store":"ruv-neural","version":null,"description":"Bounded native EEG recording adapters for rUv Neural","keywords":["edf","eeg","neural","parser","streaming"],"source":"ruv-neural/ruv-neural-io/Cargo.toml","sourceSha256":"da1411fd2e126e410dd0d0d741df95b5b5fb5cabb787afdf2e8096a36b6b34ed","variants":[],"readme":"Bounded native recording adapters for NeuroSleep research ingestion. EdfEpochSource streams one complete EDF/EDF+ data record at a time. It checks source bytes, channel count, sample rates, duration, metadata, temporary allocation, arithmetic overflow, numeric metadata, sample finiteness, truncation, annotation count, chronology, and trailing bytes before returning atomic epochs. It performs no writes. Allocation admission conservatively accounts for the raw record, decoded f64 samples, epoch/channel/annotation structs, and cloned EDF labels; the fuzz target provides an additional bounded malformed-input surface. The brainvision-compat default feature reexports the existing BrainVision API w","readmeSource":"ruv-neural/ruv-neural-io/README.md","product":"key:ruv-neural:ruv-neural-io","canonical":"ruv-neural-io","tier":"T1","t":["ruv-neural-io","neural","bound","eeg","record","adapter","edf","parser","stream"],"s":["ruv-neural-io","neural","edf","eeg","parser","stream"]},
+{"id":"ruv-neural-loop","family":"ruv-neural-loop","kind":"crate","store":"ruv-neural","version":null,"description":"Ruflo — closed-loop sensory neuromodulation controller for rUv Neural: personal state embedding, safety envelope, conservative dosing, and a tamper-evident audit trail. Detect a target state, deliver a verified stimulus, measure the response, stop safely outside the envelope.","keywords":["audit","closed-loop","controller","neuromodulation","safety"],"source":"ruv-neural/ruv-neural-loop/Cargo.toml","sourceSha256":"a1646c75c3f7a65fd8971847d53ba18984ef6dbd4317ae21733fd335d5499d90","variants":[],"product":"key:ruv-neural:ruv-neural-loop","canonical":"ruv-neural-loop","tier":"T1","t":["ruv-neural-loop","neural","loop","ruflo","closed-loop","clos","sensory","neuromodulation","controller","personal","state","embedd","safety","envelope","conservative","dosing","tamper-evident","tamper","evident","audit","trail","detect","target","deliver","verify","stimulus","measure","response","stop","safely","outside"],"s":["ruv-neural-loop","neural","loop","audit","closed-loop","clos","controller","neuromodulation","safety"]},
+{"id":"ruv-neural-memory","family":"ruv-neural-memory","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — Persistent neural state memory with vector search and longitudinal tracking","keywords":[],"source":"ruv-neural/ruv-neural-memory/Cargo.toml","sourceSha256":"729039ceab808ac1ff7fdc8c12311a01a558f403d614a5317fc2de05f3eb765f","variants":[],"readme":"Persistent neural state memory with vector search and longitudinal tracking. ruv-neural-memory provides in-memory and persistent storage for neural embeddings, supporting brute-force and HNSW-based approximate nearest neighbor search. It includes session-based memory management for organizing recordings by subject and session, longitudinal drift detection for tracking embedding distribution changes over time, and RVF/bincode persistence for durable storage. - Embedding store (store): NeuralMemoryStore for inserting, querying, and managing collections of NeuralEmbedding values with brute-force nearest neighbor search - HNSW index (hnsw): HnswIndex for approximate nearest neighbor search with","readmeSource":"ruv-neural/ruv-neural-memory/README.md","product":"key:ruv-neural:ruv-neural-memory","canonical":"ruv-neural-memory","tier":"T1","t":["ruv-neural-memory","neural","memory","persistent","state","vector","search","longitudinal","track"],"s":["ruv-neural-memory","neural","memory"]},
+{"id":"ruv-neural-mincut","family":"ruv-neural-mincut","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — Dynamic minimum cut analysis for brain network topology detection","keywords":[],"source":"ruv-neural/ruv-neural-mincut/Cargo.toml","sourceSha256":"20a47df5c14fba993809af382aaa68cafafa9be0403a93b5f2c44c21e87dbc7b","variants":[],"readme":"Dynamic minimum cut analysis for brain network topology detection. ruv-neural-mincut provides algorithms for computing minimum cuts on brain connectivity graphs, tracking topology changes over time, and detecting neural coherence events such as network formation, dissolution, merger, and split. These algorithms form the core of the rUv Neural cognitive state detection pipeline, identifying when brain network topology undergoes significant structural transitions. - Stoer-Wagner (stoer_wagner): Global minimum cut in O(V^3) time, returning cut value, partitions, and cut edges - Normalized cut (normalized): Shi-Malik spectral bisection via the Fiedler vector for balanced graph partitioning - Mul","readmeSource":"ruv-neural/ruv-neural-mincut/README.md","product":"key:ruv-neural:ruv-neural-mincut","canonical":"ruv-neural-mincut","tier":"T1","t":["ruv-neural-mincut","neural","mincut","dynamic","minimum","cut","analysis","brain","network","topology","detection"],"s":["ruv-neural-mincut","neural","mincut"]},
+{"id":"ruv-neural-neurosleep","family":"ruv-neural-neurosleep","kind":"crate","store":"ruv-neural","version":null,"description":"Stage-aware offline NeuroSleep qEEG research profile","keywords":[],"source":"ruv-neural/ruv-neural-neurosleep/Cargo.toml","sourceSha256":"12f9ab70ee6fff205023f811bd567cb70c9b9d7febcf28bd5add6326f127b50a","variants":[],"product":"key:ruv-neural:ruv-neural-neurosleep","canonical":"ruv-neural-neurosleep","tier":"T1","t":["ruv-neural-neurosleep","neural","neurosleep","stage-aware","stage","aware","offline","qeeg","research","profile"],"s":["ruv-neural-neurosleep","neural","neurosleep"]},
+{"id":"ruv-neural-sensor","family":"ruv-neural-sensor","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — Sensor data acquisition for NV diamond, OPM, EEG, and simulated sources","keywords":[],"source":"ruv-neural/ruv-neural-sensor/Cargo.toml","sourceSha256":"3fa6e8775542c042741fce7309e5d89770009f4d64794870f962160da6c779bd","variants":[],"readme":"Sensor data acquisition for NV diamond, OPM, EEG, and simulated sources. ruv-neural-sensor provides uniform sensor interfaces for multiple neural magnetometry and electrophysiology sensor types. Each sensor backend implements the SensorSource trait from ruv-neural-core, producing MultiChannelTimeSeries data. The crate also includes calibration utilities and real-time signal quality monitoring. - Simulated sensor (simulator feature, default): Synthetic multi-channel data generation with configurable alpha rhythm injection, noise floor control, and event injection (spikes, artifacts) - NV diamond (nv_diamond feature): Nitrogen-vacancy diamond magnetometer interface with configurable sensitivit","readmeSource":"ruv-neural/ruv-neural-sensor/README.md","product":"key:ruv-neural:ruv-neural-sensor","canonical":"ruv-neural-sensor","tier":"T1","t":["ruv-neural-sensor","neural","sensor","acquisition","diamond","opm","eeg","simulat","source"],"s":["ruv-neural-sensor","neural","sensor"]},
+{"id":"ruv-neural-signal","family":"ruv-neural-signal","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — Signal processing: filtering, spectral analysis, artifact rejection for neural data","keywords":[],"source":"ruv-neural/ruv-neural-signal/Cargo.toml","sourceSha256":"f6d503ea1b98e661143b3c35a16c1728c2aa2c82bae7639e46058c1dfef140e6","variants":[],"readme":"Signal processing: filtering, spectral analysis, connectivity metrics, and artifact rejection for neural time series data. ruv-neural-signal provides a complete digital signal processing pipeline for multi-channel neural magnetic field and electrophysiology data. It covers IIR filtering in second-order sections form, FFT-based spectral analysis, Hilbert transform for instantaneous phase extraction, artifact detection and rejection, cross-channel connectivity metrics, and a configurable multi-stage preprocessing pipeline. - IIR Filters (filter): Butterworth bandpass, highpass, lowpass, and notch filters in SOS (second-order sections) form for numerical stability - Spectral analysis (spectral)","readmeSource":"ruv-neural/ruv-neural-signal/README.md","product":"key:ruv-neural:ruv-neural-signal","canonical":"ruv-neural-signal","tier":"T1","t":["ruv-neural-signal","neural","signal","process","filter","spectral","analysis","artifact","rejection"],"s":["ruv-neural-signal","neural","signal"]},
+{"id":"ruv-neural-stim","family":"ruv-neural-stim","kind":"crate","store":"ruv-neural","version":null,"description":"Multi-modal sensory neuromodulation stimulus synthesis (40 Hz light/audio/haptic) with sensory-safety limits and cryptographic delivery receipts for the rUv Neural closed-loop platform.","keywords":["entrainment","gamma","neurofeedback","neuromodulation","stimulus"],"source":"ruv-neural/ruv-neural-stim/Cargo.toml","sourceSha256":"b457104256aaf50637fe6377e715be3203809c11d4ed496337f22175cb2eec8a","variants":[],"product":"key:ruv-neural:ruv-neural-stim","canonical":"ruv-neural-stim","tier":"T1","t":["ruv-neural-stim","neural","stim","multi-modal","multi","modal","sensory","neuromodulation","stimulus","synthesis","light","audio","haptic","sensory-safety","safety","limit","cryptographic","delivery","receipt","closed-loop","clos","loop","entrainment","gamma","neurofeedback"],"s":["ruv-neural-stim","neural","stim","entrainment","gamma","neurofeedback","neuromodulation","stimulus"]},
+{"id":"ruv-neural-ui","family":"ruv-neural-ui","kind":"npm","store":"ruv-neural","version":"0.1.0","description":"rUv Neural UI — static, local-first web console for the Ruflo closed-loop sensory neuromodulation platform (Demo + Replay).","keywords":[],"source":"ruv-neural/apps/ruv-neural-ui/package.json","sourceSha256":"8c9b1bdb467a39dc958c3b54953bb8887c56a8a1622efc766c7bcc6a16778dc1","variants":[],"readme":"A static, local-first web console for the Ruflo closed-loop sensory neuromodulation platform (see [ADR-0014](../../docs/adr/0014-web-console.md)). It demonstrates the whole loop — *identify a target state → deliver a verified 40 Hz stimulus → measure the response → stop safely outside the envelope* — and verifies the evidence entirely in your browser. No backend, no accounts, no health data leaves the page. - Demo — deterministic, browser-only. Plays back real evidence bundles generated by ruv-neural neuromod --bundle … --sign (relaxed converge, relaxed fail-safe stop, focused, multimodal gamma). - Replay — import any Ruflo evidence bundle (JSON) and verify it locally: schema validity, the r","readmeSource":"ruv-neural/apps/ruv-neural-ui/README.md","product":"key:ruv-neural:ruv-neural-ui","canonical":"ruv-neural-ui","tier":"T1","t":["ruv-neural-ui","neural","static","local-first","local","web","console","ruflo","closed-loop","clos","loop","sensory","neuromodulation","demo","replay"],"s":["ruv-neural-ui","neural"]},
+{"id":"ruv-neural-viz","family":"ruv-neural-viz","kind":"crate","store":"ruv-neural","version":null,"description":"rUv Neural — Brain topology visualization data structures and ASCII rendering","keywords":[],"source":"ruv-neural/ruv-neural-viz/Cargo.toml","sourceSha256":"95fd32642f99260b9e87eb7c4a7cee133386bd3ac21370860a43e760c98c9f98","variants":[],"readme":"Brain topology visualization, ASCII rendering, and export formats. ruv-neural-viz provides layout algorithms, color mapping, terminal-friendly ASCII rendering, animation frame generation, and export to standard graph visualization formats for brain connectivity graphs. It turns BrainGraph and mincut analysis results into visual output suitable for terminal dashboards, web applications, and graph analysis tools. - Layout algorithms (layout): ForceDirectedLayout for spring-based node positioning and AnatomicalLayout for MNI-coordinate-based brain region placement; circular layout variants - Color mapping (colormap): ColorMap with cool-warm, viridis, and module-color schemes for mapping scalar","readmeSource":"ruv-neural/ruv-neural-viz/README.md","product":"key:ruv-neural:ruv-neural-viz","canonical":"ruv-neural-viz","tier":"T1","t":["ruv-neural-viz","neural","viz","brain","topology","visualization","structure","ascii","render"],"s":["ruv-neural-viz","neural","viz"]},
+{"id":"ruv-swarm","family":"ruv-swarm","kind":"npm","store":"ruv-fann","version":"1.0.18","description":"High-performance neural network swarm orchestration in WebAssembly","keywords":["ai","cli","distributed","orchestration","swarm"],"source":"ruv-fann/ruv-swarm/npm/package.json","sourceSha256":"2c2f844269a7f76b34b1d119499e2a8d25dbba990a33eb46eb9f38ffecf64952","variants":["ruv-swarm-cli","ruv-swarm-core","ruv-swarm-wasm"],"readme":"What if every task, every file, every function could truly think? Just for a moment. No LLM required. That's what ruv-swarm makes real. ruv-swarm lets you spin up ultra-lightweight custom neural networks that exist just long enough to solve the problem. Tiny purpose-built brains dedicated to solving very specific challenges. Think particular coding structures, custom communications, trading optimization - neural networks built on the fly just for the task they need to exist for, long enough to solve it, then gone. Built for the GPU-poor: These agents are CPU-native and GPU-optional. Rust compiles to high-speed WASM binaries that run anywhere - browser, edge, server - with zero external depen","readmeSource":"ruv-fann/ruv-swarm/npm/README.md","product":"key:ruv-fann:ruv-swarm","canonical":"ruv-swarm","tier":"T1","t":["ruv-swarm","swarm","high","neural","network","orchestration","distribut"],"s":["ruv-swarm","swarm","distribut","orchestration"]},
+{"id":"ruv-swarm-agents","family":"ruv-swarm-agents","kind":"crate","store":"ruv-fann","version":null,"description":"Specialized AI agents for RUV Swarm","keywords":[],"source":"ruv-fann/ruv-swarm/crates/ruv-swarm-agents/Cargo.toml","sourceSha256":"71a2ca9bdd2e256a72369ef97f3405412048a75d9fbf6dca4464dfa7fbe58426","variants":[],"readme":"Specialized AI agent implementations for the RUV Swarm neural orchestration system. This crate provides cognitive diversity through intelligent agent types that leverage different thinking patterns for optimal swarm performance. The ruv-swarm-agents crate implements specialized AI agents with diverse cognitive patterns designed for high-performance swarm orchestration. Each agent type embodies different thinking approaches, enabling the swarm to tackle complex problems from multiple angles simultaneously. Built on WebAssembly with SIMD optimization, these agents provide: - Cognitive Diversity: Six distinct thinking patterns for comprehensive problem-solving - Neural Integration: Built-in neu","readmeSource":"ruv-fann/ruv-swarm/crates/ruv-swarm-agents/README.md","product":"key:ruv-fann:ruv-swarm-agents","canonical":"ruv-swarm-agents","tier":"T1","t":["ruv-swarm-agents","swarm","specializ"],"s":["ruv-swarm-agents","swarm"]},
+{"id":"ruv-swarm-benchmarking","family":"ruv-swarm-benchmarking","kind":"crate","store":"ruv-fann","version":null,"description":"Comprehensive benchmarking framework for Claude Code CLI and swarm orchestration","keywords":[],"source":"ruv-fann/ruv-swarm/benchmarking/Cargo.toml","sourceSha256":"9772985ad95783c69c4db61f27ea2bdb2083d84393fd64f0782fd21c9cab52ab","variants":[],"readme":"A comprehensive benchmarking framework for evaluating Claude Code CLI performance with ML-optimized swarm intelligence against baseline implementations. 1. BenchmarkExecutor: Runs Claude Code CLI commands with stream-json output parsing 2. MetricsCollector: Gathers detailed performance metrics including: - Task completion time - Tool invocation patterns - Thinking sequences and token counts - Error recovery tracking - Resource usage (CPU, memory, network) - Code quality metrics 3. SQLite Storage: Persistent storage with comprehensive schema for: - Benchmark runs and configurations - Stream events from Claude - Tool invocations and thinking sequences - Performance metrics and comparisons - SW","readmeSource":"ruv-fann/ruv-swarm/benchmarking/README.md","product":"key:ruv-fann:ruv-swarm-benchmarking","canonical":"ruv-swarm-benchmarking","tier":"T1","t":["ruv-swarm-benchmarking","swarm","benchmark","comprehensive","claude","code","orchestration"],"s":["ruv-swarm-benchmarking","swarm","benchmark"]},
+{"id":"ruv-swarm-daa","family":"ruv-swarm-daa","kind":"crate","store":"ruv-fann","version":null,"description":"Decentralized Autonomous Agents (DAA) integration for ruv-swarm","keywords":["ai","autonomous-agents","coordination","daa","swarm"],"source":"ruv-fann/ruv-swarm/crates/ruv-swarm-daa/Cargo.toml","sourceSha256":"ca5098131b6c759b667dd58d2f5cd53600b5d56133fda426bc0a1b7cedd9c2b8","variants":[],"product":"key:ruv-fann:ruv-swarm-daa","canonical":"ruv-swarm-daa","tier":"T1","t":["ruv-swarm-daa","swarm","daa","decentraliz","autonomous","integration","ruv-swarm","autonomous-agents","coordination"],"s":["ruv-swarm-daa","swarm","daa","autonomous-agents","autonomous","coordination"]},
+{"id":"ruv-swarm-mcp","family":"ruv-swarm-mcp","kind":"crate","store":"ruv-fann","version":null,"description":"Model Context Protocol (MCP) integration for RUV Swarm","keywords":[],"source":"ruv-fann/ruv-swarm/crates/ruv-swarm-mcp/Cargo.toml","sourceSha256":"a9a11123a9deb93316c72abaebf32f72bfbc3aeebdafe7e575009e59a84b382e","variants":[],"readme":"ruv-swarm-mcp is a powerful Model Context Protocol (MCP) server implementation for the RUV-Swarm orchestration system. It provides Claude Code and other MCP-compatible clients with seamless access to advanced swarm intelligence capabilities through a standardized JSON-RPC interface. The ruv-swarm-mcp crate bridges the gap between Claude Code's AI capabilities and RUV-Swarm's distributed agent orchestration system. By implementing the Model Context Protocol specification, it enables Claude to directly control and coordinate intelligent agent swarms for complex task execution. - 13+ Comprehensive MCP Tools - Complete swarm orchestration capabilities - Claude Code Integration - Seamless integra","readmeSource":"ruv-fann/ruv-swarm/crates/ruv-swarm-mcp/README.md","product":"key:ruv-fann:ruv-swarm-mcp","canonical":"ruv-swarm-mcp","tier":"T1","t":["ruv-swarm-mcp","swarm","mcp","model","context","protocol","integration"],"s":["ruv-swarm-mcp","swarm","mcp"]},
+{"id":"ruv-swarm-ml","family":"ruv-swarm-ml","kind":"crate","store":"ruv-fann","version":null,"description":"Machine learning integration for RUV Swarm","keywords":[],"source":"ruv-fann/ruv-swarm/crates/ruv-swarm-ml/Cargo.toml","sourceSha256":"cec30dbcb7df71e02634ac86366c46c83c3767827253588ec7c5b15e0ba707b9","variants":[],"readme":"Advanced Machine Learning and Neural Forecasting for RUV Swarm ruv-swarm-ml is a high-performance machine learning crate designed for intelligent swarm orchestration and neural forecasting. It provides agent-specific time series prediction, ensemble methods, and swarm-level forecasting coordination with support for 27+ state-of-the-art forecasting models. - Basic Models: MLP, DLinear, NLinear, MLPMultivariate - Recurrent Models: RNN, LSTM, GRU with memory optimization - Advanced Models: NBEATS, NBEATSx, NHITS, TiDE with interpretability - Transformer Models: TFT, Informer, AutoFormer, FedFormer, PatchTST, ITransformer - Specialized Models: DeepAR, DeepNPTS, TCN, BiTCN, TimesNet, StemGNN, TSM","readmeSource":"ruv-fann/ruv-swarm/crates/ruv-swarm-ml/README.md","product":"key:ruv-fann:ruv-swarm-ml","canonical":"ruv-swarm-ml","tier":"T1","t":["ruv-swarm-ml","swarm","machine","learn","integration"],"s":["ruv-swarm-ml","swarm"]},
+{"id":"ruv-swarm-ml-training","family":"ruv-swarm-ml-training","kind":"crate","store":"ruv-fann","version":null,"description":"Advanced ML training pipeline for neuro-divergent models in RUV Swarm","keywords":[],"source":"ruv-fann/ruv-swarm/ml-training/Cargo.toml","sourceSha256":"2064454e781f95826cf69dd444a47219ffb21601bce80b09553b9ea021b5bc02","variants":[],"readme":"Advanced machine learning training pipeline for neuro-divergent models in the RUV Swarm ecosystem. This crate provides comprehensive tools for training LSTM, TCN, and N-BEATS models to predict agent performance and optimize prompts. - Stream Data Loading: Efficient loading and preprocessing of time-series event data from JSON streams - Multiple Model Architectures: - LSTM (Long Short-Term Memory) for sequence modeling - TCN (Temporal Convolutional Networks) for efficient temporal processing - N-BEATS (Neural Basis Expansion Analysis) for time series forecasting - Hyperparameter Optimization: Advanced optimization methods including Random Search, Bayesian Optimization, Grid Search, and Hyperb","readmeSource":"ruv-fann/ruv-swarm/ml-training/README.md","product":"key:ruv-fann:ruv-swarm-ml-training","canonical":"ruv-swarm-ml-training","tier":"T1","t":["ruv-swarm-ml-training","swarm","train","advanc","pipeline","neuro-divergent","neuro","divergent","model"],"s":["ruv-swarm-ml-training","swarm","train"]},
+{"id":"ruv-swarm-persistence","family":"ruv-swarm-persistence","kind":"crate","store":"ruv-fann","version":null,"description":"Persistence layer with SQLite and ORM for RUV-Swarm","keywords":[],"source":"ruv-fann/ruv-swarm/crates/ruv-swarm-persistence/Cargo.toml","sourceSha256":"4976016cd9446d59e829bde1e549e8ae7387c0d748da54a8814691b771756312","variants":[],"readme":"High-performance, ACID-compliant persistence layer for RUV-Swarm with SQLite and cross-platform support. ruv-swarm-persistence is the foundational data persistence layer for the RUV-Swarm distributed agent system. It provides robust, type-safe data storage with full ACID compliance, real-time persistence, and seamless cross-platform compatibility. Built with modern Rust principles, this crate delivers memory-safe database operations, connection pooling, transaction support, and a flexible repository pattern for managing swarm agent state, task coordination, event sourcing, and performance metrics. - Full ACID transaction support with automatic rollback - WAL (Write-Ahead Logging) mode for op","readmeSource":"ruv-fann/ruv-swarm/crates/ruv-swarm-persistence/README.md","product":"key:ruv-fann:ruv-swarm-persistence","canonical":"ruv-swarm-persistence","tier":"T1","t":["ruv-swarm-persistence","swarm","persistence","layer","sqlite","orm","ruv-swarm"],"s":["ruv-swarm-persistence","swarm","persistence"]},
+{"id":"ruv-swarm-transport","family":"ruv-swarm-transport","kind":"crate","store":"ruv-fann","version":null,"description":"Transport layer for RUV-FANN swarm communication with WebSocket and SharedMemory support","keywords":["distributed","shared-memory","swarm","transport","websocket"],"source":"ruv-fann/ruv-swarm/crates/ruv-swarm-transport/Cargo.toml","sourceSha256":"bc9e1618a8db95233fbbf2f8ae898094011ac54dd9aad07785fda1b4d75ac7a2","variants":[],"readme":"High-performance transport layer for distributed agent communication in the RUV-FANN swarm intelligence framework. The ruv-swarm-transport crate provides a unified, async-first networking abstraction designed specifically for intelligent agent swarms. It delivers multiple transport backends optimized for different deployment scenarios, from high-throughput in-process communication to secure networked connections and low-latency shared memory transfers. - 🌐 WebSocket Transport: Full-duplex communication with TLS support, auto-reconnection, and connection pooling - 🔗 Shared Memory Transport: Zero-copy message passing with lock-free ring buffers for maximum throughput - ⚡ In-Process Transport","readmeSource":"ruv-fann/ruv-swarm/crates/ruv-swarm-transport/README.md","product":"key:ruv-fann:ruv-swarm-transport","canonical":"ruv-swarm-transport","tier":"T1","t":["ruv-swarm-transport","swarm","transport","layer","ruv-fann","fann","communication","websocket","sharedmemory","distribut","shared-memory","shar","memory"],"s":["ruv-swarm-transport","swarm","transport","distribut","shared-memory","shar","memory","websocket"]},
+{"id":"ruv-swarm-wasm-unified","family":"ruv-swarm-wasm-unified","kind":"crate","store":"ruv-fann","version":null,"description":"Unified WASM module exposing all ruv-FANN ecosystem capabilities","keywords":["forecasting","neural-network","ruv-fann","swarm","wasm"],"source":"ruv-fann/ruv-swarm/crates/ruv-swarm-wasm-unified/Cargo.toml","sourceSha256":"8955357a0aa3009f4b596b8c96be49a54f540f11ca19d94fcc3b4c8d1788c942","variants":[],"readme":"This crate provides a unified WebAssembly interface to the entire ruv-FANN ecosystem, including neural networks, forecasting models, and swarm orchestration capabilities. The unified WASM module exposes: - Core: Swarm orchestration, agent management, task distribution - Neural: ruv-FANN neural networks with 18 activation functions and 5 training algorithms - Forecasting: 27+ neuro-divergent forecasting models - Persistence: SQLite-based storage with WASM optimizations - SIMD Optimization: 4x speedup for vector operations (when available) - Memory Management: Smart pooling and allocation strategies - Progressive Loading: Load only required modules - TypeScript Support: Full type definitions g","readmeSource":"ruv-fann/ruv-swarm/crates/ruv-swarm-wasm-unified/README.md","product":"key:ruv-fann:ruv-swarm-wasm-unified","canonical":"ruv-swarm-wasm-unified","tier":"T1","t":["ruv-swarm-wasm-unified","swarm","unify","expos","all","ruv-fann","fann","ecosystem","capability","forecast","neural-network","neural","network"],"s":["ruv-swarm-wasm-unified","swarm","unify","forecast","neural-network","neural","network","ruv-fann","fann"]},
+{"id":"ruvbot","family":"ruvbot","kind":"npm","store":"ruvector","version":"0.3.3","description":"Enterprise-grade self-learning AI assistant with military-strength security, 150x faster vector search, and 12+ LLM models","keywords":["ai","ai-assistant","aidefence","chatbot","claude","cli","discord-bot","embeddings","enterprise","gemini","hnsw","llm","memory","multi-tenant","neural","openai","openrouter","prompt-injection","security","self-learning","slack-bot","typescript","vector-database","vector-search","wasm"],"source":"ruvector/npm/packages/ruvbot/package.json","sourceSha256":"179bb1eb566d84ab25c346c8be79606c4c44e7ff8c9b105ff0dbefdfa0c5cce0","variants":[],"readme":"Enterprise-Grade Self-Learning AI Assistant with Military-Strength Security Live Demo: https://ruvbot-875130704813.us-central1.run.app - [RVF Microkernel](#rvf-microkernel---self-contained-boot) - [Why RuvBot?](#why-ruvbot-over-clawdbot) - [Comparison](#ruvbot-vs-clawdbot-comparison) - [Requirements](#requirements) - [Quick Start](#quick-start) - [Configuration](#configuration) - [Channel Integrations](#channel-integrations) - [Template Library](#template-library) - [API Usage](#api-usage) - [Security](#security-architecture-6-layers---why-this-matters) - [LLM Providers](#llm-providers---gemini-25-default) - [TypeScript](#typescript-support) - [Events & Hooks](#events--hooks) - [Streaming](#","readmeSource":"ruvector/npm/packages/ruvbot/README.md","product":"key:ruvbot","canonical":"ruvbot","tier":"T0","t":["ruvbot","enterprise-grade","enterprise","grade","self-learning","self","learn","assistant","military-strength","military","strength","security","150x","vector","search","model","ai-assistant","aidefence","chatbot","claude","discord-bot","discord","bot","embedding","gemini","hnsw","memory","multi-tenant","multi","tenant","neural","openai","openrouter","prompt-injection","prompt","injection","slack-bot","slack","vector-database","database","vector-search"],"s":["ruvbot","ai-assistant","assistant","aidefence","chatbot","claude","discord-bot","discord","bot","embedding","enterprise","gemini","hnsw","memory","multi-tenant","multi","tenant","neural","openai","openrouter","prompt-injection","prompt","injection","security","self-learning","self","learn","slack-bot","slack","vector-database","vector","database","vector-search","search"]},
+{"id":"ruvdrone","family":"ruvdrone","kind":"npm","store":"ruv-drone","version":"0.1.0","description":"AI agent harness for ruv-drone — cooperative civilian UAV fleet coordination in Rust. Wired for all 9 hosts and verified against the real host runtimes.","keywords":["agent-harness","claude-code","codex","copilot","drone-swarm","example","formation-control","github-actions","hermes","metaharness","openclaw","opencode","path-planning","pi-dev","raft-consensus","rust","ruv-drone","rvm","uav"],"source":"ruv-drone/agent-harness/package.json","sourceSha256":"29a21f291c8801c0753b2e73b038b3faa6c100cd2c519028126cd54995d58f68","variants":[],"readme":"A repo-aware AI agent harness for ruv-drone (cooperative civilian UAV fleet coordination in Rust), minted with [MetaHarness](https://github.com/ruvnet/agent-harness-generator). It's a flagship example: one harness wired for all 9 supported hosts, and every host config was verified against the real host runtime (not just schema-checked). A small, repo-tuned agent: 4 agents (architect · implementer · reviewer · test-writer), a plan-change skill, doctor + review-diff commands, a default-deny MCP policy, and a memory namespace — all aware of ruv-drone's modules (formation, allocation, planning, marl, failsafe, …) and Rust workflow (cargo build/test/clippy/bench). The same harness drops into any","readmeSource":"ruv-drone/agent-harness/README.md","product":"key:ruv-drone:ruvdrone","canonical":"ruvdrone","tier":"T2","t":["ruvdrone","harness","ruv-drone","drone","cooperative","civilian","uav","fleet","coordination","wired","all","host","verify","against","real","runtime","agent-harness","claude-code","claude","code","codex","copilot","drone-swarm","swarm","example","formation-control","formation","control","github-actions","github","action","herme","metaharness","openclaw","opencode","path-planning","path","plann","pi-dev","dev","raft-consensus","raft","consensus","rvm"],"s":["ruvdrone","agent-harness","harness","claude-code","claude","code","codex","copilot","drone-swarm","drone","swarm","example","formation-control","formation","control","github-actions","github","action","herme","metaharness","openclaw","opencode","path-planning","path","plann","pi-dev","dev","raft-consensus","raft","consensus","ruv-drone","rvm","uav"]},
+{"id":"ruvector","family":"ruvector","kind":"npm","store":"ruvector","version":"0.1.2","description":"High-performance Rust-native vector database with AgenticDB compatibility","keywords":["agentic","ai","database","embeddings","hnsw","rust","vector"],"source":"ruvector/package.json","sourceSha256":"38f041d8325665274c26e9f5924e026e86f0426a30bbc6b912c56f9b53d5e7fd","variants":[],"readme":"RuVector is a Rust native substrate for fast local decisions and agent memory across sessions. It combines local semantic embeddings, persistent vector retrieval, graph relationships, explicit feedback learning, memory lifecycle controls, and optional shared memory. The default retrieval path runs locally. Learning happens from recorded outcomes and feedback, not from reads alone. Hosted services remain optional and create a separate data boundary. [@ruvector/typesafe](./npm/packages/typesafe) turns text into typed choice, score, and noul decisions using local embeddings and native decision heads, with a WASM fallback. It returns confidence and abstention information, supports labeled exampl","readmeSource":"ruvector/README.md","product":"repo:ruvector","canonical":"ruvector","tier":"T0","t":["high","rust-native","vector","database","agenticdb","compatibility","embedding","hnsw"],"s":["database","embedding","hnsw","vector"]},
+{"id":"ruvector-acorn","family":"acorn","kind":"crate","store":"ruvector","version":null,"description":"ACORN: Predicate-Agnostic Filtered HNSW — interleaved predicate evaluation inside the graph walk for 2-1000x QPS improvement over post-filter patterns at low selectivity","keywords":["acorn","filtered-search","hnsw","vector-search","wasm"],"source":"ruvector/crates/ruvector-acorn/Cargo.toml","sourceSha256":"484f6d03865ef8b43f64c8f4c3f8c8f00426c8e9591caf8ce3506b872d06a310","variants":["ruvector-acorn-wasm"],"product":"key:ruvector:acorn","canonical":"ruvector-acorn","tier":"T0","t":["ruvector-acorn","acorn","predicate-agnostic","predicate","agnostic","filter","hnsw","interleav","evaluation","inside","graph","walk","2-1000x","1000x","qps","improvement","post-filter","post","pattern","low","selectivity","filtered-search","search","vector-search","vector"],"s":["ruvector-acorn","acorn","filtered-search","filter","search","hnsw","vector-search","vector"]},
+{"id":"ruvector-adaptive-ann","family":"adaptive-ann","kind":"crate","store":"ruvector","version":null,"description":"Adaptive recall-targeted ANN search: automatic ef calibration to hit user-specified recall budgets without manual parameter tuning","keywords":["agent-memory","ann","hnsw","recall","vector-search"],"source":"ruvector/crates/ruvector-adaptive-ann/Cargo.toml","sourceSha256":"de9b694d18b9284ede1ef092080e0411d31b463f0bc42e28b6a43433355bac4a","variants":[],"readme":"Research implementations of empirical ef calibration for graph-based approximate nearest-neighbour search. Calibration tables are specific to the result count (k), graph, and sampled query distribution. They estimate mean recall observed during calibration and do not provide a per-query recall guarantee. Recalibrate and audit whenever the data, graph, or workload shifts. See docs/adr/ADR-272-adaptive-recall-ann.md and the associated nightly report for methodology and limitations.","readmeSource":"ruvector/crates/ruvector-adaptive-ann/README.md","product":"key:ruvector:adaptive-ann","canonical":"ruvector-adaptive-ann","tier":"T0","t":["ruvector-adaptive-ann","adaptive","ann","adaptive-ann","recall-targeted","recall","target","search","automatic","calibration","hit","user-specified","user","specify","budget","manual","parameter","tuning","agent-memory","memory","hnsw","vector-search","vector"],"s":["ruvector-adaptive-ann","adaptive","ann","adaptive-ann","agent-memory","memory","hnsw","recall","vector-search","vector","search"]},
+{"id":"ruvector-agent-memory","family":"agent-memory","kind":"crate","store":"ruvector","version":null,"description":"Coherence-weighted agent memory compaction for ruvector: retain important memories using recency, frequency, and semantic coherence","keywords":["agent-memory","memory-management","rag","ruvector","vector-search"],"source":"ruvector/crates/ruvector-agent-memory/Cargo.toml","sourceSha256":"c4f981c21d898a30a632ed369c9d8da2db630f289a1609f6fe34dc2a18a6c037","variants":[],"product":"key:ruvector:agent-memory","canonical":"ruvector-agent-memory","tier":"T0","t":["ruvector-agent-memory","memory","agent-memory","coherence-weighted","coherence","weight","compaction","retain","important","recency","frequency","semantic","memory-management","management","rag","vector-search","vector","search"],"s":["ruvector-agent-memory","memory","agent-memory","memory-management","management","rag","vector-search","vector","search"]},
+{"id":"ruvector-attention-unified-wasm","family":"attention-unified","kind":"crate","store":"ruvector","version":null,"description":"Unified WebAssembly bindings for 18+ attention mechanisms: Neural, DAG, Graph, and Mamba SSM","keywords":["attention","dag","mamba","neural","wasm"],"source":"ruvector/crates/ruvector-attention-unified-wasm/Cargo.toml","sourceSha256":"2c255c71c915fbdba83dc8f6bb80fcbcca53aeaa936743b1d601b0d6c8c7e34b","variants":[],"readme":"Unified WebAssembly bindings for 18+ attention mechanisms, combining Neural, DAG, Graph, and Mamba SSM attention types into a single npm package. Standard transformer-style attention mechanisms for sequence processing. For hierarchical data like trees and taxonomies. O(n) complexity for long sequences. Memory-efficient blocked attention for large sequences. Sparse attention with global tokens (like Longformer). Route tokens to specialized expert attention heads.","readmeSource":"ruvector/crates/ruvector-attention-unified-wasm/README.md","product":"key:ruvector:attention-unified","canonical":"ruvector-attention-unified-wasm","tier":"T0","t":["ruvector-attention-unified-wasm","attention","unify","attention-unified","mechanism","neural","dag","graph","mamba","ssm"],"s":["ruvector-attention-unified-wasm","attention","unify","attention-unified","dag","mamba","neural"]},
+{"id":"ruvector-attn-mincut","family":"attn-mincut","kind":"crate","store":"ruvector","version":null,"description":"Min-cut gating attention operator: dynamic graph-based alternative to softmax attention","keywords":["attention","gating","graph","min-cut","transformer"],"source":"ruvector/crates/ruvector-attn-mincut/Cargo.toml","sourceSha256":"c98025b8dd8840f2619345324912fc49f8d62dd61751180f6898d8575c6ba321","variants":[],"readme":"Dynamic min-cut gating as an alternative to softmax attention — prune low-value attention edges via graph theory. Standard transformer attention applies softmax uniformly across all Q*K^T logits, forcing every token to attend to every other token. This crate replaces that all-to-all pattern with a graph-theoretic approach: attention logits are modelled as a weighted directed graph, and Dinic's max-flow / min-cut algorithm partitions the graph to prune low-value attention edges. Only surviving edges pass through row-softmax before multiplying by V. The result is a sparse, structure-aware attention mask that can reduce KV-cache pressure, lower memory footprint, and cut energy-per-sample -- whi","readmeSource":"ruvector/crates/ruvector-attn-mincut/README.md","product":"key:ruvector:attn-mincut","canonical":"ruvector-attn-mincut","tier":"T0","t":["ruvector-attn-mincut","attn","mincut","attn-mincut","min-cut","min","cut","gating","attention","operator","dynamic","graph-based","graph","alternative","softmax","transformer"],"s":["ruvector-attn-mincut","attn","mincut","attn-mincut","attention","gating","graph","min-cut","min","cut","transformer"]},
+{"id":"ruvector-bounded-rag","family":"bounded-rag","kind":"crate","store":"ruvector","version":null,"description":"Bounded context RAG via MinCut graph partitioning for coherence-constrained agent memory retrieval","keywords":["agent-memory","graph","mincut","rag","vector-search"],"source":"ruvector/crates/ruvector-bounded-rag/Cargo.toml","sourceSha256":"0720c8a1fd5bc9df0ed86ea514a052923765e14a81270955d43756577a5bf278","variants":[],"readme":"Research implementations of context-budgeted retrieval for RuVector. The crate compares cosine top-k retrieval, priority traversal over a dense similarity graph, and an Edmonds–Karp min-cut partition followed by relevance ranking and budget truncation. The graph-based variants rebuild their pairwise similarity structures per query and are intended as auditable research baselines rather than production-scale indexes. See docs/adr/ADR-272-bounded-rag-mincut.md and the associated nightly research report for methodology and limitations.","readmeSource":"ruvector/crates/ruvector-bounded-rag/README.md","product":"key:ruvector:bounded-rag","canonical":"ruvector-bounded-rag","tier":"T0","t":["ruvector-bounded-rag","bound","rag","bounded-rag","context","via","mincut","graph","partition","coherence-constrained","coherence","constrain","memory","retrieval","agent-memory","vector-search","vector","search"],"s":["ruvector-bounded-rag","bound","rag","bounded-rag","agent-memory","memory","graph","mincut","vector-search","vector","search"]},
+{"id":"ruvector-capgated","family":"capgated","kind":"crate","store":"ruvector","version":null,"description":"Capability-gated ANN search for RuVector - per-vector read access control using bitset capability tokens with three measured retrieval variants","keywords":["access-control","agent-memory","ann","security","vector-search"],"source":"ruvector/crates/ruvector-capgated/Cargo.toml","sourceSha256":"c34313ad99a9cd866e71623c76176a918b61b9ef2013ee7f113acbb764b6894a","variants":[],"readme":"Capability-gated approximate nearest-neighbour search — per-vector read access control baked into the retrieval engine, using 64-bit bitset capability tokens. Pure Rust, zero external dependencies, WASM-safe. Most vector databases enforce access control at the *collection* level: everyone sharing an index can retrieve everyone else's vectors. For multi-tenant agent memory (thousands of agents on one index) that is a design gap, and one-collection-per-tenant does not scale. ruvector-capgated makes the access check part of the search itself: each vector carries a required CapMask, each query presents a held CapMask, and only vectors the querier is authorised for are ever returned. This is the","readmeSource":"ruvector/crates/ruvector-capgated/README.md","product":"key:ruvector:capgated","canonical":"ruvector-capgated","tier":"T0","t":["ruvector-capgated","capgat","capability-gated","capability","gated","ann","search","per-vector","per","vector","read","access","control","bitset","token","three","measur","retrieval","variant","access-control","agent-memory","memory","security","vector-search"],"s":["ruvector-capgated","capgat","access-control","access","control","agent-memory","memory","ann","security","vector-search","vector","search"]},
+{"id":"ruvector-cluster-rag","family":"cluster-rag","kind":"crate","store":"ruvector","version":null,"description":"Hierarchical cluster-summary retrieval for RuVector: RAPTOR-style two-level tree with coherence-weighted cluster scoring for agent memory RAG","keywords":["agent-memory","cluster","hierarchical","rag","vector-search"],"source":"ruvector/crates/ruvector-cluster-rag/Cargo.toml","sourceSha256":"2812bb9246b848a1f29dfdb31b0d21d3416b8ca3062e3faa0c1645027ef47772","variants":[],"readme":"Experimental hierarchical retrieval over a two-level tree of leaf vectors and k-means cluster summaries. It includes exhaustive, IVF-style, and coherence-weighted search variants. Clustered variants are approximate. Tune the cluster count, probe count, and coherence weight against workload-specific recall and latency targets.","readmeSource":"ruvector/crates/ruvector-cluster-rag/README.md","product":"key:ruvector:cluster-rag","canonical":"ruvector-cluster-rag","tier":"T0","t":["ruvector-cluster-rag","cluster","rag","cluster-rag","hierarchical","cluster-summary","summary","retrieval","raptor-style","raptor","style","two-level","level","tree","coherence-weighted","coherence","weight","scor","memory","agent-memory","vector-search","vector","search"],"s":["ruvector-cluster-rag","cluster","rag","cluster-rag","agent-memory","memory","hierarchical","vector-search","vector","search"]},
+{"id":"ruvector-cognitive-container","family":"cognitive-container","kind":"crate","store":"ruvector","version":null,"description":"Verifiable WASM cognitive container with canonical witness chains","keywords":["cognitive","container","deterministic","wasm","witness-chain"],"source":"ruvector/crates/ruvector-cognitive-container/Cargo.toml","sourceSha256":"e83d7b349aedd713db4d9c6294f0c76aaf957e6492325a5c11faf4c86bbbd132","variants":[],"readme":"Verifiable WASM cognitive container with canonical witness chains for the RuVector ecosystem. - Epoch Controller: Phase-budgeted tick execution (ingest/mincut/spectral/evidence/witness) - Memory Slab: Arena-based allocation for graph data - Witness Chain: Hash-linked chain of ContainerWitnessReceipt for deterministic verification - Cognitive Container: Full orchestration with snapshot/restore support MIT","readmeSource":"ruvector/crates/ruvector-cognitive-container/README.md","product":"key:ruvector:cognitive-container","canonical":"ruvector-cognitive-container","tier":"T0","t":["ruvector-cognitive-container","cognitive","container","cognitive-container","verifiable","canonical","witness","chain","deterministic","witness-chain"],"s":["ruvector-cognitive-container","cognitive","container","cognitive-container","deterministic","witness-chain","witness","chain"]},
+{"id":"ruvector-coherence","family":"coherence","kind":"crate","store":"ruvector","version":null,"description":"Coherence measurement proxies for comparing attention mechanisms","keywords":[],"source":"ruvector/crates/ruvector-coherence/Cargo.toml","sourceSha256":"bf3e00983cd6ac1c541677838871dbf3fdaea236ff76a50b8bedfa3ba9a86205","variants":[],"readme":"Quantitative coherence metrics for comparing attention mechanisms — measure what gating costs and what it preserves. When replacing softmax attention with a gated alternative (such as min-cut gating), the central question is: does the output stay coherent? This crate provides a suite of metrics, comparison utilities, quality guardrails, and batched evaluation tools to answer that question quantitatively. \"Coherence\" here means the degree to which gated attention outputs preserve the semantic and structural properties of baseline softmax outputs. The crate measures this through vector similarity, contradiction detection, mask overlap analysis, and statistical aggregation with confidence inter","readmeSource":"ruvector/crates/ruvector-coherence/README.md","product":"key:ruvector:coherence","canonical":"ruvector-coherence","tier":"T0","t":["ruvector-coherence","coherence","measurement","proxy","compar","attention","mechanism"],"s":["ruvector-coherence","coherence"]},
+{"id":"ruvector-coherence-hnsw","family":"coherence-hnsw","kind":"crate","store":"ruvector","version":null,"description":"Coherence-gated HNSW search: traversal-direction pruning for faster beam search with maintained recall","keywords":["agent-memory","ann","coherence","hnsw","vector-search"],"source":"ruvector/crates/ruvector-coherence-hnsw/Cargo.toml","sourceSha256":"822aa89c94c96d7dfa66d37363470367fc31ecad2dfce296da6411d0dc73b2b5","variants":[],"product":"key:ruvector:coherence-hnsw","canonical":"ruvector-coherence-hnsw","tier":"T0","t":["ruvector-coherence-hnsw","coherence","hnsw","coherence-hnsw","coherence-gated","gated","search","traversal-direction","traversal","direction","prun","beam","maintain","recall","agent-memory","memory","ann","vector-search","vector"],"s":["ruvector-coherence-hnsw","coherence","hnsw","coherence-hnsw","agent-memory","memory","ann","vector-search","vector","search"]},
+{"id":"ruvector-collections","family":"collections","kind":"crate","store":"ruvector","version":null,"description":"High-performance collection management for Ruvector vector databases","keywords":[],"source":"ruvector/crates/ruvector-collections/Cargo.toml","sourceSha256":"917f45ed1e9722a610d9dfceb04a048a45d91f6292b7e4be8f7f94526fa9390d","variants":[],"readme":"High-performance collection management for Ruvector vector databases. ruvector-collections provides multi-tenant collection support with isolated namespaces, schema management, and collection-level configuration. Part of the [Ruvector](https://github.com/ruvnet/ruvector) ecosystem. - Multi-Tenant: Isolated collections with separate namespaces - Schema Support: Define and enforce vector schemas - Collection Configs: Per-collection settings for dimensions, metrics - Thread-Safe: Concurrent access with DashMap - Metadata Support: Rich collection metadata and tagging - Collection CRUD: Create, read, update, delete collections - Namespace Isolation: Logical separation between collections - Schema","readmeSource":"ruvector/crates/ruvector-collections/README.md","product":"key:ruvector:collections","canonical":"ruvector-collections","tier":"T0","t":["ruvector-collections","collection","high","management","vector","databas"],"s":["ruvector-collections","collection"]},
+{"id":"ruvector-consciousness","family":"consciousness","kind":"crate","store":"ruvector","version":null,"description":"SOTA consciousness metrics: IIT Φ computation, causal emergence, effective information with SIMD acceleration and sublinear approximations","keywords":["consciousness","emergence","iit","phi","sublinear","wasm"],"source":"ruvector/crates/ruvector-consciousness/Cargo.toml","sourceSha256":"62c844e8b25aba2be62ef80f208b13d53f4723a17841afe3f1b14f56f7a0b6af","variants":["ruvector-consciousness-wasm"],"product":"key:ruvector:consciousness","canonical":"ruvector-consciousness","tier":"T0","t":["ruvector-consciousness","consciousness","sota","metric","iit","computation","causal","emergence","effective","information","acceleration","sublinear","approximation","phi"],"s":["ruvector-consciousness","consciousness","emergence","iit","phi","sublinear"]},
+{"id":"ruvector-context","family":"context","kind":"crate","store":"ruvector","version":null,"description":"Physically scope-sharded vector retrieval for governed context systems","keywords":["context","multitenancy","security","vector"],"source":"ruvector/crates/ruvector-context/Cargo.toml","sourceSha256":"5e470f46d1ba427a29a368ca92510a0efe2a94b201418409d6f71fb6a1873b5e","variants":[],"product":"key:ruvector:context","canonical":"ruvector-context","tier":"T0","t":["ruvector-context","context","physically","scope-sharded","scope","shard","vector","retrieval","govern","multitenancy","security"],"s":["ruvector-context","context","multitenancy","security","vector"]},
+{"id":"ruvector-crv","family":"crv","kind":"crate","store":"ruvector","version":null,"description":"CRV (Coordinate Remote Viewing) protocol integration for ruvector - maps 6-stage signal line methodology to vector database subsystems","keywords":["attention","crv","hyperbolic","signal-line","vector-search"],"source":"ruvector/crates/ruvector-crv/Cargo.toml","sourceSha256":"5821d73355dbe24fcce13f00cea2a941ec53bc54ffcc7022a4e894a917ae15b4","variants":[],"readme":"CRV (Coordinate Remote Viewing) protocol integration for RuVector — maps the 6-stage signal line methodology to vector database subsystems with Poincaré ball embeddings, multi-head attention, and MinCut partitioning. CRV (Coordinate Remote Viewing) protocol integration for ruvector. Maps the 6-stage CRV signal line methodology to ruvector's subsystems: The Poincaré ball embedding for Stage I gestalts encodes the hierarchical gestalt taxonomy (root → manmade/natural/movement/energy/water/land) with exponentially less distortion than Euclidean space. For AOL (Analytical Overlay) separation, the spiking neural network temporal encoding models signal-vs-noise discrimination: high-frequency spike","readmeSource":"ruvector/crates/ruvector-crv/README.md","product":"key:ruvector:crv","canonical":"ruvector-crv","tier":"T0","t":["ruvector-crv","crv","coordinate","remote","view","protocol","integration","maps","6-stage","stage","signal","line","methodology","vector","database","subsystem","attention","hyperbolic","signal-line","vector-search","search"],"s":["ruvector-crv","crv","attention","hyperbolic","signal-line","signal","line","vector-search","vector","search"]},
+{"id":"ruvector-dag","family":"dag","kind":"crate","store":"ruvector","version":null,"description":"Directed Acyclic Graph (DAG) structures for query plan optimization with neural learning","keywords":[],"source":"ruvector/crates/ruvector-dag/Cargo.toml","sourceSha256":"2fddffd13222f0e863645ebc7dd599b048d32557e62a6d3554e539d9a2455f54","variants":["ruvector-dag-wasm"],"readme":"Make your queries faster automatically. RuVector DAG learns from every query execution and continuously optimizes performance—no manual tuning required. RuVector DAG is a self-learning query optimization system. Think of it as a \"nervous system\" for your database queries that: 1. Watches how queries execute and identifies bottlenecks 2. Learns which optimization strategies work best for different query patterns 3. Adapts in real-time, switching strategies when conditions change 4. Heals itself by detecting anomalies and fixing problems before they impact users Unlike traditional query optimizers that use static rules, RuVector DAG learns from actual execution patterns and gets smarter over t","readmeSource":"ruvector/crates/ruvector-dag/README.md","product":"key:ruvector:dag","canonical":"ruvector-dag","tier":"T0","t":["ruvector-dag","dag","direct","acyclic","graph","structure","query","plan","neural","learn"],"s":["ruvector-dag","dag"]},
+{"id":"ruvector-decompiler","family":"decompiler","kind":"crate","store":"rudevolution","version":null,"description":"SOTA JavaScript bundle decompiler using MinCut graph partitioning, self-learning name inference, and RVF witness chains","keywords":["decompiler","javascript","mincut","reverse-engineering","source-map"],"source":"rudevolution/Cargo.toml","sourceSha256":"609e0522f57e77c528490911770719ee490704f51b48e4123cd16c8116017b4a","variants":[],"readme":"🧠 MinCut Module Detection &bull; 🔮 AI Name Recovery &bull; 🔗 Cryptographic Witness Chains &bull; 📊 Confidence Scoring &bull; 🧬 Self-Learning ruDevolution is a next-generation JavaScript decompiler built in pure Rust. It takes minified, obfuscated, or bundled JavaScript — the kind produced by esbuild, webpack, Terser, or any bundler — and reconstructs readable source code with original module boundaries, meaningful variable names, and full cryptographic provenance. Unlike traditional decompilers that only reformat whitespace, ruDevolution uses graph algorithms (MinCut partitioning) to detect where modules originally split apart, AI inference (neural + 210 pattern rules) to predict what v","readmeSource":"rudevolution/README.md","product":"repo:rudevolution","canonical":"ruvector-decompiler-wasm","tier":"T1","t":["ruvector-decompiler","decompiler","sota","bundle","mincut","graph","partition","self-learning","self","learn","name","inference","rvf","witness","chain","reverse-engineering","reverse","engineer","source-map","source","map"],"s":["ruvector-decompiler","decompiler","mincut","reverse-engineering","reverse","engineer","source-map","source","map"]},
+{"id":"ruvector-decompiler-wasm","family":"decompiler","kind":"npm","store":"ruvector","version":"2.1.0","description":"WASM bindings for the RuVector JavaScript bundle decompiler (Louvain pipeline)","keywords":["decompiler","javascript","louvain","mincut","wasm"],"source":"ruvector/npm/packages/ruvector/wasm/package.json","sourceSha256":"c73824071013e70000959c9b49b1db142cf65152d8982daaa508a454a6fba57f","variants":[],"product":"repo:rudevolution","canonical":"ruvector-decompiler-wasm","tier":"T0","t":["ruvector-decompiler-wasm","decompiler","bundle","louvain","pipeline","mincut"],"s":["ruvector-decompiler-wasm","decompiler","louvain","mincut"]},
+{"id":"ruvector-delta-consensus","family":"delta-consensus","kind":"crate","store":"ruvector","version":null,"description":"Distributed delta consensus using CRDTs and causal ordering","keywords":["causal","consensus","crdt","delta","distributed"],"source":"ruvector/crates/ruvector-delta-consensus/Cargo.toml","sourceSha256":"cb3011eeb8a7b0fbc6095d2cc8fcee7eeac6b707983146596d458d61c992eb32","variants":[],"product":"key:ruvector:delta-consensus","canonical":"ruvector-delta-consensus","tier":"T0","t":["ruvector-delta-consensus","delta","consensus","delta-consensus","distribut","crdt","causal","order"],"s":["ruvector-delta-consensus","delta","consensus","delta-consensus","causal","crdt","distribut"]},
+{"id":"ruvector-delta-core","family":"delta","kind":"crate","store":"ruvector","version":null,"description":"Core delta types and traits for behavioral vector change tracking","keywords":["database","delta","incremental","simd","streaming","vector","wasm"],"source":"ruvector/crates/ruvector-delta-core/Cargo.toml","sourceSha256":"31402edde3e42df37ec62fa679bd1de77d54d87ad59d22aa0b1c850a503489ea","variants":["ruvector-delta-wasm"],"product":"key:ruvector:delta","canonical":"ruvector-delta-core","tier":"T0","t":["ruvector-delta-core","delta","core","type","trait","behavioral","vector","change","track","database","incremental","stream"],"s":["ruvector-delta-core","delta","core","database","incremental","stream","vector"]},
+{"id":"ruvector-delta-graph","family":"delta-graph","kind":"crate","store":"ruvector","version":null,"description":"Delta operations for graph structures - edge and node changes","keywords":["delta","graph","incremental","streaming"],"source":"ruvector/crates/ruvector-delta-graph/Cargo.toml","sourceSha256":"605f5751d3fa7a634075fc1940a9c2237c71b0573592cf945d193021db0e06e9","variants":[],"product":"key:ruvector:delta-graph","canonical":"ruvector-delta-graph","tier":"T0","t":["ruvector-delta-graph","delta","graph","delta-graph","operation","structure","edge","change","incremental","stream"],"s":["ruvector-delta-graph","delta","graph","delta-graph","incremental","stream"]},
+{"id":"ruvector-delta-index","family":"delta-index","kind":"crate","store":"ruvector","version":null,"description":"Delta-aware HNSW index with incremental updates and repair strategies","keywords":["delta","hnsw","incremental","index","vector"],"source":"ruvector/crates/ruvector-delta-index/Cargo.toml","sourceSha256":"3a15239e1cad970431b95af66be0e371985ea7bd2d1b46aea6f5e314a1f30c87","variants":[],"product":"key:ruvector:delta-index","canonical":"ruvector-delta-index","tier":"T0","t":["ruvector-delta-index","delta","index","delta-index","delta-aware","aware","hnsw","incremental","update","repair","strategy","vector"],"s":["ruvector-delta-index","delta","index","delta-index","hnsw","incremental","vector"]},
+{"id":"ruvector-dither","family":"dither","kind":"crate","store":"ruvector","version":null,"description":"Deterministic low-discrepancy dithering for low-bit quantization: golden-ratio and π-digit sequences for blue-noise error shaping","keywords":["dither","golden-ratio","inference","quantization","wasm"],"source":"ruvector/crates/ruvector-dither/Cargo.toml","sourceSha256":"e2c29afcbefb2ad428e6b74f0ca56a6ff59291dfc69924f7b0da7d6b39fc8ee2","variants":[],"readme":"Deterministic, low-discrepancy pre-quantization dithering for low-bit neural network inference on tiny devices (WASM, Seed, STM32). Quantizers at 3/5/7 bits can align with power-of-two boundaries, producing idle tones, sticky activations, and periodic errors that degrade accuracy. A sub-LSB pre-quantization offset: - Decorrelates the signal from grid boundaries. - Pushes quantization error toward high frequencies (blue-noise-like), which average out downstream. - Uses no RNG -- outputs are deterministic, reproducible across platforms (WASM / x86 / ARM), and cache-friendly. - Golden-ratio sequence -- best 1-D equidistribution, irrational period (never repeats). - Pi-digit table -- 256-byte cy","readmeSource":"ruvector/crates/ruvector-dither/README.md","product":"key:ruvector:dither","canonical":"ruvector-dither","tier":"T0","t":["ruvector-dither","dither","deterministic","low-discrepancy","low","discrepancy","low-bit","bit","quantization","golden-ratio","golden","ratio","digit","sequence","blue-noise","blue","noise","error","shap","inference"],"s":["ruvector-dither","dither","golden-ratio","golden","ratio","inference","quantization"]},
+{"id":"ruvector-diverse-beam","family":"diverse-beam","kind":"crate","store":"ruvector","version":null,"description":"Diverse beam ANN search - MMR and coherence-pruned beam traversal with three measured variants for RuVector","keywords":["agent-memory","ann","beam-search","diversity","vector-search"],"source":"ruvector/crates/ruvector-diverse-beam/Cargo.toml","sourceSha256":"7d4b6b8a19afcd6285b06357b3f448f805ce8d03e9f6d703d1a686a17e80129f","variants":[],"readme":"Research implementations of diverse approximate-nearest-neighbour search for RuVector. The crate compares: - greedy beam traversal; - greedy traversal followed by Maximum Marginal Relevance reranking; and - coherence-pruned traversal. The implementation uses an exact flat k-nearest-neighbour graph so that the search strategies can be measured in isolation. It is intended for research and evaluation, not as a replacement for RuVector's production HNSW index. Run the tests and benchmark with: See docs/adr/ADR-272-diverse-beam-ann.md and the nightly research report for methodology, limitations, and measured results.","readmeSource":"ruvector/crates/ruvector-diverse-beam/README.md","product":"key:ruvector:diverse-beam","canonical":"ruvector-diverse-beam","tier":"T0","t":["ruvector-diverse-beam","diverse","beam","diverse-beam","ann","search","mmr","coherence-pruned","coherence","prun","traversal","three","measur","variant","agent-memory","memory","beam-search","diversity","vector-search","vector"],"s":["ruvector-diverse-beam","diverse","beam","diverse-beam","agent-memory","memory","ann","beam-search","search","diversity","vector-search","vector"]},
+{"id":"ruvector-domain-expansion","family":"domain-expansion","kind":"crate","store":"ruvector","version":null,"description":"Cross-domain transfer learning engine: Rust synthesis, structured planning, tool orchestration","keywords":["domain-expansion","generalization","planning","rust-synthesis","transfer-learning","wasm"],"source":"ruvector/crates/ruvector-domain-expansion/Cargo.toml","sourceSha256":"bd4715c35d8fb12a6f7a961e3371555e9e4086ff776a800a8853ff232e38f818","variants":["ruvector-domain-expansion-wasm"],"readme":"Cross-domain transfer learning — train on one problem, get better at a different one automatically. Most AI systems learn one task at a time. Train a model on genomics and it can't trade stocks. Teach it quantum circuits and it won't plan workflows. ruvector-domain-expansion changes that: knowledge learned in one domain automatically transfers to other domains — and it proves the transfer actually helped before committing it. Genomics priors seed molecular design. Trading risk models improve resource allocation. Quantum noise detection accelerates signal processing. This is how real generalization works. Part of the [RuVector](https://github.com/ruvnet/ruvector) ecosystem.","readmeSource":"ruvector/crates/ruvector-domain-expansion/README.md","product":"key:ruvector:domain-expansion","canonical":"ruvector-domain-expansion","tier":"T0","t":["ruvector-domain-expansion","domain","expansion","domain-expansion","cross-domain","cross","transfer","learn","synthesis","structur","plann","orchestration","generalization","rust-synthesis","transfer-learning"],"s":["ruvector-domain-expansion","domain","expansion","domain-expansion","generalization","plann","rust-synthesis","synthesis","transfer-learning","transfer","learn"]},
+{"id":"ruvector-economy-wasm","family":"economy","kind":"crate","store":"ruvector","version":null,"description":"CRDT-based autonomous credit economy for distributed compute networks - WASM optimized","keywords":["crdt","distributed-economy","p2p-credits","reputation","wasm"],"source":"ruvector/crates/ruvector-economy-wasm/Cargo.toml","sourceSha256":"784cb1ab6f29c35b6506083243b4baafbf2e48ad91d45e4028226dd11a68c71f","variants":[],"readme":"CRDT-based autonomous credit economy for distributed compute networks. Designed for WASM execution with P2P consistency guarantees. The core CRDT ledger for tracking credits earned, spent, and staked. Early adopter reward multiplier with exponential decay.","readmeSource":"ruvector/crates/ruvector-economy-wasm/README.md","product":"key:ruvector:economy","canonical":"ruvector-economy-wasm","tier":"T0","t":["ruvector-economy-wasm","economy","crdt-based","crdt","autonomous","credit","distribut","compute","network","distributed-economy","p2p-credits","p2p","reputation"],"s":["ruvector-economy-wasm","economy","crdt","distributed-economy","distribut","p2p-credits","p2p","credit","reputation"]},
+{"id":"ruvector-edge-analytics","family":"edge-analytics","kind":"crate","store":"ruvector","version":null,"description":"ruvector edge rv-mincut core (ADR-351 §3, §10, M4): tenant graphs persisted as a versioned, checksummed, chunked edge list, rebuilt through ruvector-mincut's MinCutBuilder, with exact min-cut queries (approximate requests answered exactly) under vertex/edge limits and a work/memory budget (budget_exceeded -> 413) and an async job descriptor for large graphs. Pure Rust, wasm32 clean.","keywords":[],"source":"ruvector/crates/ruvector-edge-analytics/Cargo.toml","sourceSha256":"910d707a2c27e02eb07881adc2a4b55553576a2845ef76bd851d9befcba861ee","variants":[],"product":"key:ruvector:edge-analytics","canonical":"ruvector-edge-analytics","tier":"T0","t":["ruvector-edge-analytics","edge","analytic","edge-analytics","rv-mincut","mincut","core","adr-351","adr","351","tenant","graph","persist","version","checksum","chunk","list","rebuilt","through","ruvector-mincut","mincutbuilder","exact","min-cut","min","cut","query","approximate","request","answer","exactly","vertex","limit","memory","budget","exceed","413","async","job","descriptor","large","pure","wasm32","clean"],"s":["ruvector-edge-analytics","edge","analytic","edge-analytics"]},
+{"id":"ruvector-edge-auth","family":"edge-auth","kind":"crate","store":"ruvector","version":null,"description":"ruvector edge resource-server auth (ADR-351): ES256 JWT/JWKS verification, claims and audience policy, scope->capability table, RFC 9728 metadata. Pure Rust, wasm32-unknown-unknown clean.","keywords":[],"source":"ruvector/crates/ruvector-edge-auth/Cargo.toml","sourceSha256":"d79e4261a9729a659506e92d300f359f163135b4dbb98ef0264960ad42956be5","variants":[],"product":"key:ruvector:edge-auth","canonical":"ruvector-edge-auth","tier":"T0","t":["ruvector-edge-auth","edge","auth","edge-auth","resource-server","resource","server","adr-351","adr","351","es256","jwt","jwks","verification","claim","audience","policy","scope","capability","table","rfc","9728","metadata","pure","wasm32-unknown-unknown","wasm32","unknown","clean"],"s":["ruvector-edge-auth","edge","auth","edge-auth"]},
+{"id":"ruvector-edge-auth-worker","family":"edge-auth-worker","kind":"crate","store":"ruvector","version":null,"description":"ruvector-edge-auth authorization-server Worker: federates login to auth.cognitum.one and mints resource-bound ES256 access tokens. Durable Object SQLite holds clients, codes, flows and refresh families.","keywords":[],"source":"ruvector/edge/auth-worker/Cargo.toml","sourceSha256":"6877fb5be90495c9987e05a275d8003d30d2c9b50d14f0c69c7adb2a45562f9c","variants":[],"product":"key:ruvector:edge-auth-worker","canonical":"ruvector-edge-auth-worker","tier":"T0","t":["ruvector-edge-auth-worker","edge","auth","worker","edge-auth-worker","ruvector-edge-auth","authorization-server","authorization","server","federate","login","auth.cognitum.one","mint","resource-bound","resource","bound","es256","access","token","durable","object","sqlite","hold","client","code","flow","refresh","family"],"s":["ruvector-edge-auth-worker","edge","auth","worker","edge-auth-worker"]},
+{"id":"ruvector-edge-authz","family":"edge-authz","kind":"crate","store":"ruvector","version":null,"description":"Core of the ruvector edge OAuth 2.1 authorization server: DCR (RFC 7591), PKCE S256, one-time codes, upstream auth.cognitum.one federation, resource-bound ES256 access tokens (RFC 8707/9068), refresh rotation with reuse detection, revocation (RFC 7009), RFC 8414 metadata. Pure Rust, wasm32 clean.","keywords":[],"source":"ruvector/crates/ruvector-edge-authz/Cargo.toml","sourceSha256":"27776a2906effc4283c06a02a65dbedc253cfdd415b6478bbf28f8678430bd40","variants":[],"product":"key:ruvector:edge-authz","canonical":"ruvector-edge-authz","tier":"T0","t":["ruvector-edge-authz","edge","authz","edge-authz","core","oauth","2.1","authorization","server","dcr","rfc","7591","pkce","s256","one-time","code","upstream","auth.cognitum.one","federation","resource-bound","resource","bound","es256","access","token","8707","9068","refresh","rotation","reuse","detection","revocation","7009","8414","metadata","pure","wasm32","clean"],"s":["ruvector-edge-authz","edge","authz","edge-authz"]},
+{"id":"ruvector-edge-gateway","family":"edge-gateway","kind":"crate","store":"ruvector","version":null,"description":"rv-gateway Worker (ADR-351 M1): /v1/health, RFC 9728 metadata, /v1/me, REST collections/vectors, /v1/ops and /v1/mcp over TenantLedger / VectorShard Durable Objects; accepts only edge-auth-issued, resource-bound ES256 tokens.","keywords":[],"source":"ruvector/edge/gateway/Cargo.toml","sourceSha256":"26d4e7e6e3a229c76608461f9e09fedf6d401bebbb8a864aa78e77d42e8355a8","variants":[],"product":"key:ruvector:edge-gateway","canonical":"ruvector-edge-gateway","tier":"T0","t":["ruvector-edge-gateway","edge","gateway","edge-gateway","rv-gateway","worker","adr-351","adr","351","health","rfc","9728","metadata","rest","collection","vector","ops","mcp","tenantledger","vectorshard","durable","object","accept","only","edge-auth-issued","auth","issu","resource-bound","resource","bound","es256","token"],"s":["ruvector-edge-gateway","edge","gateway","edge-gateway"]},
+{"id":"ruvector-edge-index","family":"edge-index","kind":"crate","store":"ruvector","version":null,"description":"ruvector edge in-memory vector index (ADR-351 §6.1 M2a/M2b): versioned int8 scalar quantizer, quantized flat scan with batched f32 rerank, HNSW traversing u8 codes with fixed-stride u32 adjacency, checksummed <= 1 MiB index_chunks persistence and memory accounting. Pure Rust, wasm32 clean (no getrandom, threads or clock).","keywords":[],"source":"ruvector/crates/ruvector-edge-index/Cargo.toml","sourceSha256":"bac3fed50fea55533f90d5e77e411bf667f34023d47eb4f1c29d08aa1c5d8ce4","variants":[],"product":"key:ruvector:edge-index","canonical":"ruvector-edge-index","tier":"T0","t":["ruvector-edge-index","edge","index","edge-index","in-memory","memory","vector","adr-351","adr","351","6.1","m2a","m2b","version","int8","scalar","quantizer","quantiz","flat","scan","batch","f32","rerank","hnsw","travers","code","fixed-stride","fixed","stride","u32","adjacency","checksum","mib","chunk","persistence","account","pure","wasm32","clean","getrandom","thread","clock"],"s":["ruvector-edge-index","edge","index","edge-index"]},
+{"id":"ruvector-edge-quant","family":"edge-quant","kind":"crate","store":"ruvector","version":null,"description":"ruvector edge rv-quant core (ADR-351 §3, M4): a tenant-shard RaBitQ index over ruvector-rabitq with a chunked, checksummed persist v2 format (packed codes + rotation seed, no f32 rebuild on cold load), memory/CPU budget accounting (413 budget_exceeded) and optional exact f32 rerank through a callback. Pure Rust, wasm32 clean.","keywords":[],"source":"ruvector/crates/ruvector-edge-quant/Cargo.toml","sourceSha256":"405878e09dc6bbd4db69dbaba8efa1093cf50498d443dd6577d4a30a847a0503","variants":[],"product":"key:ruvector:edge-quant","canonical":"ruvector-edge-quant","tier":"T0","t":["ruvector-edge-quant","edge","quant","edge-quant","rv-quant","core","adr-351","adr","351","tenant-shard","tenant","shard","rabitq","index","ruvector-rabitq","chunk","checksum","persist","format","pack","code","rotation","seed","f32","rebuild","cold","load","memory","cpu","budget","account","413","exceed","optional","exact","rerank","through","callback","pure","wasm32","clean"],"s":["ruvector-edge-quant","edge","quant","edge-quant"]},
+{"id":"ruvector-edge-registry","family":"edge-registry","kind":"crate","store":"ruvector","version":null,"description":"ruvector edge RVF registry core (ADR-351 §3 rv-registry, M5): scoped package names, immutable semver versions with yanking, package manifests, a streaming strict RVF validator over untrusted bytes (rvf-wire), content-addressed R2 blob keys, upload sessions, the visibility/authorization matrix and paginated listing. Pure Rust, wasm32 clean.","keywords":[],"source":"ruvector/crates/ruvector-edge-registry/Cargo.toml","sourceSha256":"a90b63276b1d10d0825a01bba656a9ad1136f89d235179c042169b108db06bc0","variants":[],"product":"key:ruvector:edge-registry","canonical":"ruvector-edge-registry","tier":"T0","t":["ruvector-edge-registry","edge","registry","edge-registry","rvf","core","adr-351","adr","351","rv-registry","scop","name","immutable","semver","version","yank","manifest","stream","strict","validator","untrust","byte","rvf-wire","wire","content-addressed","content","addres","blob","keys","upload","session","visibility","authorization","matrix","paginat","list","pure","wasm32","clean"],"s":["ruvector-edge-registry","edge","registry","edge-registry"]},
+{"id":"ruvector-edge-snapshot","family":"edge-snapshot","kind":"crate","store":"ruvector","version":null,"description":"ruvector edge durability core (ADR-351 §3 rv-snapshot/rv-embed/rv-ingest, §6.3, §15 M3): chunked RVF snapshots with sha256 manifests, per-tenant witness chain, restore validation, streamed RVF import/export, resumable bulk-import jobs and an embedding batcher. Pure Rust, wasm32 clean.","keywords":[],"source":"ruvector/crates/ruvector-edge-snapshot/Cargo.toml","sourceSha256":"58ee5ea2abe811f783e30908022c62a7ba5764dbbe5f4a1168730a9a8283bf04","variants":[],"product":"key:ruvector:edge-snapshot","canonical":"ruvector-edge-snapshot","tier":"T0","t":["ruvector-edge-snapshot","edge","snapshot","edge-snapshot","durability","core","adr-351","adr","351","rv-snapshot","rv-embed","embed","rv-ingest","ingest","6.3","chunk","rvf","sha256","manifest","per-tenant","per","tenant","witness","chain","restore","validation","stream","import","export","resumable","bulk-import","bulk","jobs","embedd","batcher","pure","wasm32","clean"],"s":["ruvector-edge-snapshot","edge","snapshot","edge-snapshot"]},
+{"id":"ruvector-edge-store","family":"edge-store","kind":"crate","store":"ruvector","version":null,"description":"ruvector edge storage core (ADR-351 §6, §16): VectorShard (int8 flat or HNSW index with f32 rerank from SQLite, chunked index persistence, op log, replay, resident cap), TenantLedger (catalog, collection_uid, memberships, quotas, op_id idempotency) and the /v1/ops dispatcher, behind a SqlStore port with an in-memory mock. Pure Rust, wasm32 clean.","keywords":[],"source":"ruvector/crates/ruvector-edge-store/Cargo.toml","sourceSha256":"2c0be870c32d05671897e5b95ce60925b4503e8d16ebb72014fb630980705057","variants":[],"product":"key:ruvector:edge-store","canonical":"ruvector-edge-store","tier":"T0","t":["ruvector-edge-store","edge","store","edge-store","storage","core","adr-351","adr","351","vectorshard","int8","flat","hnsw","index","f32","rerank","sqlite","chunk","persistence","log","replay","resident","cap","tenantledger","catalog","collection","uid","membership","quota","idempotency","ops","dispatcher","behind","sqlstore","port","in-memory","memory","mock","pure","wasm32","clean"],"s":["ruvector-edge-store","edge","store","edge-store"]},
+{"id":"ruvector-edge-tenancy","family":"edge-tenancy","kind":"crate","store":"ruvector","version":null,"description":"ruvector edge tenancy (ADR-351 §4): TenantContext from verified claims only, tenant_key and Durable Object name derivation, input validators, quota types, RFC 9457 problems. Pure Rust, wasm32 clean.","keywords":[],"source":"ruvector/crates/ruvector-edge-tenancy/Cargo.toml","sourceSha256":"3d03a4222843aa762f2ebddd56662842e03f2729a945c8979e1ccf2e03528c08","variants":[],"product":"key:ruvector:edge-tenancy","canonical":"ruvector-edge-tenancy","tier":"T0","t":["ruvector-edge-tenancy","edge","tenancy","edge-tenancy","adr-351","adr","351","tenantcontext","verify","claim","only","tenant","key","durable","object","name","derivation","input","validator","quota","type","rfc","9457","problem","pure","wasm32","clean"],"s":["ruvector-edge-tenancy","edge","tenancy","edge-tenancy"]},
+{"id":"ruvector-embed-core","family":"embed","kind":"crate","store":"ruvector","version":null,"description":"Sentence-embedding backends for @ruvector/typesafe: `ort` (native) and `tract` (wasm) implementations of the `Embedder` trait, with a hash-pinned model manifest (ADR-002, ADR-005).","keywords":[],"source":"ruvector/crates/ruvector-embed-core/Cargo.toml","sourceSha256":"2e6a04a72dae8185d76b5978c6245868c9c6b91083eaccf9c6d20a350cd5a034","variants":[],"readme":"Sentence-embedding backends for @ruvector/typesafe: two implementations of the [Embedder](../ruvector-typesafe-core/src/embedder.rs) trait, one per target (ADR-002). cargo build --workspace (no features) compiles no inference engine and downloads nothing: only the manifest, pooling and error types are present. - Weights are pinned by a content-hash [model manifest](../../npm/packages/typesafe/models/manifest.json) ({name, file, sha256, dims, license, source_url, added, review_by, pooling, tokenizer_file, tokenizer_sha256, max_tokens}) and verified at load. A mismatch is a typed EmbedError::HashMismatch and nothing is loaded — fail closed (ADR-005). The optional tokenizer_sha256 closes the \"p","readmeSource":"ruvector/crates/ruvector-embed-core/README.md","product":"key:ruvector:embed","canonical":"ruvector-embed-core","tier":"T0","t":["ruvector-embed-core","embed","core","sentence-embedding","sentence","embedd","backend","typesafe","ort","tract","implementation","embedder","trait","hash-pinned","hash","pin","model","manifest","adr-002","adr","002","adr-005","005"],"s":["ruvector-embed-core","embed","core"]},
+{"id":"ruvector-entropy-ann","family":"entropy-ann","kind":"crate","store":"ruvector","version":null,"description":"Entropy-adaptive beam search for ANN: live Shannon entropy of the candidate-heap distance distribution controls beam width during graph traversal","keywords":["adaptive","ann","entropy","hnsw","vector-search"],"source":"ruvector/crates/ruvector-entropy-ann/Cargo.toml","sourceSha256":"6155f9c6f213c1dd724fa837a6e450af2c92e57857d3355578fe365e0498e008","variants":[],"product":"key:ruvector:entropy-ann","canonical":"ruvector-entropy-ann","tier":"T0","t":["ruvector-entropy-ann","entropy","ann","entropy-ann","entropy-adaptive","adaptive","beam","search","live","shannon","candidate-heap","candidate","heap","distance","distribution","control","width","during","graph","traversal","hnsw","vector-search","vector"],"s":["ruvector-entropy-ann","entropy","ann","entropy-ann","adaptive","hnsw","vector-search","vector","search"]},
+{"id":"ruvector-exotic-wasm","family":"exotic","kind":"crate","store":"ruvector","version":null,"description":"Exotic AI mechanisms for emergent behavior - Neural Autonomous Orgs, Morphogenetic Networks, Time Crystals","keywords":[],"source":"ruvector/crates/ruvector-exotic-wasm/Cargo.toml","sourceSha256":"d0b86e823f34bbe56c08395daf883b9bd6d8e1e7036cd2e19d89a7f5ac9bc373","variants":[],"readme":"Exotic AI mechanisms for emergent behavior in distributed systems. This WASM module provides novel coordination primitives inspired by decentralized governance, developmental biology, and quantum physics. Decentralized governance for AI agent collectives using stake-weighted quadratic voting and oscillatory synchronization for coherence. Creates a new NAO with the specified quorum threshold (0.0 - 1.0). Biologically-inspired network growth with cellular differentiation through morphogen gradients, emergent network topology, and synaptic pruning. Creates a new morphogenetic network with the specified grid dimensions.","readmeSource":"ruvector/crates/ruvector-exotic-wasm/README.md","product":"key:ruvector:exotic","canonical":"ruvector-exotic-wasm","tier":"T0","t":["ruvector-exotic-wasm","exotic","mechanism","emergent","behavior","neural","autonomous","orgs","morphogenetic","network","crystal"],"s":["ruvector-exotic-wasm","exotic"]},
+{"id":"ruvector-extensions","family":"extensions","kind":"npm","store":"ruvector","version":"0.1.2","description":"Advanced features for ruvector: embeddings, UI, exports, temporal tracking, and persistence","keywords":["cohere","embeddings","graph-visualization","neo4j","openai","persistence","ruvector","temporal-tracking"],"source":"ruvector/npm/packages/ruvector-extensions/package.json","sourceSha256":"c3b85e5bb2d29f818dacdbac40139eeb56fe3fbf8aeca93d06a321bfaf91e9b8","variants":[],"readme":"Advanced persistence and extension features for the [ruvector](https://github.com/ruvnet/ruvector) vector database. - 💾 Multiple Persistence Formats: JSON, Binary (MessagePack), SQLite - 📸 Snapshot Management: Create, list, restore, and delete database snapshots - 🔄 Incremental Saves: Save only changed data for efficiency - 📤 Export/Import: Flexible data portability - 🗜️ Compression Support: Gzip and Brotli compression for smaller files - 📊 Progress Tracking: Real-time progress callbacks for large operations - ⚡ Auto-Save: Configurable automatic saves - 🔒 Data Integrity: Built-in checksum verification Main class for managing database persistence. Options: - baseDir (string): Base dire","readmeSource":"ruvector/npm/packages/ruvector-extensions/README.md","product":"key:ruvector:extensions","canonical":"ruvector-extensions","tier":"T0","t":["ruvector-extensions","extension","advanc","feature","embedding","export","temporal","track","persistence","cohere","graph-visualization","graph","visualization","neo4j","openai","temporal-tracking"],"s":["ruvector-extensions","extension","cohere","embedding","graph-visualization","graph","visualization","neo4j","openai","persistence","temporal-tracking","temporal","track"]},
+{"id":"ruvector-filter","family":"filter","kind":"crate","store":"ruvector","version":null,"description":"Advanced metadata filtering for Ruvector vector search","keywords":[],"source":"ruvector/crates/ruvector-filter/Cargo.toml","sourceSha256":"d2463581ecaf9c8b7486dc2a4612944fefbd9d7052c087a6850dba6be0280cc4","variants":[],"readme":"Advanced metadata filtering for Ruvector vector search. ruvector-filter provides a powerful filter expression language for combining vector similarity search with metadata constraints. Supports complex boolean expressions, range queries, and efficient filter evaluation. Part of the [Ruvector](https://github.com/ruvnet/ruvector) ecosystem. - Rich Expressions: Complex boolean filter expressions - Type-Safe: Strongly typed filter operations - Optimized: Filter pushdown for efficient evaluation - Extensible: Custom filter operators - JSON Compatible: Easy integration with JSON metadata - Comparison Operators: =, !=, <, >, <=, >= - Boolean Logic: AND, OR, NOT - Range Queries: BETWEEN, IN - String","readmeSource":"ruvector/crates/ruvector-filter/README.md","product":"key:ruvector:filter","canonical":"ruvector-filter","tier":"T0","t":["ruvector-filter","filter","advanc","metadata","vector","search"],"s":["ruvector-filter","filter"]},
+{"id":"ruvector-fpga-transformer","family":"fpga-transformer","kind":"crate","store":"ruvector","version":null,"description":"FPGA Transformer backend with deterministic latency, quantization-first design, and coherence gating","keywords":["fpga","inference","low-latency","quantization","transformer","wasm"],"source":"ruvector/crates/ruvector-fpga-transformer/Cargo.toml","sourceSha256":"54c84fa2d81979d2ac086b1a81b206217dc55fa2efa54f8616f3ae0ce61d1e87","variants":["ruvector-fpga-transformer-wasm"],"readme":"Run AI models on specialized hardware with predictable, ultra-low latency. FPGA Transformer is a Rust library that lets you run transformer neural networks (like those used in ChatGPT, code completion, and other AI applications) on FPGA hardware instead of GPUs. This gives you consistent, predictable response times - essential for real-time applications. Pre-defined configurations for common use cases: Skip unnecessary computation when the model is already confident: Gate Decisions: - RanFull - Model ran all layers - EarlyExit { layer } - Exited early at specified layer - Skipped { reason } - Computation was blocked Convert floating-point models to efficient integer math:","readmeSource":"ruvector/crates/ruvector-fpga-transformer/README.md","product":"key:ruvector:fpga-transformer","canonical":"ruvector-fpga-transformer","tier":"T0","t":["ruvector-fpga-transformer","fpga","transformer","fpga-transformer","backend","deterministic","latency","quantization-first","quantization","design","coherence","gating","inference","low-latency","low"],"s":["ruvector-fpga-transformer","fpga","transformer","fpga-transformer","inference","low-latency","low","latency","quantization"]},
+{"id":"ruvector-gnn-rerank","family":"gnn-rerank","kind":"crate","store":"ruvector","version":null,"description":"GNN-enhanced candidate reranking for approximate ANN search in ruvector","keywords":["ann","gnn","rag","reranking","vector-search"],"source":"ruvector/crates/ruvector-gnn-rerank/Cargo.toml","sourceSha256":"2bbe7f1b0b2b979431805bbd0de360faaa682d739b807cb57eacb5e5f1b03739","variants":[],"readme":"GNN score-diffusion reranking for approximate ANN search — recovers recall lost to quantization by smoothing candidate scores over a k-NN graph of the (full-precision) candidate vectors. Part of the [ruvector](https://github.com/ruvnet/ruvector) ecosystem. A quantized / coarse ANN index returns approximately-correct candidates but mis-ranks them near the top-k boundary. GnnDiffusion builds a k-NN graph over the candidates' full-precision vectors and diffuses each candidate's score across its graph neighbours — pulling true neighbours (which got an unluckily-low noisy score) back up. This is a recall/latency tradeoff, not free throughput. On the same benchmark: Right for a rerank stage over a","readmeSource":"ruvector/crates/ruvector-gnn-rerank/README.md","product":"key:ruvector:gnn-rerank","canonical":"ruvector-gnn-rerank","tier":"T0","t":["ruvector-gnn-rerank","gnn","rerank","gnn-rerank","gnn-enhanced","enhanc","candidate","approximate","ann","search","rag","vector-search","vector"],"s":["ruvector-gnn-rerank","gnn","rerank","gnn-rerank","ann","rag","vector-search","vector","search"]},
+{"id":"ruvector-gpu-runner","family":"gpu-runner","kind":"crate","store":"ruvector","version":null,"description":"Budget-capped, auto-destroying vast.ai GPU job runner for ruvector training jobs","keywords":[],"source":"ruvector/crates/ruvector-gpu-runner/Cargo.toml","sourceSha256":"4c552716a8c925aa678e4c36ba45e45fc3419b34a411f1839b084754cb696413","variants":[],"readme":"Runs one training job on a rented vast.ai GPU. It enforces a budget cap, destroys the instance when the job ends, and uploads artifacts to GCS. Written in Rust, with blocking HTTP via ureq and no async runtime. It does not use the Python vastai CLI. The exit code is 0 when the job succeeds or the dry run would launch. It is 2 when the launch is refused (the reasons are printed) and 1 when the job fails or a watchdog destroys the instance. The instance receives no credentials of any kind: no vast.ai API key, no GCP key and no token. Before the create call, the launcher mints four V4 signed PUT URLs locally:","readmeSource":"ruvector/crates/ruvector-gpu-runner/README.md","product":"key:ruvector:gpu-runner","canonical":"ruvector-gpu-runner","tier":"T0","t":["ruvector-gpu-runner","gpu","runner","gpu-runner","budget-capped","budget","cap","auto-destroying","auto","destroy","vast.ai","job","train","jobs"],"s":["ruvector-gpu-runner","gpu","runner","gpu-runner"]},
+{"id":"ruvector-graph-condense","family":"graph-condense","kind":"crate","store":"ruvector","version":null,"description":"Structure-preserving graph condensation: collapse large feature graphs into small synthetic graphs using dynamic min-cut community boundaries","keywords":["condensation","distillation","gnn","graph","min-cut"],"source":"ruvector/crates/ruvector-graph-condense/Cargo.toml","sourceSha256":"e2641c4e27927b22b99d94920d8372f8e3a14cc42110eb0100d8602c6a08dc54","variants":["ruvector-graph-condense-wasm"],"readme":"Training-free, structure-preserving, provenance-retaining graph condensation. *Collapse a large feature graph into a small synthetic graph of super-nodes while preserving its cut structure — plus a differentiable min-cut loss.* Graph condensation shrinks a graph + per-node embeddings (+ optional labels) into a much smaller graph that downstream tasks can still reason over. The published SOTA — GCond, SFGC, GEOM, SGDD — *synthesises* a fake graph via expensive, supervised bi-level gradient/distribution/trajectory matching, and discards the node→original mapping. ruvector-graph-condense takes the complementary, training-free route the 2024–2026 condensation surveys flag as under-explored: - Mi","readmeSource":"ruvector/crates/ruvector-graph-condense/README.md","product":"key:ruvector:graph-condense","canonical":"ruvector-graph-condense","tier":"T0","t":["ruvector-graph-condense","graph","condense","graph-condense","structure-preserving","structure","preserv","condensation","collapse","large","feature","small","synthetic","dynamic","min-cut","min","cut","community","boundary","distillation","gnn"],"s":["ruvector-graph-condense","graph","condense","graph-condense","condensation","distillation","gnn","min-cut","min","cut"]},
+{"id":"ruvector-hailo","family":"hailo","kind":"crate","store":"ruvector","version":null,"description":"ruvector embedding backend for the Hailo-8 NPU (ADR-167)","keywords":["edge-ai","embedding","hailo","npu","ruvector"],"source":"ruvector/crates/ruvector-hailo/Cargo.toml","sourceSha256":"11c34bc89eebd535ca15ba2f3a291decd3b8362747110bf0bd3715c2bad1773d","variants":[],"readme":"Embedding backend for the Hailo-8 NPU on Raspberry Pi 5 + AI HAT+. Implements ruvector_core::embeddings::EmbeddingProvider so any caller holding Arc<dyn EmbeddingProvider> can swap from CPU to NPU inference with zero code changes. The cpu-fallback feature is the floor: even in NPU builds, if the HEF artifact is missing on disk, the worker falls back to candle without needing a redeploy. Same model fingerprint is reported in both modes so cluster integrity gates still work. End-to-end pipeline: text → tokenize → host embed-lookup → NPU forward → mean-pool → l2-normalize → Vec<f32> (384 dims for all-MiniLM-L6-v2). Single-Pi cognitum-v0 (Pi 5 + AI HAT+, all-MiniLM-L6-v2 HEF sha256 cdbc89...), m","readmeSource":"ruvector/crates/ruvector-hailo/README.md","product":"key:ruvector:hailo","canonical":"ruvector-hailo","tier":"T0","t":["ruvector-hailo","hailo","embedd","backend","hailo-8","npu","adr-167","adr","167","edge-ai","edge"],"s":["ruvector-hailo","hailo","edge-ai","edge","embedd","npu"]},
+{"id":"ruvector-hailo-cluster","family":"hailo-cluster","kind":"crate","store":"ruvector","version":null,"description":"Multi-Pi cluster coordinator for ruvector Hailo embedding workers (ADR-167 §8)","keywords":["cluster","edge-ai","embedding","hailo","ruvector"],"source":"ruvector/crates/ruvector-hailo-cluster/Cargo.toml","sourceSha256":"0e72a92762ab39e1b2bbe3029a3361a41a0e31e07c9c5a359fcab715aca9a3d2","variants":[],"readme":"Multi-Pi cluster coordinator for ruvector's Hailo-8 NPU embedding workers. Implements P2C+EWMA load balancing, fingerprint enforcement, optional in-process caching, and Tailscale-tag-based discovery. Three deploy paths, pick whichever matches your hardware: Expected: 67 embeds/sec/worker, 57 ms p50. Same workflow but skip the HEF download and build with cpu-fallback only: Expected: 7 embeds/sec/worker on Pi 5, ~3-4 on Pi 4. See [ADR-177][adr177] for the full Pi 4 deploy story. Expected: 45 embeds/sec/worker on AVX2 release. After systemctl start, journalctl should show: The sim_close > sim_far line is the iter-167 ranking gate — if it fails the worker exits non-zero. Test from another host:","readmeSource":"ruvector/crates/ruvector-hailo-cluster/README.md","product":"key:ruvector:hailo-cluster","canonical":"ruvector-hailo-cluster","tier":"T0","t":["ruvector-hailo-cluster","hailo","cluster","hailo-cluster","multi-pi","multi","coordinator","embedd","worker","adr-167","adr","167","edge-ai","edge"],"s":["ruvector-hailo-cluster","hailo","cluster","hailo-cluster","edge-ai","edge","embedd"]},
+{"id":"ruvector-hnsw-repair","family":"hnsw-repair","kind":"crate","store":"ruvector","version":null,"description":"Online HNSW graph repair after vector deletions — three deletion strategies with measured recall recovery","keywords":[],"source":"ruvector/crates/ruvector-hnsw-repair/Cargo.toml","sourceSha256":"9e8ad5d3a4072c748b335fb8fc3151e6bb60955c3d031cb6252079d8185f8329","variants":[],"product":"key:ruvector:hnsw-repair","canonical":"ruvector-hnsw-repair","tier":"T0","t":["ruvector-hnsw-repair","hnsw","repair","hnsw-repair","online","graph","after","vector","deletion","three","strategy","measur","recall","recovery"],"s":["ruvector-hnsw-repair","hnsw","repair","hnsw-repair"]},
+{"id":"ruvector-hybrid","family":"hybrid","kind":"crate","store":"ruvector","version":null,"description":"Hybrid sparse-dense search: BM25 + ANN + Reciprocal Rank Fusion for ruvector","keywords":["ann","bm25","hybrid-search","rrf","vector-search"],"source":"ruvector/crates/ruvector-hybrid/Cargo.toml","sourceSha256":"0ea12973b11354ab43865c8b0d4d8573af4f531e76187222807dc22098b38e75","variants":[],"product":"key:ruvector:hybrid","canonical":"ruvector-hybrid","tier":"T0","t":["ruvector-hybrid","hybrid","sparse-dense","sparse","dense","search","bm25","ann","reciprocal","rank","fusion","hybrid-search","rrf","vector-search","vector"],"s":["ruvector-hybrid","hybrid","ann","bm25","hybrid-search","search","rrf","vector-search","vector"]},
+{"id":"ruvector-hyperbolic-hnsw","family":"hyperbolic-hnsw","kind":"crate","store":"ruvector","version":null,"description":"Hyperbolic (Poincare ball) embeddings with HNSW integration for hierarchy-aware vector search, enabling efficient similarity search in non-Euclidean spaces for taxonomies, ontologies, and hierarchical data","keywords":["hierarchy","hnsw","hyperbolic","poincare","vector-search","wasm"],"source":"ruvector/crates/ruvector-hyperbolic-hnsw/Cargo.toml","sourceSha256":"2d59c58d4744f6a83acb7dd3b279b52d7a63c9919ccf0e8e4f1a291c8fa70716","variants":["ruvector-hyperbolic-hnsw-wasm"],"readme":"Hyperbolic (Poincaré ball) embeddings with HNSW integration for hierarchy-aware vector search. Hierarchies compress naturally in hyperbolic space. Taxonomies, catalogs, ICD trees, product facets, org charts, and long-tail tags all fit better than in Euclidean space, which means higher recall on deep leaves without blowing up memory or latency. - Poincaré Ball Model: Store vectors in the Poincaré ball with clamp r < 1 − eps - HNSW Speed Trick: Prune with cheap tangent-space proxy, rank with true hyperbolic distance - Per-Shard Curvature: Different parts of the hierarchy can have different optimal curvatures - Dual-Space Index: Keep a synchronized Euclidean ANN for fallback and mutual-ranking","readmeSource":"ruvector/crates/ruvector-hyperbolic-hnsw/README.md","product":"key:ruvector:hyperbolic-hnsw","canonical":"ruvector-hyperbolic-hnsw","tier":"T0","t":["ruvector-hyperbolic-hnsw","hyperbolic","hnsw","hyperbolic-hnsw","poincare","ball","embedding","integration","hierarchy-aware","hierarchy","aware","vector","search","enabl","efficient","similarity","non-euclidean","non","euclidean","space","taxonomy","ontology","hierarchical","vector-search"],"s":["ruvector-hyperbolic-hnsw","hyperbolic","hnsw","hyperbolic-hnsw","hierarchy","poincare","vector-search","vector","search"]},
+{"id":"ruvector-kalshi","family":"kalshi","kind":"crate","store":"ruvector","version":null,"description":"Kalshi exchange integration for the RuVector Neural Trader (ADR-153)","keywords":[],"source":"ruvector/crates/ruvector-kalshi/Cargo.toml","sourceSha256":"66075405e9214e11bf5b3014270b269720a40fa705178ce2d0b604dcba5aa960","variants":[],"product":"key:ruvector:kalshi","canonical":"ruvector-kalshi","tier":"T0","t":["ruvector-kalshi","kalshi","exchange","integration","neural","trader","adr-153","adr","153"],"s":["ruvector-kalshi","kalshi"]},
+{"id":"ruvector-learning-wasm","family":"learning","kind":"crate","store":"ruvector","version":null,"description":"Ultra-fast MicroLoRA adaptation for WASM - rank-2 LoRA with <100us latency for per-operator learning","keywords":["adaptation","lora","machine-learning","neural-network","wasm"],"source":"ruvector/crates/ruvector-learning-wasm/Cargo.toml","sourceSha256":"32a1afc3bc10c4ba05a807237b1feccc95254c0138294894db0398fcf1deb73b","variants":[],"readme":"Ultra-fast MicroLoRA adaptation for WASM - rank-2 LoRA with <100us latency for per-operator learning. This package provides Low-Rank Adaptation (LoRA) matrices optimized for WebAssembly execution. It enables real-time per-operator-type learning in query optimization systems with minimal latency overhead. - Rank-2 LoRA: Minimal parameter count (2d parameters per adapter) - Per-Operator Scoping: Separate adapters for 17 different operator types - <100us Adaptation: Instant weight updates for real-time learning - WASM-Optimized: Compiled to WebAssembly for near-native performance - Zero-Allocation Hot Paths: Pre-allocated buffers for performance-critical operations The main LoRA engine for sing","readmeSource":"ruvector/crates/ruvector-learning-wasm/README.md","product":"key:ruvector:learning","canonical":"ruvector-learning-wasm","tier":"T0","t":["ruvector-learning-wasm","learn","ultra-fast","ultra","microlora","adaptation","rank-2","rank","lora","100us","latency","per-operator","per","operator","machine-learning","machine","neural-network","neural","network"],"s":["ruvector-learning-wasm","learn","adaptation","lora","machine-learning","machine","neural-network","neural","network"]},
+{"id":"ruvector-lsm-ann","family":"lsm-ann","kind":"crate","store":"ruvector","version":null,"description":"LSM-ANN: write-optimized streaming vector index with multi-tier compaction for agent memory","keywords":["agent-memory","ann","lsm-tree","streaming","vector-search"],"source":"ruvector/crates/ruvector-lsm-ann/Cargo.toml","sourceSha256":"93aa554f69f14e2cc60486c6796e2c37dfc9ac14cd9bc78ac49b9f020e37606d","variants":[],"readme":"Write-optimised streaming vector index with multi-tier LSM compaction for agent memory. ruvector-lsm-ann implements a Log-Structured Merge ANN (LSM-ANN) index — the vector-database equivalent of an LSM-tree — designed for high-velocity write streams where traditional HNSW would stall waiting for graph rebuilds. Queries fan out across all tiers and merge-rank by exact distance. Run cargo run --release -p ruvector-lsm-ann --bin benchmark for live numbers. MIT — part of the [RuVector](https://github.com/ruvnet/ruvector) project.","readmeSource":"ruvector/crates/ruvector-lsm-ann/README.md","product":"key:ruvector:lsm-ann","canonical":"ruvector-lsm-ann","tier":"T0","t":["ruvector-lsm-ann","lsm","ann","lsm-ann","write-optimized","write","stream","vector","index","multi-tier","multi","tier","compaction","memory","agent-memory","lsm-tree","tree","vector-search","search"],"s":["ruvector-lsm-ann","lsm","ann","lsm-ann","agent-memory","memory","lsm-tree","tree","stream","vector-search","vector","search"]},
+{"id":"ruvector-math","family":"math","kind":"crate","store":"ruvector","version":null,"description":"Advanced mathematics for next-gen vector search: Optimal Transport, Information Geometry, Product Manifolds","keywords":["hyperbolic","information-geometry","optimal-transport","vector-search","wasm","wasserstein"],"source":"ruvector/crates/ruvector-math/Cargo.toml","sourceSha256":"3b617c02c9cda09f799ab447c2e30d0d0373729cfcf5eef9ad0af509b5661132","variants":["ruvector-math-wasm"],"readme":"Advanced Mathematics for Next-Generation Vector Search ruvector-math brings advanced mathematical tools to vector search and AI systems. Think of it as a Swiss Army knife for working with high-dimensional data, embeddings, and neural networks. All modules in this library connect through a single unifying concept: mincut (minimum cut). Mincut measures how \"connected\" a graph is - specifically, how much you'd need to cut to separate it into parts. In AI systems, mincut tells us: - Low mincut (near 0): The system is stable - use fast, simple processing - High mincut: The system is changing - be cautious, use more careful methods - Very high mincut: Major shifts detected - pause and re-evaluate","readmeSource":"ruvector/crates/ruvector-math/README.md","product":"key:ruvector:math","canonical":"ruvector-math","tier":"T0","t":["ruvector-math","math","advanc","mathematic","next-gen","next","gen","vector","search","optimal","transport","information","geometry","product","manifold","hyperbolic","information-geometry","optimal-transport","vector-search","wasserstein"],"s":["ruvector-math","math","hyperbolic","information-geometry","information","geometry","optimal-transport","optimal","transport","vector-search","vector","search","wasserstein"]},
+{"id":"ruvector-matryoshka","family":"matryoshka","kind":"crate","store":"ruvector","version":null,"description":"Matryoshka-aware coarse-to-fine vector search: three-variant funnel ANN with measured recall and latency tradeoffs","keywords":["agent-memory","ann","coarse-to-fine","matryoshka","vector-search"],"source":"ruvector/crates/ruvector-matryoshka/Cargo.toml","sourceSha256":"2d3df453799e44524d64a1da4bf7ab7eaf2b40958d9e9428fbf134c7517952a7","variants":[],"readme":"Matryoshka-aware coarse-to-fine vector search: adaptive funnel ANN with measured recall and latency tradeoffs. All major 2026 embedding models (OpenAI text-embedding-3, Nomic nomic-embed-text-v2, Voyage 4, Cohere v4, Jina v5) use Matryoshka Representation Learning (MRL): any prefix of the full-dimension vector is a valid, lower-dimensional embedding. ruvector-matryoshka exploits this property for a coarse-to-fine search funnel: This gives 3–8× faster search with minimal recall loss versus full-dim search. Run cargo run --release -p ruvector-matryoshka --bin benchmark for live numbers. Any model trained with MRL or returning truncatable embeddings: - OpenAI text-embedding-3-small / text-embed","readmeSource":"ruvector/crates/ruvector-matryoshka/README.md","product":"key:ruvector:matryoshka","canonical":"ruvector-matryoshka","tier":"T0","t":["ruvector-matryoshka","matryoshka","matryoshka-aware","aware","coarse-to-fine","coarse","fine","vector","search","three-variant","three","variant","funnel","ann","measur","recall","latency","tradeoff","agent-memory","memory","vector-search"],"s":["ruvector-matryoshka","matryoshka","agent-memory","memory","ann","coarse-to-fine","coarse","fine","vector-search","vector","search"]},
+{"id":"ruvector-maxsim","family":"maxsim","kind":"crate","store":"ruvector","version":null,"description":"ColBERT-style multi-vector MaxSim late interaction search for ruvector: higher recall for multi-facet documents without single-embedding averaging loss","keywords":[],"source":"ruvector/crates/ruvector-maxsim/Cargo.toml","sourceSha256":"fa6a78dfe9b11bd69a3a03a67777e36ebbf1db68ed5928a031b97b66d2e208ef","variants":[],"product":"key:ruvector:maxsim","canonical":"ruvector-maxsim","tier":"T0","t":["ruvector-maxsim","maxsim","colbert-style","colbert","style","multi-vector","multi","vector","late","interaction","search","higher","recall","multi-facet","facet","document","single-embedding","single","embedd","averag","loss"],"s":["ruvector-maxsim","maxsim"]},
+{"id":"ruvector-memory-admission","family":"memory-admission","kind":"crate","store":"ruvector","version":null,"description":"Global min-cut gated write-time cluster admission for streaming agent memory in RuVector","keywords":["agent-memory","clustering","mincut","streaming","vector-search"],"source":"ruvector/crates/ruvector-memory-admission/Cargo.toml","sourceSha256":"eb9951bede5ca71a41b4bcef00ef62854203a9c51dd4c69634123de874ccc5ae","variants":[],"readme":"Global-min-cut gated write-time cluster admission for streaming agent memory. ruvector-namespace-merge (ADR-299) uses S-T max-flow/min-cut to answer a read-time question: given a query and a fixed namespace set, which namespaces should be searched? This crate answers the dual write-time question, which has no natural source/sink to fix the flow terminals on: given a stream of incoming agent-memory vectors and a growing set of clusters, should the next vector merge into an existing cluster, or does it belong to a new one? A fixed cosine-to-nearest-centroid threshold (the obvious baseline) only looks at one edge. The global minimum cut (Stoer-Wagner, no source/sink) of the (existing centroids","readmeSource":"ruvector/crates/ruvector-memory-admission/README.md","product":"key:ruvector:memory-admission","canonical":"ruvector-memory-admission","tier":"T0","t":["ruvector-memory-admission","memory","admission","memory-admission","global","min-cut","min","cut","gated","write-time","write","cluster","stream","agent-memory","mincut","vector-search","vector","search"],"s":["ruvector-memory-admission","memory","admission","memory-admission","agent-memory","cluster","mincut","stream","vector-search","vector","search"]},
+{"id":"ruvector-metrics","family":"metrics","kind":"crate","store":"ruvector","version":null,"description":"Prometheus-compatible metrics collection for Ruvector vector databases","keywords":[],"source":"ruvector/crates/ruvector-metrics/Cargo.toml","sourceSha256":"a6e524e72e17c894f1741145ac17c21d51032bfd7cc5b0795d610759a803e067","variants":[],"readme":"Prometheus-compatible metrics collection for Ruvector vector databases. ruvector-metrics provides comprehensive observability with counters, gauges, histograms, and exporters for monitoring Ruvector performance and health. Part of the [Ruvector](https://github.com/ruvnet/ruvector) ecosystem. - Prometheus Native: Direct Prometheus integration - Zero Overhead: Lazy initialization, minimal impact - Comprehensive: Operation latencies, throughput, memory - Customizable: Add custom metrics for your use case - Standard Format: OpenMetrics-compatible output - Operation Counters: Insert, search, delete counts - Latency Histograms: p50, p95, p99 latencies - Throughput Gauges: Queries per second - Memo","readmeSource":"ruvector/crates/ruvector-metrics/README.md","product":"key:ruvector:metrics","canonical":"ruvector-metrics","tier":"T0","t":["ruvector-metrics","metric","prometheus-compatible","prometheus","compatible","collection","vector","databas"],"s":["ruvector-metrics","metric"]},
+{"id":"ruvector-mincut-brain-node","family":"mincut-brain","kind":"crate","store":"ruvector","version":null,"description":"Minimal WASM binary for pi.ruv.io brain node: canonical min-cut with V1 ABI stubs","keywords":[],"source":"ruvector/crates/ruvector-mincut-brain-node/Cargo.toml","sourceSha256":"c804c4be9a73aa335d8f539609e99938857792a25226c2cb6b89313cc57aaa60","variants":[],"product":"key:ruvector:mincut-brain","canonical":"ruvector-mincut-brain-node","tier":"T0","t":["ruvector-mincut-brain-node","mincut","brain","mincut-brain","minimal","binary","pi.ruv.io","canonical","min-cut","min","cut","abi","stub"],"s":["ruvector-mincut-brain-node","mincut","brain","mincut-brain"]},
+{"id":"ruvector-mincut-gated-transformer","family":"mincut-gated-transformer","kind":"crate","store":"ruvector","version":null,"description":"Ultra low latency transformer inference with mincut-gated coherence control","keywords":["coherence","inference","low-latency","mincut","transformer","wasm","webassembly"],"source":"ruvector/crates/ruvector-mincut-gated-transformer/Cargo.toml","sourceSha256":"48bad3714ed2a5322f5b389dd27e623c4436dae90b4b7a58743d70179ae19226","variants":["ruvector-mincut-gated-transformer-wasm"],"readme":"The Mincut-Gated Transformer is a production-grade inference engine that combines minimum cut (mincut) graph partitioning with adaptive compute allocation to achieve deterministic, ultra-low latency inference. Unlike traditional transformers that execute all layers uniformly, this architecture uses graph-theoretic coherence signals to dynamically skip computation, exit early, and control state updates—all while maintaining explainability and safety guarantees. Why Mincut? The minimum cut value (λ) of an attention graph provides a principled measure of information flow coherence. When λ is high and stable, the model can safely reduce computation. When λ drops or becomes unstable, the system c","readmeSource":"ruvector/crates/ruvector-mincut-gated-transformer/README.md","product":"key:ruvector:mincut-gated-transformer","canonical":"ruvector-mincut-gated-transformer","tier":"T0","t":["ruvector-mincut-gated-transformer","mincut","gated","transformer","mincut-gated-transformer","ultra","low","latency","inference","mincut-gated","coherence","control","low-latency"],"s":["ruvector-mincut-gated-transformer","mincut","gated","transformer","mincut-gated-transformer","coherence","inference","low-latency","low","latency"]},
+{"id":"ruvector-mmwave","family":"mmwave","kind":"crate","store":"ruvector","version":null,"description":"Shared parser for Seeed MR60BHA2 + HLK-LD2410 60/24 GHz mmWave radar UART streams (ADR-063)","keywords":["edge-ai","mmwave","presence","radar","ruvector"],"source":"ruvector/crates/ruvector-mmwave/Cargo.toml","sourceSha256":"12a082b5ecbedb5d4a8f079bea34a887dab2e6d44defa2acc404669109b0df6f","variants":[],"readme":"Streaming parser for the Seeed MR60BHA2 60 GHz radar's UART protocol. Pure-Rust, no_std-compatible, zero-allocation hot path. ADR-178 Gap H: keeping the parser separate from ruvector-hailo-cluster (the bridge's home crate) means a regression in either side surfaces independently. The parser is byte-for-byte deterministic against fuzzed inputs; the bridge layers transport, TLS, fingerprinting on top. Frame types currently parsed: The closure signature is FnMut(Event); the parser invokes it zero-or-more times per byte fed. State machine resyncs cleanly on checksum failure or unexpected SOF — no manual reset needed. Run cargo bench -p ruvector-mmwave for the full criterion sweep. Steady-state o","readmeSource":"ruvector/crates/ruvector-mmwave/README.md","product":"key:ruvector:mmwave","canonical":"ruvector-mmwave","tier":"T0","t":["ruvector-mmwave","mmwave","shar","parser","seeed","mr60bha2","hlk-ld2410","hlk","ld2410","ghz","radar","uart","stream","adr-063","adr","063","edge-ai","edge","presence"],"s":["ruvector-mmwave","mmwave","edge-ai","edge","presence","radar"]},
+{"id":"ruvector-namespace-merge","family":"namespace-merge","kind":"crate","store":"ruvector","version":null,"description":"S-T mincut namespace routing for multi-namespace agent memory vector search in RuVector","keywords":["agent-memory","ann","mincut","namespace-routing","vector-search"],"source":"ruvector/crates/ruvector-namespace-merge/Cargo.toml","sourceSha256":"9d920efb5bfdd806fdaaa37fa81c4d1d272d392cf849c2f2f75c39559de34bf9","variants":[],"readme":"Experimental routing for vector search across multiple namespaces. The crate compares exhaustive search, centroid filtering, and an S-T min-cut router that balances query relevance with inter-namespace cohesion. Routing can omit relevant namespaces. Evaluate recall against exhaustive search on representative data before deploying a selective strategy.","readmeSource":"ruvector/crates/ruvector-namespace-merge/README.md","product":"key:ruvector:namespace-merge","canonical":"ruvector-namespace-merge","tier":"T0","t":["ruvector-namespace-merge","namespace","merge","namespace-merge","s-t","mincut","rout","multi-namespace","multi","memory","vector","search","agent-memory","ann","namespace-routing","vector-search"],"s":["ruvector-namespace-merge","namespace","merge","namespace-merge","agent-memory","memory","ann","mincut","namespace-routing","rout","vector-search","vector","search"]},
+{"id":"ruvector-nervous-system","family":"nervous-system","kind":"crate","store":"ruvector","version":null,"description":"Bio-inspired neural system with spiking networks, BTSP learning, and EWC plasticity","keywords":["btsp","hdc","learning","neural","neuromorphic","plasticity","spiking","wasm"],"source":"ruvector/crates/ruvector-nervous-system/Cargo.toml","sourceSha256":"1e70828e40f7b56b9c5dc175254f575c51e7bd64e4dbfdb51c564bfd365e1c56","variants":["ruvector-nervous-system-wasm"],"readme":"A five-layer bio-inspired nervous system for AI applications. Think less \"smart algorithm\" and more \"living organism.\" Most AI systems are like assembly lines: data goes in, predictions come out, repeat forever. This crate takes a different approach. It gives your software a *nervous system* - the same kind of layered architecture that lets living creatures sense danger, react instantly, learn from experience, and rest when they need to. The result? Systems that: - React in microseconds instead of waiting for batch processing - Learn from single examples instead of retraining on millions - Stay quiet when nothing changes instead of burning compute continuously - Know when they're struggling","readmeSource":"ruvector/crates/ruvector-nervous-system/README.md","product":"key:ruvector:nervous-system","canonical":"ruvector-nervous-system","tier":"T0","t":["ruvector-nervous-system","nervous","nervous-system","bio-inspired","bio","inspir","neural","spik","network","btsp","learn","ewc","plasticity","hdc","neuromorphic"],"s":["ruvector-nervous-system","nervous","nervous-system","btsp","hdc","learn","neural","neuromorphic","plasticity","spik"]},
+{"id":"ruvector-perception","family":"perception","kind":"crate","store":"ruvector","version":null,"description":"The layer under classification: physical delta -> boundary -> coherence -> proof -> action. A trusted-physical-memory engine that emits structured delta witnesses, not class labels.","keywords":["anomaly","coherence","edge-ai","min-cut","sensing"],"source":"ruvector/crates/ruvector-perception/Cargo.toml","sourceSha256":"c12cd237886dcddc7229e59795e138c66bcc255d4cba20fca0dba4e6c41b0e31","variants":[],"product":"key:ruvector:perception","canonical":"ruvector-perception","tier":"T0","t":["ruvector-perception","perception","layer","classification","physical","delta","boundary","coherence","proof","action","trusted-physical-memory","trust","memory","emit","structur","witness","class","label","anomaly","edge-ai","edge","min-cut","min","cut","sens"],"s":["ruvector-perception","perception","anomaly","coherence","edge-ai","edge","min-cut","min","cut","sens"]},
+{"id":"ruvector-pq-search","family":"pq-search","kind":"crate","store":"ruvector","version":null,"description":"Product Quantization with Asymmetric Distance Computation (ADC) for compressed approximate nearest-neighbor search — Flat PQ, IVF+PQ, and Residual-corrected PQ variants in safe Rust","keywords":["ann","approximate-search","compression","product-quantization","vector-search"],"source":"ruvector/crates/ruvector-pq-search/Cargo.toml","sourceSha256":"f00f9e07b841948fe5c2fa731c3025414c545986c8e71f2ba393048c7b283313","variants":[],"readme":"Product Quantization with Asymmetric Distance Computation (ADC) for compressed approximate nearest-neighbor search — 64× compression in safe Rust. Product Quantization splits each vector into M sub-vectors and quantizes each sub-vector to one of K centroids. A 128-dim f32 vector (512 bytes) can be stored as 16 bytes — 64× compression. Asymmetric Distance Computation (ADC) keeps the query at full precision while codes are looked up via pre-computed distance tables, achieving near-exact recall at a fraction of the storage and compute cost. - Memory-constrained edge deployment (IoT, mobile, Pi 5) - Large-scale agent memory where 64× storage reduction enables more history - Tiered retrieval: PQ-","readmeSource":"ruvector/crates/ruvector-pq-search/README.md","product":"key:ruvector:pq-search","canonical":"ruvector-pq-search","tier":"T0","t":["ruvector-pq-search","search","pq-search","product","quantization","asymmetric","distance","computation","adc","compres","approximate","nearest-neighbor","nearest","neighbor","flat","ivf+pq","residual-corrected","residual","correct","variant","safe","ann","approximate-search","compression","product-quantization","vector-search","vector"],"s":["ruvector-pq-search","search","pq-search","ann","approximate-search","approximate","compression","product-quantization","product","quantization","vector-search","vector"]},
+{"id":"ruvector-profiler","family":"profiler","kind":"crate","store":"ruvector","version":null,"description":"Memory, power, and latency profiling hooks with CSV emitters for benchmarking attention mechanisms","keywords":["attention","benchmarking","latency","memory","profiling"],"source":"ruvector/crates/ruvector-profiler/Cargo.toml","sourceSha256":"b518580ba8931477e61084d2bb6570bfb978a25797cdbc47ef85f0520c1f94cb","variants":[],"readme":"Memory, power, and latency profiling hooks with CSV emitters — the observability layer for attention benchmarking. This crate instruments benchmark runs with three profiling dimensions -- memory pressure, energy consumption, and latency distribution -- and exports results to CSV files for downstream analysis. It is the observability layer in the ruvector attention benchmarking pipeline, sitting between the attention operators (ruvector-attn-mincut) and the analysis/plotting stage. Every benchmark run is tagged with a SHA-256 config fingerprint so that results are reproducible and auditable across machines. MemoryTracker captures RSS snapshots via /proc/self/status on Linux (zero fallback on","readmeSource":"ruvector/crates/ruvector-profiler/README.md","product":"key:ruvector:profiler","canonical":"ruvector-profiler","tier":"T0","t":["ruvector-profiler","profiler","memory","power","latency","profil","hook","csv","emitter","benchmark","attention","mechanism"],"s":["ruvector-profiler","profiler","attention","benchmark","latency","memory","profil"]},
+{"id":"ruvector-proof-gate","family":"proof-gate","kind":"crate","store":"ruvector","version":null,"description":"Proof-gated vector writes with Merkle-accumulating witness logs for RuVector. Tamper-evident write receipts for agent memory substrates.","keywords":["agent-memory","merkle","proof","vector-database","witness"],"source":"ruvector/crates/ruvector-proof-gate/Cargo.toml","sourceSha256":"49089ab24db432c1f27c45689ace2989efce17ff17a89d0d660493967fefe758","variants":[],"readme":"Tamper-evident vector writes — a Merkle / hash-chain write-ahead log that gives every write to a vector store cryptographic evidence of *what* was stored, *when*, and *by whom*. Part of the [ruvector](https://github.com/ruvnet/ruvector) ecosystem. Admit a WritePayload, get a WriteReceipt you can store alongside the index entry and verify offline. The gate's chain root is a cryptographic commitment to the entire ordered write log: any mutation, insertion, deletion, or reorder of writes produces a different root. Measured (DIM=128): HashChainGate.admit ≈ 700 ns/write (~1.4 M/s), verify_receipt ≈ 6 ns (157 M/s). The integrity tax is ~675 ns/write — negligible next to the HNSW insertion a real w","readmeSource":"ruvector/crates/ruvector-proof-gate/README.md","product":"key:ruvector:proof-gate","canonical":"ruvector-proof-gate","tier":"T0","t":["ruvector-proof-gate","proof","gate","proof-gate","proof-gated","gated","vector","write","merkle-accumulating","merkle","accumulat","witness","logs","tamper-evident","tamper","evident","receipt","memory","substrate","agent-memory","vector-database","database"],"s":["ruvector-proof-gate","proof","gate","proof-gate","agent-memory","memory","merkle","vector-database","vector","database","witness"]},
+{"id":"ruvector-query-cache","family":"query-cache","kind":"crate","store":"ruvector","version":null,"description":"Semantic query cache for RuVector ANN: exact-hash, cosine-similarity, and adaptive-threshold caching for agent-memory workloads","keywords":["agent-memory","ann","cache","semantic","vector-search"],"source":"ruvector/crates/ruvector-query-cache/Cargo.toml","sourceSha256":"be8b0ef3248fc91310d7929434072a242d1a5101439a62a031acf510cf7279f6","variants":[],"readme":"Experimental query-result caching for vector search. The crate compares an uncached brute-force baseline, exact-query caching, and cosine-based semantic caching with bounded capacity. Semantic hits deliberately trade exact result fidelity for latency. Select a threshold using workload-specific recall measurements before production use.","readmeSource":"ruvector/crates/ruvector-query-cache/README.md","product":"key:ruvector:query-cache","canonical":"ruvector-query-cache","tier":"T0","t":["ruvector-query-cache","query","cache","query-cache","semantic","ann","exact-hash","exact","hash","cosine-similarity","cosine","similarity","adaptive-threshold","adaptive","threshold","cach","agent-memory","memory","workload","vector-search","vector","search"],"s":["ruvector-query-cache","query","cache","query-cache","agent-memory","memory","ann","semantic","vector-search","vector","search"]},
+{"id":"ruvector-rabitq","family":"rabitq","kind":"crate","store":"ruvector","version":null,"description":"RaBitQ: rotation-based 1-bit quantization for ultra-fast approximate nearest-neighbor search with theoretical error bounds","keywords":["embeddings","quantization","rabitq","vector-search","wasm"],"source":"ruvector/crates/ruvector-rabitq/Cargo.toml","sourceSha256":"8aeb99ae9ae4426cb3d30a4ec422253e877cd590caffdefaf2479f8968250762","variants":["ruvector-rabitq-wasm"],"product":"key:ruvector:rabitq","canonical":"ruvector-rabitq","tier":"T0","t":["ruvector-rabitq","rabitq","rotation-based","rotation","1-bit","bit","quantization","ultra-fast","ultra","approximate","nearest-neighbor","nearest","neighbor","search","theoretical","error","bound","embedding","vector-search","vector"],"s":["ruvector-rabitq","rabitq","embedding","quantization","vector-search","vector","search"]},
+{"id":"ruvector-rairs","family":"rairs","kind":"crate","store":"ruvector","version":null,"description":"RAIRS IVF: Redundant Assignment with Amplified Inverse Residual — ruvector's first IVF index family","keywords":["ann","ivf","nearest-neighbor","ruvector","vector-search"],"source":"ruvector/crates/ruvector-rairs/Cargo.toml","sourceSha256":"f4c71441e4477173ee096b23d0334c67b6ae95a18d4537986fab7886eecfec2c","variants":[],"product":"key:ruvector:rairs","canonical":"ruvector-rairs","tier":"T0","t":["ruvector-rairs","rair","ivf","redundant","assignment","amplify","inverse","residual","index","family","ann","nearest-neighbor","nearest","neighbor","vector-search","vector","search"],"s":["ruvector-rairs","rair","ann","ivf","nearest-neighbor","nearest","neighbor","vector-search","vector","search"]},
+{"id":"ruvector-recall-bounded","family":"recall-bounded","kind":"crate","store":"ruvector","version":null,"description":"Recall-bounded approximate similarity search for RuVector - threshold-driven retrieval with three measured variants and honest quality accounting","keywords":["agent-memory","ann","hnsw","threshold-search","vector-search"],"source":"ruvector/crates/ruvector-recall-bounded/Cargo.toml","sourceSha256":"821ff497736cd1554303b6c64093c4b29dac46645d509967aa770e9b8a6d6868","variants":[],"readme":"Research implementations of threshold-driven similarity search for RuVector. The crate compares an exact linear scan with two bounded graph-search heuristics and reports empirical recall against the exact baseline. The approximate variants do not provide a formal recall guarantee. Their expansion parameters must be calibrated and audited on representative data. See docs/adr/ADR-272-recall-bounded-ann.md and the associated nightly report for methodology and limitations.","readmeSource":"ruvector/crates/ruvector-recall-bounded/README.md","product":"key:ruvector:recall-bounded","canonical":"ruvector-recall-bounded","tier":"T0","t":["ruvector-recall-bounded","recall","bound","recall-bounded","approximate","similarity","search","threshold-driven","threshold","driven","retrieval","three","measur","variant","honest","quality","account","agent-memory","memory","ann","hnsw","threshold-search","vector-search","vector"],"s":["ruvector-recall-bounded","recall","bound","recall-bounded","agent-memory","memory","ann","hnsw","threshold-search","threshold","search","vector-search","vector"]},
+{"id":"ruvector-retrieval-receipt","family":"retrieval-receipt","kind":"crate","store":"ruvector","version":null,"description":"Witness-chained provenance receipts for ANN retrieval results, binding query evidence to ruvector-proof-gate write history.","keywords":["agent-memory","merkle","provenance","rag","vector-database"],"source":"ruvector/crates/ruvector-retrieval-receipt/Cargo.toml","sourceSha256":"4ad1a9c70d06e5e05f18fe2afc3e07e7e7824dd43d1e515ee0fa8c822723bb96","variants":[],"readme":"Witness-chained provenance for ANN retrieval results — cryptographic receipts that commit a query's top-k results (together with copies of each vector's ingestion WriteReceipt) so that a receipt/result pair, once issued, cannot be silently mutated in transit or in storage. Part of the [ruvector](https://github.com/ruvnet/ruvector) ecosystem. Search a RetrievalIndex (a brute-force cosine index whose ingestion path is a real ruvector_proof_gate::HashChainGate), wrap the result set in a RetrievalReceipt, and a later holder of the receipt can check — offline, without talking to the query engine — that the results they hold are the ones the engine committed to at query time. Threat model, stated","readmeSource":"ruvector/crates/ruvector-retrieval-receipt/README.md","product":"key:ruvector:retrieval-receipt","canonical":"ruvector-retrieval-receipt","tier":"T0","t":["ruvector-retrieval-receipt","retrieval","receipt","retrieval-receipt","witness-chained","witness","chain","provenance","ann","result","query","evidence","ruvector-proof-gate","proof","gate","write","history","agent-memory","memory","merkle","rag","vector-database","vector","database"],"s":["ruvector-retrieval-receipt","retrieval","receipt","retrieval-receipt","agent-memory","memory","merkle","provenance","rag","vector-database","vector","database"]},
+{"id":"ruvector-robotics","family":"robotics","kind":"crate","store":"ruvector","version":null,"description":"Cognitive robotics platform: bridge types, perception pipeline, cognitive architecture, and MCP tools","keywords":["behavior-tree","cognitive","perception","robotics","spatial-indexing"],"source":"ruvector/crates/ruvector-robotics/Cargo.toml","sourceSha256":"e51f5e06db8321ea077ce9927a14dd892350b8518f85ece6ef9edefd245d57cf","variants":[],"readme":"Unified cognitive robotics platform built on ruvector's vector database, graph neural networks, and self-learning infrastructure. - Types: Point3D, PointCloud, RobotState, Pose, Quaternion, SensorFrame, OccupancyGrid, SceneObject, SceneGraph, Trajectory - SpatialIndex: Brute-force kNN and radius search with Euclidean/Cosine/Manhattan metrics - Converters: Bidirectional conversion between robotics messages and flat vectors - Pipeline: Lightweight perception pipeline with obstacle detection and trajectory prediction - SceneGraphBuilder: Spatial hash clustering with union-find for point cloud segmentation - ObstacleDetector: Grid-based obstacle detection with heuristic classification (Static/Dy","readmeSource":"ruvector/crates/ruvector-robotics/README.md","product":"key:ruvector:robotics","canonical":"ruvector-robotics","tier":"T0","t":["ruvector-robotics","robotic","cognitive","bridge","type","perception","pipeline","architecture","mcp","behavior-tree","behavior","tree","spatial-indexing","spatial","index"],"s":["ruvector-robotics","robotic","behavior-tree","behavior","tree","cognitive","perception","spatial-indexing","spatial","index"]},
+{"id":"ruvector-rulake","family":"rulake","kind":"crate","store":"ruvector","version":null,"description":"ruLake — vector-native federation intermediary over heterogeneous backends (ADR-155)","keywords":[],"source":"ruvector/crates/ruvector-rulake/Cargo.toml","sourceSha256":"d383903998e592ef5911609ca5fc3529935efe8c76ccfdf891a796b521435697","variants":[],"readme":"A cache layer for vector search — sits in front of whatever database, lakehouse, or file store already holds your vectors, and makes every query fast. You already have vectors somewhere: Parquet files on S3, rows in BigQuery, an Iceberg table, a Snowflake column, or .bin files on a disk. You want semantic search that's fast, consistent, and cheap. ruLake is the piece in the middle. - Your app asks ruLake for the nearest K vectors. - ruLake serves hits from a compressed in-memory cache at ~1 % over raw library speed (measured 1.00–1.02× intermediary tax). - On a cache miss, ruLake pulls from your backend, compresses with [RaBitQ](../ruvector-rabitq/) 1-bit quantization, and serves the result.","readmeSource":"ruvector/crates/ruvector-rulake/README.md","product":"repo:rulake","canonical":"rulake","tier":"T0","t":["ruvector-rulake","rulake","vector-native","vector","federation","intermediary","heterogeneous","backend","adr-155","adr","155"],"s":["ruvector-rulake","rulake"]},
+{"id":"ruvector-rulake-gcs","family":"rulake-gcs","kind":"crate","store":"rulake","version":null,"description":"GCS Parquet BackendAdapter for ruLake (ADR-155 §M2)","keywords":[],"source":"rulake/crates/gcs-backend/Cargo.toml","sourceSha256":"b2e59a39147cff56bd413d0b59e284a77f6cd1e43916085fb8975a28b93c954e","variants":[],"readme":"Implements the first cloud backend per [ADR-155 §M2](../docs/adrs/ADR-155-rulake-datalake-layer.md): read vectors from Parquet files on Google Cloud Storage, with cache coherence riding GCS's per-object generation token. Sibling Cargo package per [ADR-001](../docs/adrs/ADR-001-standalone-repo-strategy.md). - GcsParquetBackend — BackendAdapter trait impl over GCS+Parquet. - Cache coherence via GCS object generation — every reupload bumps the generation; ruLake's witness picks up the change automatically through the existing Generation::Num variant. - Cheap current_bundle() override — does a HEAD on the GCS object + a Parquet-footer-only schema read instead of the default impl's full pull_vect","readmeSource":"rulake/crates/gcs-backend/README.md","product":"key:rulake:rulake-gcs","canonical":"ruvector-rulake-gcs","tier":"T1","t":["ruvector-rulake-gcs","rulake","gcs","rulake-gcs","parquet","backendadapter","adr-155","adr","155"],"s":["ruvector-rulake-gcs","rulake","gcs","rulake-gcs"]},
+{"id":"ruvector-rulake-ipfs","family":"rulake-ipfs","kind":"crate","store":"rulake","version":null,"description":"IPFS BackendAdapter for ruLake — bundle distribution + content addressing (ADR-005)","keywords":[],"source":"rulake/crates/ipfs-backend/Cargo.toml","sourceSha256":"c59e9fa02c1f3a43a48c0a165ad2385d66560c469d4e219e9164de76e4e2f322","variants":[],"readme":"Implements [ADR-005](../docs/adrs/sdk/ADR-005-ipfs-backend-and-deploy.md): a BackendAdapter that publishes and resolves ruLake bundles (table.rulake.json) over IPFS, addressed by CIDv1. This is the bundle distribution layer, not the bulk vector-data layer. Vector bodies stay on the original backend ([gcs-backend/](../gcs-backend/), FsBackend, etc.); IPFS handles the cross-process witness anchor. Sibling Cargo package per [ADR-001](../docs/adrs/ADR-001-standalone-repo-strategy.md). - IpfsBackend — BackendAdapter trait impl over kubo HTTP RPC. - Three operating modes (ADR-005 §2): - Mode::Kubo — talks to a kubo daemon (default http://127.0.0.1:5001); read + write + pin. - Mode::Gateway — read-","readmeSource":"rulake/crates/ipfs-backend/README.md","product":"key:rulake:rulake-ipfs","canonical":"ruvector-rulake-ipfs","tier":"T1","t":["ruvector-rulake-ipfs","rulake","ipfs","rulake-ipfs","backendadapter","bundle","distribution","content","address","adr-005","adr","005"],"s":["ruvector-rulake-ipfs","rulake","ipfs","rulake-ipfs"]},
+{"id":"ruvector-rulake-kernel-avx512","family":"rulake-kernel-avx512","kind":"crate","store":"rulake","version":null,"description":"AVX-512 VectorKernel impl for ruLake (ADR-157)","keywords":[],"source":"rulake/crates/kernel-avx512/Cargo.toml","sourceSha256":"6e5296a81c74e668569579bd807a193f20ec25e36e4f05d1dbb6d122d9344051","variants":[],"readme":"AVX-512 implementation of the ADR-157 VectorKernel trait for ruLake. This crate is standalone (no workspace) per ADR-001. It depends on the root rulake crate by relative path (path = \"..\"). Provides an Avx512Kernel that accelerates: - L2² distance top-K — _mm512_sub_ps + _mm512_fmadd_ps per 16-lane chunk, horizontal reduce on the tail. - rabitq popcount Hamming top-K — _mm512_popcnt_epi64 (the avx512vpopcntdq extension) over 8-lane u64 chunks, scalar tail. Both paths are bit-equal to rulake::kernel::CpuNaiveKernel on the ADR-157 conformance fixture (assert_kernel_conformant); promotion past experimental is gated on that. Avx512Kernel::new() returns Option<Self>: - Some iff the host CPU adver","readmeSource":"rulake/crates/kernel-avx512/README.md","product":"key:rulake:rulake-kernel-avx512","canonical":"ruvector-rulake-kernel-avx512","tier":"T1","t":["ruvector-rulake-kernel-avx512","rulake","kernel","avx512","rulake-kernel-avx512","avx-512","avx","512","vectorkernel","impl","adr-157","adr","157"],"s":["ruvector-rulake-kernel-avx512","rulake","kernel","avx512","rulake-kernel-avx512"]},
+{"id":"ruvector-rulake-kernel-wgpu","family":"rulake-kernel-wgpu","kind":"crate","store":"rulake","version":null,"description":"Portable GPU VectorKernel impl for ruLake via wgpu (ADR-157)","keywords":[],"source":"rulake/crates/kernel-wgpu/Cargo.toml","sourceSha256":"b1e95a9e2d7254ff2c444e4e90fe679a767e2bf6d551b11fa058cfdf06feca9f","variants":[],"readme":"Portable GPU implementation of the ADR-157 VectorKernel trait for ruLake, via [wgpu](https://wgpu.rs). This crate is standalone (no workspace) per ADR-001. It depends on the root rulake crate by relative path (path = \"..\"). Provides a WgpuKernel that runs the two hot inner loops on whatever compute-capable backend wgpu finds (Vulkan / Metal / DX12 / GL / WebGPU): - L2² distance top-K — one workgroup over the candidate batch, one thread per (query, candidate) distance, results scored on host. - rabitq popcount Hamming top-K — packed 1-bit codes XORed and popcounted on-GPU per u32 lane (WGSL countOneBits). Both shaders compute only the per-candidate distance on GPU; the top-K sort runs on the","readmeSource":"rulake/crates/kernel-wgpu/README.md","product":"key:rulake:rulake-kernel-wgpu","canonical":"ruvector-rulake-kernel-wgpu","tier":"T1","t":["ruvector-rulake-kernel-wgpu","rulake","kernel","wgpu","rulake-kernel-wgpu","portable","gpu","vectorkernel","impl","via","adr-157","adr","157"],"s":["ruvector-rulake-kernel-wgpu","rulake","kernel","wgpu","rulake-kernel-wgpu"]},
+{"id":"ruvector-rulake-mcp","family":"rulake-mcp","kind":"crate","store":"rulake","version":null,"description":"MCP server for ruLake — agent-callable, witness-aware vector cache (ADR-004)","keywords":[],"source":"rulake/crates/mcp-server/Cargo.toml","sourceSha256":"acc5ef5d4430f68ebc2e0ea15f56e9cd0de9ab24e5a9eb5326fc5cb5b5b2c5bf","variants":[],"readme":"Implements [ADR-004](../docs/adrs/sdk/ADR-004-rulake-mcp-server.md) — the control-plane / decision-layer over [rulake](..) for agent-callable governed memory. v0.1 landed the architecture, v0.2 added remote-agent deployability (Streamable HTTP + bearer), v0.3 ships the rest of the ADR-004 §4 + §5 + §7 surface: mutation tools behind capability tiers, all four intents (search/verify/explain/refresh), the §7 JSONL audit file, and rulake://stats resources. Per [ADR-001](../docs/adrs/ADR-001-standalone-repo-strategy.md) this is a sibling Cargo package — no root workspace; the parent rulake is reached via path = \"..\". Wire to an MCP client over stdio (e.g. Claude Desktop's claude_desktop_config.js","readmeSource":"rulake/crates/mcp-server/README.md","product":"key:rulake:rulake-mcp","canonical":"ruvector-rulake-mcp","tier":"T1","t":["ruvector-rulake-mcp","rulake","mcp","rulake-mcp","server","agent-callable","callable","witness-aware","witness","aware","vector","cache","adr-004","adr","004"],"s":["ruvector-rulake-mcp","rulake","mcp","rulake-mcp"]},
+{"id":"ruvector-rulake-mcp-ruqu","family":"rulake-mcp-ruqu","kind":"crate","store":"rulake","version":null,"description":"MCP companion server for ruQu v2 — agent-callable, witness-aware quantum execution surface (ADR-008 §6)","keywords":[],"source":"rulake/crates/mcp-ruqu/Cargo.toml","sourceSha256":"f4b1df905991a58c338c7a5c1d9451bc53ee25c5dbf120e56f8c9e70347aff76","variants":[],"product":"key:rulake:rulake-mcp-ruqu","canonical":"ruvector-rulake-mcp-ruqu","tier":"T1","t":["ruvector-rulake-mcp-ruqu","rulake","mcp","ruqu","rulake-mcp-ruqu","companion","server","agent-callable","callable","witness-aware","witness","aware","quantum","execution","surface","adr-008","adr","008"],"s":["ruvector-rulake-mcp-ruqu","rulake","mcp","ruqu","rulake-mcp-ruqu"]},
+{"id":"ruvector-rulake-mcp-rvdna","family":"rulake-mcp-rvdna","kind":"crate","store":"rulake","version":null,"description":"MCP companion server for the rvDNA v2 substrate (ADR-007 §6) — five witness-pinned read-tier tools","keywords":[],"source":"rulake/crates/mcp-rvdna/Cargo.toml","sourceSha256":"ac45d0c576138aff5a6baf5ef6bc2e4be17dcc2a331119ba5415f655bd3b5063","variants":[],"product":"key:rulake:rulake-mcp-rvdna","canonical":"ruvector-rulake-mcp-rvdna","tier":"T1","t":["ruvector-rulake-mcp-rvdna","rulake","mcp","rvdna","rulake-mcp-rvdna","companion","server","substrate","adr-007","adr","007","five","witness-pinned","witness","pin","read-tier","read","tier"],"s":["ruvector-rulake-mcp-rvdna","rulake","mcp","rvdna","rulake-mcp-rvdna"]},
+{"id":"ruvector-rulake-ruqu","family":"rulake-ruqu","kind":"crate","store":"rulake","version":null,"description":"ruQu v2 BackendAdapter for ruLake — quantum execution substrate, StateVector + Stabilizer + TensorNetwork backends (ADR-008)","keywords":[],"source":"rulake/crates/ruqu-backend/Cargo.toml","sourceSha256":"602f4dbc8681032b459179ec384c1edffcac6bc96b85fac4f45cb89af96058fc","variants":[],"product":"key:rulake:rulake-ruqu","canonical":"ruvector-rulake-ruqu","tier":"T1","t":["ruvector-rulake-ruqu","rulake","ruqu","rulake-ruqu","backendadapter","quantum","execution","substrate","statevector","stabilizer","tensornetwork","backend","adr-008","adr","008"],"s":["ruvector-rulake-ruqu","rulake","ruqu","rulake-ruqu"]},
+{"id":"ruvector-rulake-rvdna","family":"rulake-rvdna","kind":"crate","store":"rulake","version":null,"description":"rvDNA v2 BackendAdapter for ruLake — genomic substrate, T0 hot-tier + T1 warm-mmap tier + T2 cold-lazy tier (ADR-007)","keywords":[],"source":"rulake/crates/rvdna-backend/Cargo.toml","sourceSha256":"26782931e492e931c61de7aae1e6bdf41595ee469c9c8eca80dcb9f27a5dce28","variants":[],"product":"key:rulake:rulake-rvdna","canonical":"ruvector-rulake-rvdna","tier":"T1","t":["ruvector-rulake-rvdna","rulake","rvdna","rulake-rvdna","backendadapter","genomic","substrate","hot-tier","hot","tier","warm-mmap","warm","mmap","cold-lazy","cold","lazy","adr-007","adr","007"],"s":["ruvector-rulake-rvdna","rulake","rvdna","rulake-rvdna"]},
+{"id":"ruvector-snapshot","family":"snapshot","kind":"crate","store":"ruvector","version":null,"description":"Point-in-time snapshots and backup for Ruvector vector databases","keywords":[],"source":"ruvector/crates/ruvector-snapshot/Cargo.toml","sourceSha256":"87d562b896d25999a8bcbf1f89c0d610569d264593d0e734134ec5953809b34c","variants":[],"readme":"Point-in-time snapshots and backup for Ruvector vector databases. ruvector-snapshot provides efficient snapshot creation, storage, and restoration for Ruvector databases. Supports incremental snapshots, compression, and integrity verification. Part of the [Ruvector](https://github.com/ruvnet/ruvector) ecosystem. - Point-in-Time Recovery: Restore to any snapshot - Incremental Snapshots: Only store changed data - Compression: GZIP compression for storage efficiency - Integrity Verification: SHA-256 checksums - Async I/O: Non-blocking snapshot operations - Full Snapshots: Complete database backup - Incremental Snapshots: Delta-based backups - Compression: GZIP compression support - Checksums: S","readmeSource":"ruvector/crates/ruvector-snapshot/README.md","product":"key:ruvector:snapshot","canonical":"ruvector-snapshot","tier":"T0","t":["ruvector-snapshot","snapshot","point-in-time","point","backup","vector","databas"],"s":["ruvector-snapshot","snapshot"]},
+{"id":"ruvector-spann","family":"spann","kind":"crate","store":"ruvector","version":null,"description":"SPANN-inspired partition spilling for boundary-safe ANN in RuVector","keywords":[],"source":"ruvector/crates/ruvector-spann/Cargo.toml","sourceSha256":"3f947bab700996c37d76c045ff882a0ee6088290e6a772a4e1b2be9d521dcde6","variants":[],"product":"key:ruvector:spann","canonical":"ruvector-spann","tier":"T0","t":["ruvector-spann","spann","spann-inspired","inspir","partition","spill","boundary-safe","boundary","safe","ann"],"s":["ruvector-spann","spann"]},
+{"id":"ruvector-sparse-inference","family":"sparse-inference","kind":"crate","store":"ruvector","version":null,"description":"PowerInfer-style sparse inference engine for efficient neural network inference on edge devices","keywords":["edge-ai","neural-network","quantization","simd","sparse-inference"],"source":"ruvector/crates/ruvector-sparse-inference/Cargo.toml","sourceSha256":"e7295bcc00045c460d61e78a2e2fac73c1c3d564c7ba9782e440c3d6a79bbe62","variants":["ruvector-sparse-inference-wasm"],"readme":"PowerInfer-style Activation Locality Inference Engine for RuVector. A high-performance sparse inference engine that exploits neural network activation patterns to achieve 2×–10× speedups with <1% accuracy loss. - Activation Locality: Exploits power-law distribution where ~10% of neurons handle ~90% of activations - Low-Rank Prediction: Fast P·Q matrix factorization predicts active neurons in O(r·d) time - Sparse FFN: Computes only active neurons, skipping cold weights entirely - SIMD Optimization: AVX2/FMA (GELU, SiLU, axpy), SSE4.1, NEON, and WASM SIMD backends - GGUF Support: Full compatibility with quantized Llama models (Q4_0 through Q6_K) - Hot/Cold Caching: LRU/LFU strategies for intel","readmeSource":"ruvector/crates/ruvector-sparse-inference/README.md","product":"key:ruvector:sparse-inference","canonical":"ruvector-sparse-inference","tier":"T0","t":["ruvector-sparse-inference","sparse","inference","sparse-inference","powerinfer-style","powerinfer","style","efficient","neural","network","edge","device","edge-ai","neural-network","quantization"],"s":["ruvector-sparse-inference","sparse","inference","sparse-inference","edge-ai","edge","neural-network","neural","network","quantization"]},
+{"id":"ruvector-sparsifier","family":"sparsifier","kind":"crate","store":"ruvector","version":null,"description":"Dynamic spectral graph sparsification: always-on compressed world model for real-time graph analytics","keywords":["dynamic","dynamic-graph","graph","laplacian","sparsification","spectral","wasm"],"source":"ruvector/crates/ruvector-sparsifier/Cargo.toml","sourceSha256":"1fbaecebd1f4acf0f99d1f55d669b2eccb5104f07e387fba628bda2db92d1a78","variants":["ruvector-sparsifier-wasm"],"readme":"An always-on compressed world model for real-time graph analytics. *Dynamic spectral sparsification for HNSW health monitoring, structural diagnostics, and continuous graph reasoning.* Every vector database, similarity index, and memory graph is backed by a dense web of connections. Analyzing the full graph is expensive — often too expensive for real-time use. RuVector Sparsifier maintains a small shadow graph that provably preserves the spectral structure of your full graph, enabling: - Continuous monitoring instead of batch analysis - 3-10x faster graph diagnostics (min-cut, clustering, Laplacian solves) - 10-50x less memory for analytics workloads - Early anomaly detection via structural","readmeSource":"ruvector/crates/ruvector-sparsifier/README.md","product":"key:ruvector:sparsifier","canonical":"ruvector-sparsifier","tier":"T0","t":["ruvector-sparsifier","sparsifier","dynamic","spectral","graph","sparsification","always-on","alway","compres","world","model","real","analytic","dynamic-graph","laplacian"],"s":["ruvector-sparsifier","sparsifier","dynamic","dynamic-graph","graph","laplacian","sparsification","spectral"]},
+{"id":"ruvector-speculative-ann","family":"speculative-ann","kind":"crate","store":"ruvector","version":null,"description":"Speculative ANN search for RuVector: draft-verify retrieval inspired by speculative decoding, with adaptive candidate multiplier and scalar quantization","keywords":["agent-memory","ann","quantization","speculative","vector-search"],"source":"ruvector/crates/ruvector-speculative-ann/Cargo.toml","sourceSha256":"c0635d9fb8da1a1e11df45b4d0c7369445cc745e1055df99ebeb697346a15eb8","variants":[],"readme":"Experimental draft-then-verify approximate nearest-neighbour search for RuVector. A scalar-quantized scan proposes candidates and exact f32 distances rerank them. The adaptive controller changes its candidate multiplier only when callers provide sampled ground-truth recall feedback. Calls without audited feedback use the current multiplier unchanged. This crate is research-tier. See docs/adr/ADR-272-speculative-ann-search.md in the RuVector repository for benchmarks, limitations, and the production roadmap.","readmeSource":"ruvector/crates/ruvector-speculative-ann/README.md","product":"key:ruvector:speculative-ann","canonical":"ruvector-speculative-ann","tier":"T0","t":["ruvector-speculative-ann","speculative","ann","speculative-ann","search","draft-verify","draft","verify","retrieval","inspir","decod","adaptive","candidate","multiplier","scalar","quantization","agent-memory","memory","vector-search","vector"],"s":["ruvector-speculative-ann","speculative","ann","speculative-ann","agent-memory","memory","quantization","vector-search","vector","search"]},
+{"id":"ruvector-staged-workspace","family":"staged-workspace","kind":"crate","store":"ruvector","version":null,"description":"StagedWorkspace (arXiv:2608.18050) content-hash + revision-id state binding for RuV artifacts: stale views fail closed (ADR-318, PIR WP16)","keywords":["agent","content-hash","provenance","ruvector","workspace"],"source":"ruvector/crates/ruvector-staged-workspace/Cargo.toml","sourceSha256":"70464fe4e154389cfda133089f4c170a37277da7daa86d7249b1f1ec8bdcb20c","variants":[],"product":"key:ruvector:staged-workspace","canonical":"ruvector-staged-workspace","tier":"T0","t":["ruvector-staged-workspace","stag","workspace","staged-workspace","stagedworkspace","arxiv","2608.18050","content-hash","content","hash","revision-id","revision","state","artifact","stale","view","fail","clos","adr-318","adr","318","pir","wp16","provenance"],"s":["ruvector-staged-workspace","stag","workspace","staged-workspace","content-hash","content","hash","provenance"]},
+{"id":"ruvector-streaming-qng","family":"streaming-qng","kind":"crate","store":"ruvector","version":null,"description":"Online reservoir-sampled product quantization for streaming ANN: three measurable variants — full-precision, static-PQ, and adaptive streaming-PQ with distribution-drift resilience","keywords":["agent-memory","ann","product-quantization","streaming","vector-search"],"source":"ruvector/crates/ruvector-streaming-qng/Cargo.toml","sourceSha256":"bb98cfacd89a7de021196c1b9b74b3e23d17c5b0fe6f39e59f99712513184ac0","variants":[],"readme":"Experimental product quantization for streaming vector collections under distribution drift. It provides full-precision, static-PQ, and reservoir-based adaptive PQ variants behind a common interface. The adaptive implementation periodically retrains its codebook and re-encodes stored vectors. This improves drift resilience at the cost of insertion time and retaining full-precision vectors.","readmeSource":"ruvector/crates/ruvector-streaming-qng/README.md","product":"key:ruvector:streaming-qng","canonical":"ruvector-streaming-qng","tier":"T0","t":["ruvector-streaming-qng","stream","qng","streaming-qng","online","reservoir-sampled","reservoir","sampl","product","quantization","ann","three","measurable","variant","full-precision","full","precision","static-pq","static","adaptive","streaming-pq","distribution-drift","distribution","drift","resilience","agent-memory","memory","product-quantization","vector-search","vector","search"],"s":["ruvector-streaming-qng","stream","qng","streaming-qng","agent-memory","memory","ann","product-quantization","product","quantization","vector-search","vector","search"]},
+{"id":"ruvector-temporal-coherence","family":"temporal-coherence","kind":"crate","store":"ruvector","version":null,"description":"Temporal coherence decay for agent memory retrieval — three scored variants with graph-coherence gating","keywords":["agent-memory","coherence","ruvector","temporal","vector-search"],"source":"ruvector/crates/ruvector-temporal-coherence/Cargo.toml","sourceSha256":"3ebdf87e080213b0ad617499728f9bd80dd4be533f600aef3625fe53b6bb4670","variants":[],"product":"key:ruvector:temporal-coherence","canonical":"ruvector-temporal-coherence","tier":"T0","t":["ruvector-temporal-coherence","temporal","coherence","temporal-coherence","decay","memory","retrieval","three","scor","variant","graph-coherence","graph","gating","agent-memory","vector-search","vector","search"],"s":["ruvector-temporal-coherence","temporal","coherence","temporal-coherence","agent-memory","memory","vector-search","vector","search"]},
+{"id":"ruvector-temporal-tensor","family":"temporal-tensor","kind":"crate","store":"ruvector","version":null,"description":"Temporal tensor compression with tiered quantization for RuVector","keywords":[],"source":"ruvector/crates/ruvector-temporal-tensor/Cargo.toml","sourceSha256":"60aff197a9c6c4e0fec6edbe9f2725c09daeb283e037d0ed7ccd99d4d3194799","variants":["ruvector-temporal-tensor-wasm"],"readme":"Shrink your vector data 4-10x without losing the signal. ruvector-temporal-tensor compresses streams of floating-point tensors by exploiting two properties that most vector workloads share: 1. Values within a group are similar — so a single scale factor per group captures the range, and a small integer code captures the value. This is *groupwise symmetric quantization*. 2. Consecutive frames barely change — so the same scale factors can be reused across many frames until the data drifts. This is *temporal segment reuse*. The crate automatically picks the right bit-width based on how \"hot\" (frequently accessed) the tensor is, giving you aggressive compression on cold data while preserving acc","readmeSource":"ruvector/crates/ruvector-temporal-tensor/README.md","product":"key:ruvector:temporal-tensor","canonical":"ruvector-temporal-tensor","tier":"T0","t":["ruvector-temporal-tensor","temporal","tensor","temporal-tensor","compression","tier","quantization"],"s":["ruvector-temporal-tensor","temporal","tensor","temporal-tensor"]},
+{"id":"ruvector-turboquant","family":"turboquant","kind":"crate","store":"ruvector","version":null,"description":"Turbo4: 4-bit Lloyd-Max quantized vector datatype with deterministic Hadamard rotation and direct packed SIMD scoring (ADR-296)","keywords":[],"source":"ruvector/crates/ruvector-turboquant/Cargo.toml","sourceSha256":"27f00dcc161a6e63e138a57b840d827c2336ba8ed162c83defeff395e409a297","variants":[],"product":"key:ruvector:turboquant","canonical":"ruvector-turboquant","tier":"T0","t":["ruvector-turboquant","turboquant","turbo4","4-bit","bit","lloyd-max","lloyd","max","quantiz","vector","datatype","deterministic","hadamard","rotation","direct","pack","scor","adr-296","adr","296"],"s":["ruvector-turboquant","turboquant"]},
+{"id":"ruvector-turbovec","family":"turbovec","kind":"crate","store":"ruvector","version":null,"description":"TurboVec: multi-bit TurboQuant ANN scalar reference with 2/3/4-bit Lloyd-Max quantization, TQ+ calibration, and length-renormalized scoring.","keywords":[],"source":"ruvector/crates/ruvector-turbovec/Cargo.toml","sourceSha256":"d910408f540371e7a5c47348e8b7523309e8b3af426fb0de84acd880224f5350","variants":[],"product":"key:ruvector:turbovec","canonical":"ruvector-turbovec","tier":"T0","t":["ruvector-turbovec","turbovec","multi-bit","multi","bit","turboquant","ann","scalar","reference","4-bit","lloyd-max","lloyd","max","quantization","calibration","length-renormalized","length","renormaliz","scor"],"s":["ruvector-turbovec","turbovec"]},
+{"id":"ruvector-typesafe-train","family":"typesafe-train","kind":"crate","store":"ruvector","version":null,"description":"OpenJev v0 trainer (ADR-008): fine-tunes bge-small-en-v1.5 in candle and transplants the weights into the pinned ONNX graph","keywords":[],"source":"ruvector/crates/ruvector-typesafe-train/Cargo.toml","sourceSha256":"aef9d9db5227e80a826b7e17c45d8250d8e22d2f1208faba6bc7ce2adefacf18","variants":[],"product":"key:ruvector:typesafe-train","canonical":"ruvector-typesafe-train","tier":"T0","t":["ruvector-typesafe-train","typesafe","train","typesafe-train","openjev","trainer","adr-008","adr","008","fine-tunes","fine","tune","bge-small-en-v1.5","bge","small","v1.5","candle","transplant","weight","pin","onnx","graph"],"s":["ruvector-typesafe-train","typesafe","train","typesafe-train"]},
+{"id":"ruvector-verified","family":"verified","kind":"crate","store":"ruvector","version":null,"description":"Formal verification layer for RuVector: proof-carrying vector operations with sub-microsecond overhead using lean-agentic dependent types","keywords":["dependent-types","formal-methods","proof-carrying","vector-database","verification","wasm","webassembly"],"source":"ruvector/crates/ruvector-verified/Cargo.toml","sourceSha256":"762639319f81b2f8dcfa53b719fb1707422db595d42aed77accede24688b9c43","variants":["ruvector-verified-wasm"],"readme":"Proof-carrying vector operations for Rust. Every dimension check, HNSW insert, and pipeline composition produces a machine-checked proof witness -- catching bugs that assert! misses, with less than 2% runtime overhead. Built on [lean-agentic](https://crates.io/crates/lean-agentic) dependent types. Part of the [RuVector](https://github.com/ruvnet/ruvector) ecosystem. Vector databases silently corrupt results when dimensions mismatch. A 384-dim query against a 768-dim index doesn't panic -- it returns wrong answers. Traditional approaches either: - Runtime assert! -- panics in production, no proof trail - Const generics -- catches errors at compile time, but can't handle dynamic dimensions fro","readmeSource":"ruvector/crates/ruvector-verified/README.md","product":"key:ruvector:verified","canonical":"ruvector-verified","tier":"T0","t":["ruvector-verified","verify","formal","verification","layer","proof-carrying","proof","carry","vector","operation","sub-microsecond","sub","microsecond","overhead","lean-agentic","lean","dependent","type","dependent-types","formal-methods","method","vector-database","database"],"s":["ruvector-verified","verify","dependent-types","dependent","type","formal-methods","formal","method","proof-carrying","proof","carry","vector-database","vector","database","verification"]},
+{"id":"ruview-auth","family":"ruview-auth","kind":"crate","store":"ruview","version":null,"description":"Cognitum OAuth access-token verification for RuView (ADR-271)","keywords":[],"source":"ruview/v2/crates/ruview-auth/Cargo.toml","sourceSha256":"32619c5c0028b1419dd7560e9dbf140140dfbe92ba944f3e14fb37482066ab64","variants":[],"product":"key:ruview:ruview-auth","canonical":"ruview-auth","tier":"T0","t":["ruview-auth","ruview","auth","cognitum","oauth","access-token","access","token","verification","adr-271","adr","271"],"s":["ruview-auth","ruview","auth"]},
+{"id":"ruview-cognitum-spaces","family":"ruview-cognitum-spaces","kind":"crate","store":"ruview","version":null,"description":"Bounded, privacy-preserving Cognitum Spaces client for RuView","keywords":[],"source":"ruview/v2/crates/ruview-cognitum-spaces/Cargo.toml","sourceSha256":"98f1230b97284109e69af6b490b520a5d20f71b946116a7763ffec191ed98023","variants":[],"product":"key:ruview:ruview-cognitum-spaces","canonical":"ruview-cognitum-spaces","tier":"T0","t":["ruview-cognitum-spaces","ruview","cognitum","space","bound","privacy-preserving","privacy","preserv","client"],"s":["ruview-cognitum-spaces","ruview","cognitum","space"]},
+{"id":"ruview-offaxis","family":"ruview-offaxis","kind":"crate","store":"ruview","version":null,"description":"Clean-room generalized off-axis (head-coupled) perspective projection for RuView demos (ADR-324)","keywords":[],"source":"ruview/v2/crates/ruview-offaxis/Cargo.toml","sourceSha256":"ef70ecc90f10383af2db1ae8a258ffde1926153da3345c5cee4e86f30250b877","variants":[],"readme":"Clean-room off-axis (head-coupled) perspective projection in Rust, with a wasm-bindgen surface for browser demos. Implements ADR-324. The screen becomes a window: given the physical screen's corners and the viewer's eye position, the crate produces the asymmetric frustum and screen-aligned view matrix that keep the screen plane fixed while everything behind and in front of it moves with true parallax. ADR-324 records that the prior-art repository (icurtis1/off-axis-sneaker) publishes no license. No code, assets, or derived text from it appear in this crate. The implementation follows the published math only: - Robert Kooima, *Generalized Perspective Projection* (2008) — the pa/pb/pc screen-c","readmeSource":"ruview/v2/crates/ruview-offaxis/README.md","product":"key:ruview:ruview-offaxis","canonical":"ruview-offaxis","tier":"T0","t":["ruview-offaxis","ruview","offaxis","clean-room","clean","room","generaliz","off-axis","off","axis","head-coupled","head","coupl","perspective","projection","demo","adr-324","adr","324"],"s":["ruview-offaxis","ruview","offaxis"]},
+{"id":"ruview-spatial-memory","family":"ruview-spatial-memory","kind":"crate","store":"ruview","version":null,"description":"Tenant-scoped encrypted RuVector spatial memory for Cognitum Spaces","keywords":[],"source":"ruview/v2/crates/ruview-spatial-memory/Cargo.toml","sourceSha256":"2c333971a8ac6851b692237626048a50c528834d33620dff0a1a3c5a0f573a57","variants":[],"product":"key:ruview:ruview-spatial-memory","canonical":"ruview-spatial-memory","tier":"T0","t":["ruview-spatial-memory","ruview","spatial","memory","tenant-scoped","tenant","scop","encrypt","cognitum","space"],"s":["ruview-spatial-memory","ruview","spatial","memory"]},
+{"id":"ruview-swarm","family":"ruview-swarm","kind":"crate","store":"ruv-drone","version":null,"description":"Industrial cooperative-UAV coordination — formation keeping, collision avoidance, Raft consensus, cooperative task allocation, MAPPO navigation (ADR-148)","keywords":[],"source":"ruv-drone/Cargo.toml","sourceSha256":"228bab572a4026afd4ea1cc5652420968005add1d53e061f6ad757b76f45faf8","variants":[],"readme":"Industrial cooperative-UAV fleet coordination, in Rust. ruv-drone is a coordination layer that sits *above* a per-vehicle autopilot (PX4 / ArduPilot) and turns a set of drones into a coordinated fleet — formation keeping, distributed consensus, cooperative task allocation, collision-avoidant planning, and learned multi-agent navigation. It targets civilian missions: search-and-rescue, infrastructure inspection, agriculture, mapping, and emergency telecom relay. Part of the [RuView / wifi-densepose](https://github.com/ruvnet/wifi-densepose) ecosystem (ADR-148), with an optional WiFi-CSI sensing payload for through-structure presence detection. Pure Rust, async, edge-deployable. The line this","readmeSource":"ruv-drone/README.md","product":"repo:ruv-drone","canonical":"ruview-swarm","tier":"T2","t":["ruview-swarm","ruview","swarm","industrial","cooperative-uav","cooperative","uav","coordination","formation","collision","avoidance","raft","consensus","task","allocation","mappo","navigation","adr-148","adr","148"],"s":["ruview-swarm","ruview","swarm"]},
+{"id":"ruview-unified","family":"ruview-unified","kind":"crate","store":"ruview","version":null,"description":"Unified RF spatial world model (ADR-273): canonical RF tensor + hardware adapters, universal RF foundation encoder, RF-aware Gaussian spatial memory, physics-guided synthetic RF worlds, and the edge sensing control plane","keywords":["csi","gaussian-splatting","rf-sensing","wifi","world-model"],"source":"ruview/v2/crates/ruview-unified/Cargo.toml","sourceSha256":"cea64e7714a692235216f73ee32f05942a48cb4000f0d17591c34a7b3437219c","variants":[],"product":"key:ruview:ruview-unified","canonical":"ruview-unified","tier":"T0","t":["ruview-unified","ruview","unify","spatial","world","model","adr-273","adr","273","canonical","tensor","hardware","adapter","universal","foundation","encoder","rf-aware","aware","gaussian","memory","physics-guided","physic","guid","synthetic","edge","sens","control","plane","csi","gaussian-splatting","splatt","rf-sensing","wifi","world-model"],"s":["ruview-unified","ruview","unify","csi","gaussian-splatting","gaussian","splatt","rf-sensing","sens","wifi","world-model","world","model"]},
+{"id":"ruvix-bcm2711","family":"ruvix-bcm2711","kind":"crate","store":"ruvector","version":null,"description":"BCM2711/BCM2712 SoC drivers for Raspberry Pi 4/5 (RuVix Phase D)","keywords":["bcm2711","embedded","kernel","no_std","rpi"],"source":"ruvector/crates/ruvix/crates/bcm2711/Cargo.toml","sourceSha256":"7812c5ff67b79f4a8aa5ee436f814282712e843c719d884b7e4df876a2622d9c","variants":[],"readme":"BCM2711/BCM2712 SoC drivers for Raspberry Pi 4/5 as part of the RuVix Cognition Kernel (ADR-087). This crate provides low-level, no_std drivers for the Broadcom BCM2711 (Raspberry Pi 4) and BCM2712 (Raspberry Pi 5) System-on-Chip devices. - Raspberry Pi 4 Model B (1GB, 2GB, 4GB, 8GB) - Raspberry Pi 400 - Raspberry Pi Compute Module 4 - Raspberry Pi 5 (BCM2712, with rpi5 feature) - 58 GPIO pins with configurable functions - Pull-up/pull-down resistor control (BCM2711 new register layout) - Edge detection and event status - Digital input/output - Serial console I/O via auxiliary UART - Configurable baud rate (default 115200) - 8N1 configuration - Implements core::fmt::Write - Communication wit","readmeSource":"ruvector/crates/ruvix/crates/bcm2711/README.md","product":"key:ruvector:ruvix-bcm2711","canonical":"ruvix-bcm2711","tier":"T0","t":["ruvix-bcm2711","ruvix","bcm2711","bcm2712","soc","driver","raspberry","phase","embed","kernel","std","rpi"],"s":["ruvix-bcm2711","ruvix","bcm2711","embed","kernel","std","rpi"]},
+{"id":"ruvix-boot","family":"ruvix-boot","kind":"crate","store":"ruvector","version":null,"description":"RVF boot loading for RuVix Cognition Kernel (ADR-087)","keywords":["attestation","boot","kernel","rvf","signature"],"source":"ruvector/crates/ruvix/crates/boot/Cargo.toml","sourceSha256":"e276b37d158cb23a824eeb3483f49f5bff798ac3de576727f50a99917a4fa841","variants":[],"readme":"RVF boot loading for the RuVix Cognition Kernel (ADR-087). This crate provides the RVF (RuVector Format) boot loading infrastructure. The kernel boot follows a strict five-stage process with cryptographic verification at every step. Critical security fixes implemented: - Signature failure: PANIC IMMEDIATELY, no fallback boot path - Root task capability drop: After Stage 3, root task drops to minimum set - Witness log integrity: Append-only, cryptographically linked Main boot orchestrator: RVF package manifest structure: ML-DSA-65 signature verification (NIST FIPS 204): Append-only witness log for boot attestation: Post-boot capability restriction: - std (default): Enable standard library sup","readmeSource":"ruvector/crates/ruvix/crates/boot/README.md","product":"key:ruvector:ruvix-boot","canonical":"ruvix-boot","tier":"T0","t":["ruvix-boot","ruvix","boot","rvf","load","cognition","kernel","adr-087","adr","087","attestation","signature"],"s":["ruvix-boot","ruvix","boot","attestation","kernel","rvf","signature"]},
+{"id":"ruvix-cap","family":"ruvix-cap","kind":"crate","store":"ruvector","version":null,"description":"seL4-inspired capability management for the RuVix Cognition Kernel (ADR-087)","keywords":["capability","kernel","no_std","security","sel4"],"source":"ruvector/crates/ruvix/crates/cap/Cargo.toml","sourceSha256":"f89cd443e04087debd9a51a7e70397ce388cdbb5649a97e7da909a1490b9913f","variants":[],"readme":"seL4-inspired capability management for the RuVix Cognition Kernel (ADR-087). This crate provides the capability manager that enforces all access control in RuVix. Every kernel object is accessed exclusively through capabilities, following the principle: \"No syscall succeeds without an appropriate capability handle.\" 1. A task can only grant capabilities it holds 2. Granted rights must be equal or fewer than held rights 3. Revocation propagates through the derivation tree 4. GRANT_ONCE provides non-transitive delegation 5. Epoch-based invalidation detects stale handles Central capability management: Delegate capabilities with restricted rights: Revoke capabilities and all derivatives: Track","readmeSource":"ruvector/crates/ruvix/crates/cap/README.md","product":"key:ruvector:ruvix-cap","canonical":"ruvix-cap","tier":"T0","t":["ruvix-cap","ruvix","cap","sel4-inspired","sel4","inspir","capability","management","cognition","kernel","adr-087","adr","087","std","security"],"s":["ruvix-cap","ruvix","cap","capability","kernel","std","security","sel4"]},
+{"id":"ruvix-cli","family":"ruvix","kind":"crate","store":"ruvector","version":null,"description":"Host-side CLI tool for RuVix cognition kernel","keywords":["capability","cognition","kernel","no_std","proof-gated"],"source":"ruvector/crates/ruvix/crates/cli/Cargo.toml","sourceSha256":"ef8d73d94d49965ca18fb7130bec49ad31111c105e115f877ccdceb40e4ccac6","variants":["ruvix-types"],"product":"key:ruvector:ruvix","canonical":"ruvix-cli","tier":"T0","t":["ruvix-cli","ruvix","host-side","host","side","cognition","kernel","capability","std","proof-gated","proof","gated"],"s":["ruvix-cli","ruvix","capability","cognition","kernel","std","proof-gated","proof","gated"]},
+{"id":"ruvix-dma","family":"ruvix-dma","kind":"crate","store":"ruvector","version":null,"description":"DMA controller abstraction for RuVix Cognition Kernel (ADR-087)","keywords":["dma","embedded","kernel","memory","no_std"],"source":"ruvector/crates/ruvix/crates/dma/Cargo.toml","sourceSha256":"bc835bff1a483af9d8e250eefaf8d31a2c878fd14feaf00109ed29de7d56c55f","variants":[],"readme":"DMA (Direct Memory Access) controller abstraction for the RuVix Cognition Kernel. This crate provides a hardware-agnostic DMA controller interface for efficient zero-copy data transfers between memory regions and peripheral devices. It is designed as part of ADR-087 for the RuVix Cognition Kernel. - No-std compatible: Works in bare-metal environments - Zero unsafe code: All public APIs are safe Rust - Scatter-gather support: Descriptor chains for non-contiguous transfers - Cache coherent buffers: Proper cache management for DMA operations - Platform-agnostic: Works across ARM64, RISC-V, and x86_64 Represents a single DMA channel with state tracking: Transfer direction enumeration: Cache-cohe","readmeSource":"ruvector/crates/ruvix/crates/dma/README.md","product":"key:ruvector:ruvix-dma","canonical":"ruvix-dma","tier":"T0","t":["ruvix-dma","ruvix","dma","controller","abstraction","cognition","kernel","adr-087","adr","087","embed","memory","std"],"s":["ruvix-dma","ruvix","dma","embed","kernel","memory","std"]},
+{"id":"ruvix-drivers","family":"ruvix-drivers","kind":"crate","store":"ruvector","version":null,"description":"Device drivers for RuVix Cognition Kernel (ADR-087)","keywords":["drivers","gic","kernel","no_std","uart"],"source":"ruvector/crates/ruvix/crates/drivers/Cargo.toml","sourceSha256":"cc0de0ff9b47ed49fdfb3fb7dc2fa646d3320a9de6314c4061cf5ebf45aab612","variants":[],"readme":"Device drivers for the RuVix Cognition Kernel (ADR-087). This crate provides hardware drivers for the QEMU virt machine on AArch64. It implements the Hardware Abstraction Layer (HAL) traits defined in ruvix-hal.","readmeSource":"ruvector/crates/ruvix/crates/drivers/README.md","product":"key:ruvector:ruvix-drivers","canonical":"ruvix-drivers","tier":"T0","t":["ruvix-drivers","ruvix","driver","device","cognition","kernel","adr-087","adr","087","gic","std","uart"],"s":["ruvix-drivers","ruvix","driver","gic","kernel","std","uart"]},
+{"id":"ruvix-dtb","family":"ruvix-dtb","kind":"crate","store":"ruvector","version":null,"description":"Device Tree Blob parser for RuVix Cognition Kernel (ADR-087)","keywords":["devicetree","dtb","embedded","kernel","no_std"],"source":"ruvector/crates/ruvix/crates/dtb/Cargo.toml","sourceSha256":"faf82e8258af8d207fc646d8efa2836212d1424a80d8cae9fdc3eaf2a1e01f3c","variants":[],"readme":"Device Tree Blob (FDT) parser for the RuVix Cognition Kernel. This crate provides a zero-copy parser for Flattened Device Tree (FDT) blobs, enabling hardware discovery at boot time. It is designed as part of ADR-087 for the RuVix Cognition Kernel. - No-std compatible: Works in bare-metal environments - Zero-copy parsing: References directly into the FDT blob - No heap allocation: Fixed-size internal buffers - Safe parsing: Validates all data before access - FDT version 17: Compatible with standard device trees Main parser and accessor for device tree data: Reference to a node in the tree: Access to node properties: Parsed memory region entry: A Flattened Device Tree consists of: 1. Header (4","readmeSource":"ruvector/crates/ruvix/crates/dtb/README.md","product":"key:ruvector:ruvix-dtb","canonical":"ruvix-dtb","tier":"T0","t":["ruvix-dtb","ruvix","dtb","device","tree","blob","parser","cognition","kernel","adr-087","adr","087","devicetree","embed","std"],"s":["ruvix-dtb","ruvix","dtb","devicetree","embed","kernel","std"]},
+{"id":"ruvix-fs","family":"ruvix-fs","kind":"crate","store":"ruvector","version":null,"description":"Minimal filesystem for the RuVix Cognition Kernel (ADR-087 Phase E)","keywords":["fat32","filesystem","kernel","no_std","vfs"],"source":"ruvector/crates/ruvix/crates/fs/Cargo.toml","sourceSha256":"3572cdfbb65511bc11f97e96c87113ba3a9b2452845a1c2d72f1302b5ca24b4c","variants":[],"readme":"Minimal filesystem implementation for the RuVix Cognition Kernel (ADR-087 Phase E). This crate provides a VFS (Virtual Filesystem) layer with pluggable filesystem backends: - RamFS: In-memory filesystem for /tmp and temporary storage - FAT32: Read-only FAT32 filesystem for boot partitions and external media","readmeSource":"ruvector/crates/ruvix/crates/fs/README.md","product":"key:ruvector:ruvix-fs","canonical":"ruvix-fs","tier":"T0","t":["ruvix-fs","ruvix","minimal","filesystem","cognition","kernel","adr-087","adr","087","phase","fat32","std","vfs"],"s":["ruvix-fs","ruvix","fat32","filesystem","kernel","std","vfs"]},
+{"id":"ruvix-hal","family":"ruvix-hal","kind":"crate","store":"ruvector","version":null,"description":"Hardware Abstraction Layer for RuVix Cognition Kernel (ADR-087)","keywords":["aarch64","embedded","hal","kernel","no_std"],"source":"ruvector/crates/ruvix/crates/hal/Cargo.toml","sourceSha256":"dd2b6bf5db854abe22b5ef41030afc169dd3053dd87e069c469681b4a33d7b4d","variants":[],"readme":"Hardware Abstraction Layer (HAL) for the RuVix Cognition Kernel (ADR-087). ruvix-hal provides platform-independent traits that hardware-specific implementations must satisfy. This crate is the foundation for building portable kernel code that works across ARM64, RISC-V, and x86_64 platforms. - #![no_std] - No standard library dependency - #![forbid(unsafe_code)] - Trait definitions are completely safe - Zero dependencies - Only depends on ruvix-types - Platform-agnostic - Traits work across all supported architectures - Result-based errors - All fallible operations return Result Serial I/O for kernel debugging and early boot logging. Key Features: - Blocking and buffered writes - IRQ-safe op","readmeSource":"ruvector/crates/ruvix/crates/hal/README.md","product":"key:ruvector:ruvix-hal","canonical":"ruvix-hal","tier":"T0","t":["ruvix-hal","ruvix","hal","hardware","abstraction","layer","cognition","kernel","adr-087","adr","087","aarch64","embed","std"],"s":["ruvix-hal","ruvix","hal","aarch64","embed","kernel","std"]},
+{"id":"ruvix-net","family":"ruvix-net","kind":"crate","store":"ruvector","version":null,"description":"Minimal networking stack for the RuVix Cognition Kernel (ADR-087 Phase E)","keywords":["ethernet","kernel","networking","no_std","tcp-ip"],"source":"ruvector/crates/ruvix/crates/net/Cargo.toml","sourceSha256":"657cadaed08c81296de19c0a8f305025933862297856841511910afc11ffac81","variants":[],"readme":"Minimal networking stack for the RuVix Cognition Kernel (ADR-087 Phase E). This crate provides a complete no_std networking stack suitable for embedded systems and OS kernels. It implements: - Ethernet II frame handling - ARP address resolution with cache management - IPv4 routing and header processing - UDP datagram transport - ICMP echo (ping) support - std - Enable standard library support - alloc - Enable heap allocation via the alloc crate By default, the crate is #![no_std] compatible with zero heap allocations. - MacAddress - 6-byte hardware address with broadcast/multicast detection - EthernetFrame - Parsing and serialization of Ethernet II frames - EtherType - Protocol identifiers (","readmeSource":"ruvector/crates/ruvix/crates/net/README.md","product":"key:ruvector:ruvix-net","canonical":"ruvix-net","tier":"T0","t":["ruvix-net","ruvix","net","minimal","network","stack","cognition","kernel","adr-087","adr","087","phase","ethernet","std","tcp-ip","tcp"],"s":["ruvix-net","ruvix","net","ethernet","kernel","network","std","tcp-ip","tcp"]},
+{"id":"ruvix-nucleus","family":"ruvix-nucleus","kind":"crate","store":"ruvector","version":null,"description":"Integration crate for RuVix Cognition Kernel - syscall dispatch, deterministic replay, and full acceptance tests (ADR-087)","keywords":["cognition","deterministic","kernel","replay","syscall"],"source":"ruvector/crates/ruvix/crates/nucleus/Cargo.toml","sourceSha256":"04adb4da61db8e31c1395eb1fbf8e8f4e7029b1d03bdfa3629f428f6b79d9ad6","variants":[],"readme":"Integration crate for the RuVix Cognition Kernel (ADR-087). The ruvix-nucleus crate brings all RuVix subsystems together and provides: - Syscall dispatch table for all 12 syscalls defined in ADR-087 - Kernel struct coordinating all subsystems - Deterministic replay support for checkpoint/restore - Witness log for attestation and auditability - Proof-gated mutations with multi-tier verification The kernel provides 12 syscalls as defined in ADR-087 Section 3.1: 1. Capability-gated: No syscall succeeds without appropriate capability 2. Proof-required for mutation: vector_put_proved, graph_apply_proved, rvf_mount 3. Bounded latency: No unbounded loops in syscall path 4. Witness-logged: Every mut","readmeSource":"ruvector/crates/ruvix/crates/nucleus/README.md","product":"key:ruvector:ruvix-nucleus","canonical":"ruvix-nucleus","tier":"T0","t":["ruvix-nucleus","ruvix","nucleus","integration","crate","cognition","kernel","syscall","dispatch","deterministic","replay","full","acceptance","test","adr-087","adr","087"],"s":["ruvix-nucleus","ruvix","nucleus","cognition","deterministic","kernel","replay","syscall"]},
+{"id":"ruvix-physmem","family":"ruvix-physmem","kind":"crate","store":"ruvector","version":null,"description":"Physical memory allocator for RuVix Cognition Kernel (ADR-087)","keywords":["allocator","buddy","kernel","memory","no_std"],"source":"ruvector/crates/ruvix/crates/physmem/Cargo.toml","sourceSha256":"a49bdcee9abe51f374ff04938c1af041778401dea145f85c5ea9a5d550d8a321","variants":[],"readme":"Physical memory allocator for the RuVix Cognition Kernel (ADR-087). This crate provides a buddy allocator for physical page frame allocation. The buddy system uses power-of-two block sizes (4KB to 2MB) for efficient allocation and deallocation with minimal external fragmentation. - Buddy Allocation: Power-of-two block sizes with automatic splitting and coalescing - No External Dependencies: Pure no_std implementation - Type-Safe Addresses: PhysAddr newtype wrapper for physical addresses - Comprehensive Statistics: Track allocations, frees, splits, and coalesces - Page Frame Abstraction: PageFrame type for safer memory management","readmeSource":"ruvector/crates/ruvix/crates/physmem/README.md","product":"key:ruvector:ruvix-physmem","canonical":"ruvix-physmem","tier":"T0","t":["ruvix-physmem","ruvix","physmem","physical","memory","allocator","cognition","kernel","adr-087","adr","087","buddy","std"],"s":["ruvix-physmem","ruvix","physmem","allocator","buddy","kernel","memory","std"]},
+{"id":"ruvix-proof","family":"ruvix-proof","kind":"crate","store":"ruvector","version":null,"description":"Proof engine with 3-tier routing for the RuVix Cognition Kernel (ADR-087)","keywords":["capability","kernel","no_std","proof","verification"],"source":"ruvector/crates/ruvix/crates/proof/Cargo.toml","sourceSha256":"f87207425557f441a9c4cccbb3c6ad905a0a169c856f9af94becea7c0cff2b5d","variants":[],"readme":"Proof engine with 3-tier routing for the RuVix Cognition Kernel (ADR-087). This crate implements the proof engine that enforces proof-gated mutation in RuVix. Every state mutation requires a valid proof token, ensuring the kernel physically prevents unauthorized changes. - Time-bounded validity: Proofs expire after a configurable window (default 100ms) - Single-use nonces: Each nonce can only be consumed once - Capability-gated: PROVE rights required on target object - Cache limits: Maximum 64 entries with 100ms TTL Generates proof tokens with automatic tier routing based on operation complexity. Verifies proof tokens with security checks. Manages proof caching with single-use nonce semantic","readmeSource":"ruvector/crates/ruvix/crates/proof/README.md","product":"key:ruvector:ruvix-proof","canonical":"ruvix-proof","tier":"T0","t":["ruvix-proof","ruvix","proof","3-tier","tier","rout","cognition","kernel","adr-087","adr","087","capability","std","verification"],"s":["ruvix-proof","ruvix","proof","capability","kernel","std","verification"]},
+{"id":"ruvix-qemu-swarm","family":"ruvix-qemu-swarm","kind":"crate","store":"ruvector","version":null,"description":"QEMU swarm simulation for distributed RuVix cluster testing","keywords":["distributed","qemu","ruvix","swarm","testing"],"source":"ruvector/crates/ruvix/qemu-swarm/Cargo.toml","sourceSha256":"f01530585b5547fc2f6ab303e9a11610f86539c3b64efbe01eef45e3fe2e886e","variants":[],"readme":"A comprehensive simulation environment for testing distributed RuVix clusters using multiple QEMU instances. Enables swarm intelligence testing, multi-node consensus verification, and chaos engineering for the RuVix Cognition Kernel. - QEMU 8.0+ with AArch64 support (qemu-system-aarch64) - Rust 1.77+ (for building) - socat (for shell scripts)","readmeSource":"ruvector/crates/ruvix/qemu-swarm/README.md","product":"key:ruvector:ruvix-qemu-swarm","canonical":"ruvix-qemu-swarm","tier":"T0","t":["ruvix-qemu-swarm","ruvix","qemu","swarm","simulation","distribut","cluster","test"],"s":["ruvix-qemu-swarm","ruvix","qemu","swarm","distribut","test"]},
+{"id":"ruvix-queue","family":"ruvix-queue","kind":"crate","store":"ruvector","version":null,"description":"io_uring-style ring buffer IPC for RuVix Cognition Kernel (ADR-087)","keywords":["ipc","kernel","queue","ring-buffer","zero-copy"],"source":"ruvector/crates/ruvix/crates/queue/Cargo.toml","sourceSha256":"436d0ac14d672dcc5075a9f2437d76f28246cb2a6abc3611d0df977a13d65fff","variants":[],"readme":"io_uring-style ring buffer IPC for the RuVix Cognition Kernel (ADR-087). This crate implements the Queue primitive from ADR-087 Section 7. All inter-task communication in RuVix goes through queues. There are no synchronous IPC calls, no shared memory without explicit region grants, and no signals. Queues use io_uring-style ring buffers with separate submission (SQ) and completion (CQ) queues: Main queue implementation: Lock-free ring buffer with atomic operations: Zero-copy message passing: When sender and receiver share a region, queue_send places a descriptor (offset + length) in the ring rather than copying bytes. This is critical for high-throughput vector streaming where copying 768-dim","readmeSource":"ruvector/crates/ruvix/crates/queue/README.md","product":"key:ruvector:ruvix-queue","canonical":"ruvix-queue","tier":"T0","t":["ruvix-queue","ruvix","queue","uring-style","uring","style","ring","buffer","ipc","cognition","kernel","adr-087","adr","087","ring-buffer","zero-copy","zero","copy"],"s":["ruvix-queue","ruvix","queue","ipc","kernel","ring-buffer","ring","buffer","zero-copy","zero","copy"]},
+{"id":"ruvix-region","family":"ruvix-region","kind":"crate","store":"ruvector","version":null,"description":"Memory region management for RuVix Cognition Kernel (ADR-087)","keywords":["allocator","kernel","memory","no_std","slab"],"source":"ruvector/crates/ruvix/crates/region/Cargo.toml","sourceSha256":"50cc173cfff3b7c6581c034c4c1ee534040585c83485381fb96c49058d6c8fb9","variants":[],"readme":"Memory region management for the RuVix Cognition Kernel (ADR-087). This crate implements the Region primitive. A region is a contiguous, capability-protected memory object with one of three policies. RuVix does not implement demand paging—all regions are physically backed at region_map time, eliminating page faults, swap, and copy-on-write complexity. Creates and manages regions: Fast fixed-size allocation with zero fragmentation: Append-only semantics for logs and event streams: Write-once semantics for constant data: From ADR-087 Section 5: - No shared memory without explicit grants: Regions require capability to access - TOCTOU protection: Slab regions cannot be used for zero-copy IPC des","readmeSource":"ruvector/crates/ruvix/crates/region/README.md","product":"key:ruvector:ruvix-region","canonical":"ruvix-region","tier":"T0","t":["ruvix-region","ruvix","region","memory","management","cognition","kernel","adr-087","adr","087","allocator","std","slab"],"s":["ruvix-region","ruvix","region","allocator","kernel","memory","std","slab"]},
+{"id":"ruvix-rpi-boot","family":"ruvix-rpi-boot","kind":"crate","store":"ruvector","version":null,"description":"Raspberry Pi boot support for RuVix Cognition Kernel (Phase D)","keywords":["boot","embedded","kernel","no_std","rpi"],"source":"ruvector/crates/ruvix/crates/rpi-boot/Cargo.toml","sourceSha256":"cd3a4f9eb20cb5798ac97d917b367c78fde8b2283f119b99a76cc22677ffa059","variants":[],"readme":"Raspberry Pi boot support for the RuVix Cognition Kernel (ADR-087 Phase D). This crate provides early boot support for Raspberry Pi 4/5, including: - DTB (Device Tree Blob) parsing - Early UART initialization - Secondary CPU wake via spin table - Boot configuration parsing - EL2 to EL1 transition 1. GPU ROM - Loads bootcode.bin from SD card 2. bootcode.bin - Initializes SDRAM, loads start4.elf 3. start4.elf - GPU firmware, parses config.txt, loads kernel8.img 4. kernel8.img - Our kernel starts executing When the kernel starts: CPU State: - MMU is OFF - D-cache is OFF - I-cache may be ON or OFF - Interrupts are masked (DAIF) - Running in EL2 (hypervisor) or EL1 (kernel) Required settings for","readmeSource":"ruvector/crates/ruvix/crates/rpi-boot/README.md","product":"key:ruvector:ruvix-rpi-boot","canonical":"ruvix-rpi-boot","tier":"T0","t":["ruvix-rpi-boot","ruvix","rpi","boot","raspberry","cognition","kernel","phase","embed","std"],"s":["ruvix-rpi-boot","ruvix","rpi","boot","embed","kernel","std"]},
+{"id":"ruvix-sched","family":"ruvix-sched","kind":"crate","store":"ruvector","version":null,"description":"Coherence-aware scheduler for the RuVix Cognition Kernel (ADR-087)","keywords":["coherence","kernel","no_std","real-time","scheduler"],"source":"ruvector/crates/ruvix/crates/sched/Cargo.toml","sourceSha256":"fa69eeaf2098669f0d24b75c09f23111ab5ce643509a5a6348ec1bda1edee75a","variants":[],"readme":"Coherence-aware scheduler for the RuVix Cognition Kernel (ADR-087). ruvix-sched implements a real-time scheduler that combines three scheduling signals to determine task priority: 1. Deadline Pressure (EDF): Tasks with tighter deadlines receive higher urgency 2. Novelty Signal: Tasks processing genuinely new information get a priority boost 3. Structural Risk: Tasks that might destabilize system coherence receive a penalty The priority formula from ADR-087 Section 5.1: - No-std compatible: Works in embedded and kernel environments - Partition scheduling: Guaranteed time slices per RVF mount - Capability-based: Integrates with ruvix-cap for access control - Zero allocation: Uses fixed-size ar","readmeSource":"ruvector/crates/ruvix/crates/sched/README.md","product":"key:ruvector:ruvix-sched","canonical":"ruvix-sched","tier":"T0","t":["ruvix-sched","ruvix","sched","coherence-aware","coherence","aware","scheduler","cognition","kernel","adr-087","adr","087","std","real"],"s":["ruvix-sched","ruvix","sched","coherence","kernel","std","real","scheduler"]},
+{"id":"ruvix-shell","family":"ruvix-shell","kind":"crate","store":"ruvector","version":null,"description":"In-kernel debug shell for the RuVix Cognition Kernel (ADR-087)","keywords":["debug","embedded","kernel","no_std","shell"],"source":"ruvector/crates/ruvix/crates/shell/Cargo.toml","sourceSha256":"b3ecac09705fee19e28a5f216c41a0bb528bd4f7e09edf42923febe98d463b7c","variants":[],"readme":"In-kernel debug shell for the RuVix Cognition Kernel (ADR-087). ruvix-shell provides an interactive command-line interface accessible over UART for runtime inspection and debugging of the RuVix kernel. - Command Parser: History support, argument parsing, whitespace handling - 13 Commands: help, info, mem, tasks, caps, vectors, witness, proofs, queues, perf, cpu, trace, reboot - Trace Mode: Enable/disable syscall tracing with visual indicator - no_std Compatible: Works in bare-metal environments with alloc MIT OR Apache-2.0","readmeSource":"ruvector/crates/ruvix/crates/shell/README.md","product":"key:ruvector:ruvix-shell","canonical":"ruvix-shell","tier":"T0","t":["ruvix-shell","ruvix","shell","in-kernel","kernel","debug","cognition","adr-087","adr","087","embed","std"],"s":["ruvix-shell","ruvix","shell","debug","embed","kernel","std"]},
+{"id":"ruvix-smp","family":"ruvix-smp","kind":"crate","store":"ruvector","version":null,"description":"Symmetric Multi-Processing support for RuVix Cognition Kernel (ADR-087 Phase C)","keywords":["kernel","multicore","no_std","smp","spinlock"],"source":"ruvector/crates/ruvix/crates/smp/Cargo.toml","sourceSha256":"4ff696d000d3f92d7631a0c0b31191a528119b2cfb2b5b2e2811303f52fb544e","variants":[],"readme":"Symmetric Multi-Processing support for the RuVix Cognition Kernel (ADR-087 Phase C). This crate provides the fundamental primitives for multi-core operation in the RuVix kernel: - no_std compatible with zero external dependencies - Up to 256 CPUs supported - Fair FIFO spinlock ordering via ticket algorithm - Power-efficient WFE/SEV synchronization - Cache-line aligned per-CPU data to prevent false sharing The primary CPU (CPU 0) starts first and initializes the system: Each secondary CPU runs this code after being released:","readmeSource":"ruvector/crates/ruvix/crates/smp/README.md","product":"key:ruvector:ruvix-smp","canonical":"ruvix-smp","tier":"T0","t":["ruvix-smp","ruvix","smp","symmetric","multi-processing","multi","process","cognition","kernel","adr-087","adr","087","phase","multicore","std","spinlock"],"s":["ruvix-smp","ruvix","smp","kernel","multicore","std","spinlock"]},
+{"id":"ruvix-vecgraph","family":"ruvix-vecgraph","kind":"crate","store":"ruvector","version":null,"description":"Kernel-resident vector and graph stores for RuVix Cognition Kernel (ADR-087)","keywords":["graph","hnsw","kernel","proof-gated","vector"],"source":"ruvector/crates/ruvix/crates/vecgraph/Cargo.toml","sourceSha256":"c8a202b8df4e3892ff2957137a89d4e7d16e3e6c6667ed4bdb2fca9a8a09d105","variants":[],"readme":"Kernel-resident vector and graph stores for the RuVix Cognition Kernel (ADR-087). This crate implements the Vector/Graph Kernel Objects from ADR-087 Section 4.3. Unlike conventional kernels where all data structures are userspace constructs, RuVix makes vector stores and graph stores kernel-resident objects. Main vector storage engine: Graph storage with mincut-aware mutations: HNSW index backed by slab allocation: Track coherence metadata for vectors: Record attestations for all mutations: Optimized distance computations: With the stats feature enabled:","readmeSource":"ruvector/crates/ruvix/crates/vecgraph/README.md","product":"key:ruvector:ruvix-vecgraph","canonical":"ruvix-vecgraph","tier":"T0","t":["ruvix-vecgraph","ruvix","vecgraph","kernel-resident","kernel","resident","vector","graph","store","cognition","adr-087","adr","087","hnsw","proof-gated","proof","gated"],"s":["ruvix-vecgraph","ruvix","vecgraph","graph","hnsw","kernel","proof-gated","proof","gated","vector"]},
+{"id":"ruvllm_retrieval_diffusion","family":"ruvllm_retrieval_diffusion","kind":"crate","store":"ruvector","version":null,"description":"Corpus-agnostic training-free retrieval LM and masked discrete diffusion built on ruvllm_sparse_attention. Works on any small-vocab token domain — game levels, drum patterns, configs, MIDI, visual tokens.","keywords":["diffusion","no-llm","pcg","retrieval","sparse-attention"],"source":"ruvector/crates/ruvllm_retrieval_diffusion/Cargo.toml","sourceSha256":"e9ad2bf0535550891e827b425afbb761f6cc33c953378bda08cd07bfa8ecdc87","variants":[],"readme":"Training-free retrieval LM and masked discrete diffusion that work on any small-vocab token domain — game levels, drum patterns, configs, MIDI loops, visual tokens. Built on the [ruvllm_sparse_attention](https://crates.io/crates/ruvllm_sparse_attention) kernel; no autograd, no learned weights, no Python in the loop. This is the corpus-agnostic generalisation of the [Sparse-Mario gist](https://gist.github.com/ruvnet/d3e0aaa7af2745b678a9eecddf610301). Two pipelines from one kernel, parameterised by a runtime RetrievalConfig (vocab size, embedding dim, mask sentinel, sampling controls): - Retriever::generate_fast — autoregressive next-token retrieval via KvCache + decode_step. O(log T) per gene","readmeSource":"ruvector/crates/ruvllm_retrieval_diffusion/README.md","product":"key:ruvector:ruvllm_retrieval_diffusion","canonical":"ruvllm_retrieval_diffusion","tier":"T0","t":["ruvllm","retrieval","diffusion","corpus-agnostic","corpus","agnostic","training-free","train","free","mask","discrete","built","sparse","attention","any","small-vocab","small","vocab","token","domain","game","level","drum","pattern","config","midi","visual","no-llm","pcg","sparse-attention"],"s":["ruvllm","retrieval","diffusion","no-llm","pcg","sparse-attention","sparse","attention"]},
+{"id":"ruvllm_sparse_attention","family":"ruvllm_sparse_attention","kind":"crate","store":"ruvector","version":null,"description":"Subquadratic O(N log N) sparse attention kernel for Rust LLM inference on edge devices, with optional FastGRNN salience gating for near-linear O(N) scaling","keywords":["attention","edge","inference","llm","transformer"],"source":"ruvector/crates/ruvllm_sparse_attention/Cargo.toml","sourceSha256":"de0023139431f50b4e9cb20c7db5177b7cadd40f6a629fc55bc9c42096393f7c","variants":[],"readme":"Subquadratic sparse attention for Rust LLM inference on edge devices — runs transformer attention in O(N log N) instead of O(N²), with an optional FastGRNN salience gate that pushes it to near-linear O(N). Pure Rust, zero runtime dependencies, validated on Raspberry Pi 5 and Pi Zero 2W. - [Why sparse attention?](#why-sparse-attention) - [What this crate gives you](#what-this-crate-gives-you) - [Quick start (60 seconds)](#quick-start-60-seconds) - [Near-linear with FastGRNN gating (new)](#near-linear-with-fastgrnn-gating-new) - [Supported platforms](#supported-platforms) - [Performance](#performance) - [Feature flags](#feature-flags) - [How it works](#how-it-works) - [FAQ](#faq) - [Sparse-Mar","readmeSource":"ruvector/crates/ruvllm_sparse_attention/README.md","product":"key:ruvector:ruvllm_sparse_attention","canonical":"ruvllm_sparse_attention","tier":"T0","t":["ruvllm","sparse","attention","subquadratic","log","kernel","inference","edge","device","optional","fastgrnn","salience","gating","near-linear","near","linear","scal","transformer"],"s":["ruvllm","sparse","attention","edge","inference","transformer"]},
+{"id":"ruvnet-brain","family":"ruvnet-brain","kind":"npm","store":"ruvnet-brain","version":"4.3.40","description":"One-command installer for RuvNet Brain — a portable, source-grounded brain over rUv's RuvNet building blocks, delivered as a Claude Code plugin so Claude uses the stack instead of fighting it.","keywords":[],"source":"ruvnet-brain/package.json","sourceSha256":"cfcab0725a3daa7a2c95b21a90b6a4ede71b4676dc9b82535ffb156a6f5f71de","variants":[],"readme":"Updated: 2026-09-11 08:00:00 EDT | Version 4.3.21 Created: 2026-06-29 22:36:38 EDT A portable, source-grounded brain over Reuven Cohen's (rUv's) RuvNet stack — delivered as a Claude Code plugin that makes Claude _use_ the stack instead of fighting it. The 4.3.10 candidate is an incremental reliability repair, not a completed North Star or a new 95-point certification. Publication is established by the protected release transaction and public installation receipt, not this source version. Historical release narratives and dated scores below are not measurements of the current candidate. The recovery covers shared QA producers, explicit continuation objectives, safer private-data handling, and","readmeSource":"ruvnet-brain/README.md","product":"repo:ruvnet-brain","canonical":"ruvnet-brain","tier":null,"t":["ruvnet-brain","brain","one-command","command","installer","portable","source-grounded","source","ground","block","deliver","claude","code","plugin","stack","fight"],"s":["ruvnet-brain","brain"]},
+{"id":"ruvnet-brain-kb","family":"ruvnet-brain-kb","kind":"npm","store":"ruvnet-brain","version":"4.3.40","description":"Self-contained RVF knowledge base for ruvnet-brain, built by rvf-kb-forge. Run `npm i` here, then use forge-ask.mjs (CLI) or forge-mcp.mjs (MCP stdio server). Vectors live in ruvnet-brain.rvf; full passage text in ruvnet-brain.passages.jsonl.","keywords":[],"source":"ruvnet-brain/kb/package.json","sourceSha256":"5c92b803cb4efd4726fe24867e567de07fa2e13ce88366485bcff3f374de58fa","variants":[],"product":"key:ruvnet-brain:ruvnet-brain-kb","canonical":"ruvnet-brain-kb","tier":null,"t":["ruvnet-brain-kb","brain","self-contained","self","contain","rvf","knowledge","base","ruvnet-brain","built","rvf-kb-forge","forge","npm","here","forge-ask.mjs","ask.mj","forge-mcp.mjs","mcp.mj","mcp","stdio","server","vector","live","ruvnet-brain.rvf","brain.rvf","full","passage","text","ruvnet-brain.passages.jsonl","brain.passages.jsonl"],"s":["ruvnet-brain-kb","brain"]},
+{"id":"rvagent-a2a","family":"rvagent-a2a","kind":"crate","store":"ruvector","version":null,"description":"rvAgent A2A — Agent2Agent peer-to-peer protocol (ADR-159)","keywords":[],"source":"ruvector/crates/rvAgent/rvagent-a2a/Cargo.toml","sourceSha256":"a3b8cf6480650fe11d3715f227e0322061639a4d7c3a0cd194b6b452fa4c7ddc","variants":[],"product":"key:ruvector:rvagent-a2a","canonical":"rvagent-a2a","tier":"T0","t":["rvagent-a2a","rvagent","a2a","agent2agent","peer-to-peer","peer","protocol","adr-159","adr","159"],"s":["rvagent-a2a","rvagent","a2a"]},
+{"id":"rvagent-acp","family":"rvagent-acp","kind":"crate","store":"ruvector","version":null,"description":"rvAgent ACP server — Agent Communication Protocol with auth, rate limiting, TLS","keywords":[],"source":"ruvector/crates/rvAgent/rvagent-acp/Cargo.toml","sourceSha256":"c2c49c5d76ff67d7614a1943aeacbbf1c6d08bf0cbcb9e1d23238e1762b68a93","variants":[],"product":"key:ruvector:rvagent-acp","canonical":"rvagent-acp","tier":"T0","t":["rvagent-acp","rvagent","acp","server","communication","protocol","auth","rate","limit","tls"],"s":["rvagent-acp","rvagent","acp"]},
+{"id":"rvagent-backends","family":"rvagent-backends","kind":"crate","store":"ruvector","version":null,"description":"rvAgent backends — filesystem, shell, composite, state, store, sandbox protocols","keywords":[],"source":"ruvector/crates/rvAgent/rvagent-backends/Cargo.toml","sourceSha256":"35f97fbdf12e8342f63ab6e7ba1bfe5a9dfab4c2e928f4f49a035e4a124f42f4","variants":[],"product":"key:ruvector:rvagent-backends","canonical":"rvagent-backends","tier":"T0","t":["rvagent-backends","rvagent","backend","filesystem","shell","composite","state","store","sandbox","protocol"],"s":["rvagent-backends","rvagent","backend"]},
+{"id":"rvagent-cli","family":"rvagent","kind":"crate","store":"ruvector","version":null,"description":"rvAgent CLI — terminal coding agent with TUI, session management, MCP tools","keywords":[],"source":"ruvector/crates/rvAgent/rvagent-cli/Cargo.toml","sourceSha256":"880405bb27ab822b989ef810cf4da3c7b7b5d4ed1a51584f93ffad91ee00380c","variants":["rvagent-core","rvagent-wasm"],"product":"key:ruvector:rvagent","canonical":"rvagent-cli","tier":"T0","t":["rvagent-cli","rvagent","terminal","coding","tui","session","management","mcp"],"s":["rvagent-cli","rvagent"]},
+{"id":"rvagent-mcp","family":"rvagent-mcp","kind":"crate","store":"ruvector","version":null,"description":"rvAgent MCP — Model Context Protocol tools, resources, and transport layer","keywords":[],"source":"ruvector/crates/rvAgent/rvagent-mcp/Cargo.toml","sourceSha256":"7705a2b3c89dc6ac2f5534e0297d502edfaa5f8bf39dbe4a557f4b3e01f9a11a","variants":[],"product":"key:ruvector:rvagent-mcp","canonical":"rvagent-mcp","tier":"T0","t":["rvagent-mcp","rvagent","mcp","model","context","protocol","resource","transport","layer"],"s":["rvagent-mcp","rvagent","mcp"]},
+{"id":"rvagent-middleware","family":"rvagent-middleware","kind":"crate","store":"ruvector","version":null,"description":"rvAgent middleware — pipeline, todolist, filesystem, subagents, summarization, memory, skills, prompt caching, HITL, witness, tool sanitizer","keywords":[],"source":"ruvector/crates/rvAgent/rvagent-middleware/Cargo.toml","sourceSha256":"b56baf2acc88529d51886b9f6c882bc7baac1388d225e7e939e382c8362a0126","variants":[],"product":"key:ruvector:rvagent-middleware","canonical":"rvagent-middleware","tier":"T0","t":["rvagent-middleware","rvagent","middleware","pipeline","todolist","filesystem","subagent","summarization","memory","skill","prompt","cach","hitl","witness","sanitizer"],"s":["rvagent-middleware","rvagent","middleware"]},
+{"id":"rvagent-subagents","family":"rvagent-subagents","kind":"crate","store":"ruvector","version":null,"description":"rvAgent subagents — spec compilation, builder, orchestration, result validation","keywords":[],"source":"ruvector/crates/rvAgent/rvagent-subagents/Cargo.toml","sourceSha256":"c396a81e923cef3c97a5506c8d27be471acb9b9775df9c153fd7ab48384d37de","variants":[],"product":"key:ruvector:rvagent-subagents","canonical":"rvagent-subagents","tier":"T0","t":["rvagent-subagents","rvagent","subagent","spec","compilation","builder","orchestration","result","validation"],"s":["rvagent-subagents","rvagent","subagent"]},
+{"id":"rvagent-tools","family":"rvagent-tools","kind":"crate","store":"ruvector","version":null,"description":"rvAgent tools — ls, read, write, edit, glob, grep, execute, todos, task (enum dispatch)","keywords":[],"source":"ruvector/crates/rvAgent/rvagent-tools/Cargo.toml","sourceSha256":"2438a4383f50e07025bcacda837607fad3c588e379c48567984b823dd08c4b7f","variants":[],"product":"key:ruvector:rvagent-tools","canonical":"rvagent-tools","tier":"T0","t":["rvagent-tools","rvagent","read","write","edit","glob","grep","execute","todo","task","enum","dispatch"],"s":["rvagent-tools","rvagent"]},
+{"id":"rvcsi-adapter-file","family":"rvcsi-adapter-file","kind":"crate","store":"rvcsi","version":null,"description":"rvCSI file/replay adapter — records and replays .rvcsi capture sessions deterministically (ADR-095 FR1/FR10, D9)","keywords":["csi","replay","rvcsi","wifi"],"source":"rvcsi/crates/rvcsi-adapter-file/Cargo.toml","sourceSha256":"8c9c6cfe280b3785a7d5f76e7bfa38e34d29e5a99c17e8795b03d4084cfaee3e","variants":[],"readme":"The .rvcsi capture container and a deterministic replay CsiSource for [rvCSI](https://github.com/ruvnet/rvcsi) (ADR-095 FR1/FR10, D9). A .rvcsi file is plain JSONL: the first line is a CaptureHeader (capture version, session/source ids, the source's AdapterProfile, the ValidationPolicy in force, calibration version, an opaque runtime-config blob, creation time); every subsequent line is one rvcsi_core::CsiFrame serialized as compact JSON. Simple, deterministic, append-friendly, debuggable with head / jq. - FileRecorder — create(path, &header) writes the header line; write_frame(&CsiFrame) appends one frame; finish() / flush() makes sure it hit disk. - FileReplayAdapter — a CsiSource that rep","readmeSource":"rvcsi/crates/rvcsi-adapter-file/README.md","product":"key:rvcsi:rvcsi-adapter-file","canonical":"rvcsi-adapter-file","tier":"T1","t":["rvcsi-adapter-file","rvcsi","adapter","file","replay","record","capture","session","deterministically","adr-095","adr","095","fr1","fr10","csi","wifi"],"s":["rvcsi-adapter-file","rvcsi","adapter","file","csi","replay","wifi"]},
+{"id":"rvcsi-adapter-nexmon","family":"rvcsi-adapter-nexmon","kind":"crate","store":"rvcsi","version":null,"description":"rvCSI Nexmon adapter — wraps the isolated napi-c shim that parses Nexmon CSI UDP/PCAP records into normalized CsiFrames (ADR-095 D2/D15, ADR-096)","keywords":["csi","nexmon","rvcsi","wifi"],"source":"rvcsi/crates/rvcsi-adapter-nexmon/Cargo.toml","sourceSha256":"e7069e527d48473fbc703ddc5e6fa4c470a617e4a810d470edf6412b14db8908","variants":[],"readme":"The Nexmon CSI adapter for [rvCSI](https://github.com/ruvnet/rvcsi) — and the napi-c seam: the only crate in the runtime with unsafe, and within it unsafe is confined to one ffi module wrapping a thin C shim (ADR-095 D2/D15, ADR-096). - native/rvcsi_nexmon_shim.{c,h} — the only C in the runtime, compiled via build.rs + cc. Allocation-free, global-free, bounds-checked against caller-supplied lengths, never panics, ABI-versioned (0x0001_0001). Handles two byte formats: the compact self-describing \"rvCSI Nexmon record\" ('RVNX' magic, version, flags, RSSI/noise, channel, bandwidth, timestamp, then int16 I/Q in Q8.8), and the real nexmon_csi UDP payload (the 18-byte magic 0x1111 · rssi int8 · fct","readmeSource":"rvcsi/crates/rvcsi-adapter-nexmon/README.md","product":"key:rvcsi:rvcsi-adapter-nexmon","canonical":"rvcsi-adapter-nexmon","tier":"T1","t":["rvcsi-adapter-nexmon","rvcsi","adapter","nexmon","wrap","isolat","napi-c","shim","pars","csi","udp","pcap","record","normaliz","csiframe","adr-095","adr","095","d15","adr-096","096","wifi"],"s":["rvcsi-adapter-nexmon","rvcsi","adapter","nexmon","csi","wifi"]},
+{"id":"rvcsi-dsp","family":"rvcsi-dsp","kind":"crate","store":"rvcsi","version":null,"description":"rvCSI DSP — reusable signal-processing stages (DC removal, phase unwrap, smoothing, Hampel, variance, baseline, motion energy, presence) (ADR-095 FR4)","keywords":["csi","dsp","rvcsi","wifi"],"source":"rvcsi/crates/rvcsi-dsp/Cargo.toml","sourceSha256":"bc3b42b3fe74dcab8b3bc206fcd89b2212e43bcec8d105127a33cb158ddf3131","variants":[],"readme":"The dependency-light DSP layer of [rvCSI](https://github.com/ruvnet/rvcsi) (ADR-095 FR4). Everything here is deterministic and hand-rolled — no rustfft, no ndarray. - stages — pure per-vector primitives on &[f32] / &mut [f32]: mean, variance, std_dev, median, remove_dc_offset, unwrap_phase (1-D phase unwrap), moving_average, ewma, hampel_filter / hampel_filter_count (MAD outlier rejection), short_window_variance, subtract_baseline. Failable stages report DspError. - features — frame/window-level scalar features: motion_energy / motion_energy_series, presence_score (logistic squash of motion energy), confidence_score (mean − 0.5·std of per-frame quality), and breathing_band_estimate — a heuri","readmeSource":"rvcsi/crates/rvcsi-dsp/README.md","product":"key:rvcsi:rvcsi-dsp","canonical":"rvcsi-dsp","tier":"T1","t":["rvcsi-dsp","rvcsi","dsp","reusable","signal-processing","signal","process","stage","removal","phase","unwrap","smooth","hampel","variance","baseline","motion","energy","presence","adr-095","adr","095","fr4","csi","wifi"],"s":["rvcsi-dsp","rvcsi","dsp","csi","wifi"]},
+{"id":"rvcsi-events","family":"rvcsi-events","kind":"crate","store":"rvcsi","version":null,"description":"rvCSI events — window aggregation + presence/motion/anomaly state machines producing CsiEvent (ADR-095 FR5)","keywords":["csi","events","rvcsi","wifi"],"source":"rvcsi/crates/rvcsi-events/Cargo.toml","sourceSha256":"fa57c236a7b05f9cd4ab2b3d18e8e945846f7139ff908b122c6c564e9f216b38","variants":[],"readme":"Window aggregation and event detection for [rvCSI](https://github.com/ruvnet/rvcsi) (ADR-095 FR5). - WindowBuffer — buffers exposable CsiFrames from one (session_id, source_id) and emits a CsiWindow every N frames or T nanoseconds: per-subcarrier mean amplitude + phase variance, scalar motion energy (mean RMS amplitude delta over consecutive frames), a logistic presence score, and the mean quality. Frames with a different subcarrier count from the window's first frame are skipped (mixed legacy/HT streams aren't aggregated together). - EventDetector trait + four state machines: - PresenceDetector — hysteresis on presence_score → PresenceStarted / PresenceEnded. - MotionDetector — debounced ed","readmeSource":"rvcsi/crates/rvcsi-events/README.md","product":"key:rvcsi:rvcsi-events","canonical":"rvcsi-events","tier":"T1","t":["rvcsi-events","rvcsi","event","window","aggregation","presence","motion","anomaly","state","machine","produc","csievent","adr-095","adr","095","fr5","csi","wifi"],"s":["rvcsi-events","rvcsi","event","csi","wifi"]},
+{"id":"rvcsi-runtime","family":"rvcsi-runtime","kind":"crate","store":"rvcsi","version":null,"description":"rvCSI runtime composition — wires a CsiSource + DSP + the event pipeline + RuVector export; the shared layer under rvcsi-node and rvcsi-cli (ADR-096)","keywords":["csi","runtime","rvcsi","wifi"],"source":"rvcsi/crates/rvcsi-runtime/Cargo.toml","sourceSha256":"1db4700c4e453f6e3e9008bcd88a2eef641f68aeb1aa511064954c2045ee68a0","variants":[],"readme":"The composition layer of [rvCSI](https://github.com/ruvnet/rvcsi) — no FFI of its own; the shared layer that rvcsi-node and rvcsi-cli are built on (ADR-096). This is the crate most applications want. - CaptureRuntime — wires a CsiSource + rvcsi_core::validate_frame + rvcsi_dsp::SignalPipeline + rvcsi_events::EventPipeline into one stream: next_validated_frame() (validated only), next_clean_frame() (validated and DSP-cleaned), and a clean-frame iterator that also drains events. - One-shot helpers for the offline path: summarize_capture (frame count, channels, subcarrier counts, quality, validation breakdown), decode_nexmon_records / decode_nexmon_pcap (+ _for, per-chip) / summarize_nexmon_pca","readmeSource":"rvcsi/crates/rvcsi-runtime/README.md","product":"key:rvcsi:rvcsi-runtime","canonical":"rvcsi-runtime","tier":"T1","t":["rvcsi-runtime","rvcsi","runtime","composition","wire","csisource","dsp","event","pipeline","export","shar","layer","rvcsi-node","rvcsi-cli","adr-096","adr","096","csi","wifi"],"s":["rvcsi-runtime","rvcsi","runtime","csi","wifi"]},
+{"id":"rvcsi-ruvector","family":"rvcsi-ruvector","kind":"crate","store":"rvcsi","version":null,"description":"rvCSI RuVector bridge — exports temporal RF embeddings + event metadata as a queryable RF-memory store (ADR-095 FR8, D8)","keywords":["csi","ruvector","rvcsi","wifi"],"source":"rvcsi/crates/rvcsi-ruvector/Cargo.toml","sourceSha256":"c672780aab50e4859e0f2ab31b31965e2c641a24055f099a1be61409bdd4bbb9","variants":[],"readme":"The [RuVector](https://github.com/ruvnet/ruvector) RF-memory bridge for [rvCSI](https://github.com/ruvnet/rvcsi) (ADR-095 FR8, D8) — CSI becomes far more valuable stored as temporal embeddings and room signatures. - Deterministic window_embedding and event_embedding — fixed-length vectors derived from a CsiWindow / CsiEvent (same input → same vector), so storage and similarity search are reproducible. - cosine_similarity on those embeddings. - The RfMemoryStore trait — store windows/events as embedding records, query by similarity, track a running baseline, detect drift. - InMemoryRfMemory and JsonlRfMemory — a usable standin until the production RuVector binding lands; JsonlRfMemory persist","readmeSource":"rvcsi/crates/rvcsi-ruvector/README.md","product":"key:rvcsi:rvcsi-ruvector","canonical":"rvcsi-ruvector","tier":"T1","t":["rvcsi-ruvector","rvcsi","bridge","export","temporal","embedding","event","metadata","queryable","rf-memory","memory","store","adr-095","adr","095","fr8","csi","wifi"],"s":["rvcsi-ruvector","rvcsi","csi","wifi"]},
+{"id":"rvf-adapter-agentdb","family":"rvf-adapter-agentdb","kind":"crate","store":"ruvector","version":null,"description":"AgentDB adapter for RuVector Format -- maps agent memory to RVF segments","keywords":[],"source":"ruvector/crates/rvf/rvf-adapters/agentdb/Cargo.toml","sourceSha256":"daa6129b219ed716e41427430905d195e3dde5d88c059da6481df8ac4ca537b5","variants":[],"product":"key:ruvector:rvf-adapter-agentdb","canonical":"rvf-adapter-agentdb","tier":"T0","t":["rvf-adapter-agentdb","rvf","adapter","agentdb","format","maps","memory","segment"],"s":["rvf-adapter-agentdb","rvf","adapter","agentdb"]},
+{"id":"rvf-adapter-agentic-flow","family":"rvf-adapter-agentic-flow","kind":"crate","store":"ruvector","version":null,"description":"Agentic-flow swarm adapter for RuVector Format -- maps inter-agent memory, coordination state, and learning patterns to RVF segments","keywords":[],"source":"ruvector/crates/rvf/rvf-adapters/agentic-flow/Cargo.toml","sourceSha256":"066d63a2b7557f362138b9fb009f35b4c4eb6af9ef029041ef1e674948f86204","variants":[],"product":"key:ruvector:rvf-adapter-agentic-flow","canonical":"rvf-adapter-agentic-flow","tier":"T0","t":["rvf-adapter-agentic-flow","rvf","adapter","flow","agentic-flow","swarm","format","maps","inter-agent","inter","memory","coordination","state","learn","pattern","segment"],"s":["rvf-adapter-agentic-flow","rvf","adapter","flow"]},
+{"id":"rvf-adapter-claude-flow","family":"rvf-adapter-claude-flow","kind":"crate","store":"ruvector","version":null,"description":"RVF adapter for claude-flow memory subsystem — stores memory entries as RVF files with WITNESS_SEG audit trails","keywords":[],"source":"ruvector/crates/rvf/rvf-adapters/claude-flow/Cargo.toml","sourceSha256":"613cf02d89af56f1cdc937bfcbd7dd9ea82e442f94f7a06000a431b55bb02566","variants":[],"product":"key:ruvector:rvf-adapter-claude-flow","canonical":"rvf-adapter-claude-flow","tier":"T0","t":["rvf-adapter-claude-flow","rvf","adapter","claude","flow","claude-flow","memory","subsystem","store","entry","file","witness","seg","audit","trail"],"s":["rvf-adapter-claude-flow","rvf","adapter","claude","flow"]},
+{"id":"rvf-adapter-ospipe","family":"rvf-adapter-ospipe","kind":"crate","store":"ruvector","version":null,"description":"OSpipe adapter for RuVector Format -- maps observation state vectors to RVF with META_SEG","keywords":[],"source":"ruvector/crates/rvf/rvf-adapters/ospipe/Cargo.toml","sourceSha256":"d792ac9f0220603c748d9ec2a5fa7cfc7bbd2c1bb63794b4d99f3f5e43bf6e00","variants":[],"product":"key:ruvector:rvf-adapter-ospipe","canonical":"rvf-adapter-ospipe","tier":"T0","t":["rvf-adapter-ospipe","rvf","adapter","ospipe","format","maps","observation","state","vector","meta","seg"],"s":["rvf-adapter-ospipe","rvf","adapter","ospipe"]},
+{"id":"rvf-adapter-rvlite","family":"rvf-adapter-rvlite","kind":"crate","store":"ruvector","version":null,"description":"Lightweight embedded vector store adapter for RuVector Format -- minimal API over RVF Core Profile","keywords":[],"source":"ruvector/crates/rvf/rvf-adapters/rvlite/Cargo.toml","sourceSha256":"27f7626b9b133a22a689992c3cd90ac28d35d701f757adc6b5f449f8b5e460fd","variants":[],"product":"key:ruvector:rvf-adapter-rvlite","canonical":"rvf-adapter-rvlite","tier":"T0","t":["rvf-adapter-rvlite","rvf","adapter","rvlite","lightweight","embed","vector","store","format","minimal","core","profile"],"s":["rvf-adapter-rvlite","rvf","adapter","rvlite"]},
+{"id":"rvf-adapter-sona","family":"rvf-adapter-sona","kind":"crate","store":"ruvector","version":null,"description":"SONA adapter for RuVector Format -- stores learning trajectories, neural patterns, and experience replay buffers as RVF segments","keywords":[],"source":"ruvector/crates/rvf/rvf-adapters/sona/Cargo.toml","sourceSha256":"1d24f1bafdc134fa1bc7774ea8ee7e9f98bbaadde92501a9fe2e7d09d79f3fcb","variants":[],"product":"key:ruvector:rvf-adapter-sona","canonical":"rvf-adapter-sona","tier":"T0","t":["rvf-adapter-sona","rvf","adapter","sona","format","store","learn","trajectory","neural","pattern","experience","replay","buffer","segment"],"s":["rvf-adapter-sona","rvf","adapter","sona"]},
+{"id":"rvf-crypto","family":"rvf-crypto","kind":"crate","store":"ruvector","version":null,"description":"RuVector Format cryptographic primitives -- SHA-3 hashing and Ed25519 signing","keywords":["crypto","ed25519","rvf","sha3","vector"],"source":"ruvector/crates/rvf/rvf-crypto/Cargo.toml","sourceSha256":"026c7848aa86f2510deeae8a73328d02d7714dbee045442eb1425d9b7d7eebda","variants":[],"readme":"Tamper-proof hashing and signing for every RVF segment -- SHA-3 digests, Ed25519 signatures, and lineage witness chains. Every operation on an RVF file gets recorded in a cryptographic witness chain. rvf-crypto provides the primitives that make this possible: SHA-3 (SHAKE-256) content hashing for segment identity, Ed25519 digital signatures for provenance, and lineage verification functions that ensure no record in the chain has been altered. If you are building tools that read, write, or transform .rvf files, this crate handles all the cryptography so you do not have to. For no_std or WASM targets that only need hashing and witness chains (no signing), disable defaults:","readmeSource":"ruvector/crates/rvf/rvf-crypto/README.md","product":"key:ruvector:rvf-crypto","canonical":"rvf-crypto","tier":"T0","t":["rvf-crypto","rvf","crypto","format","cryptographic","primitive","sha-3","sha","hash","ed25519","sign","sha3","vector"],"s":["rvf-crypto","rvf","crypto","ed25519","sha3","vector"]},
+{"id":"rvf-ebpf","family":"rvf-ebpf","kind":"crate","store":"ruvector","version":null,"description":"Real eBPF programs for RVF vector distance computation","keywords":["bpf","ebpf","rvf","vector-acceleration","xdp"],"source":"ruvector/crates/rvf/rvf-ebpf/Cargo.toml","sourceSha256":"a1af8f49e120e5fedfffedf90dbcf3115fc59fc3c10df759491d0560826b4d29","variants":[],"readme":"Real eBPF program compiler and embedder for RVF cognitive containers. rvf-ebpf compiles real BPF C programs with clang and embeds them into .rvf files as EBPF_SEG segments. These programs provide kernel-level acceleration for vector operations. - clang with BPF target support (for compilation) - Programs can also be pre-compiled and embedded as raw ELF bytes MIT OR Apache-2.0","readmeSource":"ruvector/crates/rvf/rvf-ebpf/README.md","product":"key:ruvector:rvf-ebpf","canonical":"rvf-ebpf","tier":"T0","t":["rvf-ebpf","rvf","ebpf","real","program","vector","distance","computation","bpf","vector-acceleration","acceleration","xdp"],"s":["rvf-ebpf","rvf","ebpf","bpf","vector-acceleration","vector","acceleration","xdp"]},
+{"id":"rvf-federation","family":"rvf-federation","kind":"crate","store":"ruvector","version":null,"description":"Federated RVF transfer learning -- PII stripping, differential privacy, federated averaging","keywords":["differential-privacy","federated-learning","rvf","transfer-learning"],"source":"ruvector/crates/rvf/rvf-federation/Cargo.toml","sourceSha256":"1ad1d3b82c84d3639a3a548afd9fc276b55123079108704b3bba4b3b1c44b3e8","variants":[],"readme":"Privacy-preserving federated transfer learning for the RVF format. RuVector users independently accumulate learning patterns -- SONA weight trajectories, policy kernel configurations, domain expansion priors, HNSW tuning parameters. Today that learning is siloed. rvf-federation implements the inter-user federation layer defined in [ADR-057](../../../docs/adr/ADR-057-federated-rvf-transfer-learning.md): it strips PII, injects differential privacy noise, packages transferable learning as RVF segments, and merges incoming learning with formal privacy guarantees.","readmeSource":"ruvector/crates/rvf/rvf-federation/README.md","product":"key:ruvector:rvf-federation","canonical":"rvf-federation","tier":"T0","t":["rvf-federation","rvf","federation","federat","transfer","learn","pii","stripp","differential","privacy","averag","differential-privacy","federated-learning","transfer-learning"],"s":["rvf-federation","rvf","federation","differential-privacy","differential","privacy","federated-learning","federat","learn","transfer-learning","transfer"]},
+{"id":"rvf-forge-core","family":"rvf-forge","kind":"crate","store":"ruvector","version":null,"description":"RVForge packaging and verification library: inspection-only RVF reading, Ed25519 verification, canonical build manifests, and provenance (ADR-283, ADR-290, ADR-291)","keywords":[],"source":"ruvector/crates/rvf-forge-core/Cargo.toml","sourceSha256":"75a36b9efeeebc71fc8e7a57780ce4c1aaae7816dd4c35334e493a74b299a44e","variants":[],"product":"key:ruvector:rvf-forge","canonical":"rvf-forge-core","tier":"T0","t":["rvf-forge-core","rvf","forge","core","rvf-forge","rvforge","packag","verification","inspection-only","inspection","only","read","ed25519","canonical","manifest","provenance","adr-283","adr","283","adr-290","290","adr-291","291"],"s":["rvf-forge-core","rvf","forge","core","rvf-forge"]},
+{"id":"rvf-import","family":"rvf-import","kind":"crate","store":"ruvector","version":null,"description":"Import tools for migrating data from JSON, CSV, and NumPy formats into RVF stores","keywords":["csv","import","json","numpy","rvf"],"source":"ruvector/crates/rvf/rvf-import/Cargo.toml","sourceSha256":"8d69e563ee9541a4e37e66ba63791c4f29084d755fbfdec9b84b6b4e14d84e39","variants":[],"readme":"Data import tools for migrating vectors from JSON, CSV, and NumPy formats into RVF stores. rvf-import provides both a library API and a CLI binary for importing vector data from common formats into .rvf files. Supports automatic ID generation, metadata extraction, and batch ingestion. MIT OR Apache-2.0","readmeSource":"ruvector/crates/rvf/rvf-import/README.md","product":"key:ruvector:rvf-import","canonical":"rvf-import","tier":"T0","t":["rvf-import","rvf","import","migrat","json","csv","numpy","format","store"],"s":["rvf-import","rvf","import","csv","json","numpy"]},
+{"id":"rvf-index","family":"rvf-index","kind":"crate","store":"ruvector","version":null,"description":"RuVector Format progressive HNSW indexing with Layer A/B/C tiered search","keywords":["hnsw","indexing","nearest-neighbor","rvf","vector"],"source":"ruvector/crates/rvf/rvf-index/Cargo.toml","sourceSha256":"2ff66ba254edb9ef71d1603c7391cc2e6f8c119afbc8cb8d396358c16573ab17","variants":[],"readme":"Progressive HNSW indexing with tiered Layer A/B/C search for RuVector Format. rvf-index implements a Hierarchical Navigable Small World (HNSW) index optimized for the RVF storage model: - Layer A -- hot vectors, full-precision, in-memory graph - Layer B -- warm vectors, quantized, memory-mapped - Layer C -- cold vectors, compressed, on-disk with lazy loading - Progressive build -- index grows incrementally without full rebuilds - Vamana alpha-pruning -- diversity-aware neighbor selection during build (recall@10 0.986 -> 0.996 at ef_search=30, measured at 100k x 64-dim, Windows x64 criterion release) - Hardened codec -- INDEX_SEG decoding validates lengths and counts against the payload size","readmeSource":"ruvector/crates/rvf/rvf-index/README.md","product":"key:ruvector:rvf-index","canonical":"rvf-index","tier":"T0","t":["rvf-index","rvf","index","format","progressive","hnsw","layer","tier","search","nearest-neighbor","nearest","neighbor","vector"],"s":["rvf-index","rvf","index","hnsw","nearest-neighbor","nearest","neighbor","vector"]},
+{"id":"rvf-kernel","family":"rvf-kernel","kind":"crate","store":"ruvector","version":null,"description":"Real Linux microkernel builder for RVF cognitive containers","keywords":["cognitive-container","initramfs","kernel","microvm","rvf"],"source":"ruvector/crates/rvf/rvf-kernel/Cargo.toml","sourceSha256":"a1689648a9d76fa627c0dfb5af7e816b54e78ea22e38fb31b78d3689ca4430a1","variants":[],"readme":"Real Linux microkernel builder for RVF cognitive containers. rvf-kernel builds production-grade Linux kernel images and initramfs archives for embedding into .rvf files. A single .rvf file with KERNEL_SEG boots as a standalone Linux microservice on QEMU, Firecracker, or bare metal. - VirtIO PCI/BLK/NET for VM I/O - BPF + JIT for eBPF programs - KASLR + stack protector for security - No modules, USB, DRM, or wireless (minimal attack surface) - ~1.5 MB bzImage, ~512 KB initramfs MIT OR Apache-2.0","readmeSource":"ruvector/crates/rvf/rvf-kernel/README.md","product":"key:ruvector:rvf-kernel","canonical":"rvf-kernel","tier":"T0","t":["rvf-kernel","rvf","kernel","real","linux","microkernel","builder","cognitive","container","cognitive-container","initramf","microvm"],"s":["rvf-kernel","rvf","kernel","cognitive-container","cognitive","container","initramf","microvm"]},
+{"id":"rvf-launch","family":"rvf-launch","kind":"crate","store":"ruvector","version":null,"description":"QEMU microVM launcher for RVF cognitive containers","keywords":["cognitive-container","launcher","microvm","qemu","rvf"],"source":"ruvector/crates/rvf/rvf-launch/Cargo.toml","sourceSha256":"ebb7d787a25d656fd8c318bd8b721b050337c77d88b776bd4fd89686b5254bc9","variants":[],"readme":"QEMU microVM launcher for RVF cognitive containers. rvf-launch boots an .rvf file as a standalone Linux microservice inside a QEMU microVM. It extracts the embedded kernel from KERNEL_SEG, launches QEMU with virtio I/O, and provides a programmatic API for lifecycle management. - QEMU installed (qemu-system-x86_64 or qemu-system-aarch64) - KVM optional (falls back to TCG emulation) MIT OR Apache-2.0","readmeSource":"ruvector/crates/rvf/rvf-launch/README.md","product":"key:ruvector:rvf-launch","canonical":"rvf-launch","tier":"T0","t":["rvf-launch","rvf","launch","qemu","microvm","launcher","cognitive","container","cognitive-container"],"s":["rvf-launch","rvf","launch","cognitive-container","cognitive","container","launcher","microvm","qemu"]},
+{"id":"rvf-manifest","family":"rvf-manifest","kind":"crate","store":"ruvector","version":null,"description":"RuVector Format two-level manifest system for segment tracking and compaction","keywords":["compaction","database","manifest","rvf","vector"],"source":"ruvector/crates/rvf/rvf-manifest/Cargo.toml","sourceSha256":"59429b127112576379347642a66ef96ea80ff0f8953a4f9ff06e56fc25cc5ceb","variants":[],"readme":"Two-level manifest system for tracking RVF segments and coordinating compaction. rvf-manifest manages the metadata layer that tracks which segments are active, tombstoned, or pending compaction: - Level-0 manifest -- per-file segment directory with CRC32c integrity - Level-1 manifest -- cross-file global index for multi-file stores - Compaction coordination -- tracks merge candidates and generation numbers - std (default) -- enable std I/O support The FileIdentity struct (68 bytes) is stored at offset 0xF00 within the Level0Root reserved area (252 bytes starting at the end of the signature region). This placement is backward compatible: old readers that ignore the reserved area see zeros and","readmeSource":"ruvector/crates/rvf/rvf-manifest/README.md","product":"key:ruvector:rvf-manifest","canonical":"rvf-manifest","tier":"T0","t":["rvf-manifest","rvf","manifest","format","two-level","level","segment","track","compaction","database","vector"],"s":["rvf-manifest","rvf","manifest","compaction","database","vector"]},
+{"id":"rvf-quant","family":"rvf-quant","kind":"crate","store":"ruvector","version":null,"description":"RuVector Format temperature-tiered vector quantization (f32/f16/u8/binary)","keywords":["compression","embedding","quantization","rvf","vector"],"source":"ruvector/crates/rvf/rvf-quant/Cargo.toml","sourceSha256":"599359dc4362f5712750f7a4a6823d4c54327192a16bfc7f5f900db8345d806b","variants":[],"readme":"Temperature-tiered vector quantization for RuVector Format. rvf-quant provides quantization codecs that reduce vector storage size based on access temperature: - f32 -- full precision for hot vectors - f16 -- half precision for warm vectors - u8 -- scalar quantization for cool vectors - binary -- 1-bit quantization for cold/archive vectors - RaBitQ -- randomized 1-bit codes (32x code compression) with unbiased distance estimation, used by the runtime's opt-in two-stage query path (recall@10 0.972 after exact rescore at 100k x 64-dim, Windows x64 criterion release) - Automatic tiering -- promote/demote vectors based on access patterns - decode_sketch_seg now returns Result instead of panickin","readmeSource":"ruvector/crates/rvf/rvf-quant/README.md","product":"key:ruvector:rvf-quant","canonical":"rvf-quant","tier":"T0","t":["rvf-quant","rvf","quant","format","temperature-tiered","temperature","tier","vector","quantization","f32","f16","binary","compression","embedd"],"s":["rvf-quant","rvf","quant","compression","embedd","quantization","vector"]},
+{"id":"rvf-runtime","family":"rvf-runtime","kind":"crate","store":"ruvector","version":null,"description":"RuVector Format runtime -- RvfStore API, compaction, and streaming I/O","keywords":["compaction","database","rvf","storage","vector"],"source":"ruvector/crates/rvf/rvf-runtime/Cargo.toml","sourceSha256":"6c245550f5a5617d04c01045841d32537c35a5b7bcee09e8040cac97738a0e46","variants":[],"readme":"RuVector Format runtime providing the RvfStore API, background compaction, and streaming I/O. rvf-runtime is the main entry point for applications that want to read and write RVF files: - RvfStore -- high-level API for storing and retrieving vectors - HNSW query path -- queries route through the persisted HNSW index when one is available - Compaction -- background merge of segments to reclaim space - Streaming I/O -- append-only writes with configurable flush policy - std (default) -- enable std I/O support - wasm -- enable WASM-compatible runtime paths RvfStore::query routes through the persisted HNSW index when an INDEX_SEG is present in the file; otherwise it falls back to an exact brute-","readmeSource":"ruvector/crates/rvf/rvf-runtime/README.md","product":"key:ruvector:rvf-runtime","canonical":"rvf-runtime","tier":"T0","t":["rvf-runtime","rvf","runtime","format","rvfstore","compaction","stream","database","storage","vector"],"s":["rvf-runtime","rvf","runtime","compaction","database","storage","vector"]},
+{"id":"rvf-server","family":"rvf-server","kind":"crate","store":"ruvector","version":null,"description":"RuVector Format TCP/HTTP streaming server with REST API","keywords":["database","http","rvf","server","vector"],"source":"ruvector/crates/rvf/rvf-server/Cargo.toml","sourceSha256":"68096c25b88fa8e1337b0c770bc1f653c25ab0edbf7972baa42e6e7d398e9f7e","variants":[],"readme":"TCP/HTTP streaming server exposing the RuVector Format runtime via a REST API. rvf-server wraps rvf-runtime in an Axum-based HTTP server for networked vector operations: - REST API -- CRUD endpoints for vectors and collections - Streaming -- chunked transfer for large batch imports/exports - CLI -- configurable via clap command-line arguments MIT OR Apache-2.0","readmeSource":"ruvector/crates/rvf/rvf-server/README.md","product":"key:ruvector:rvf-server","canonical":"rvf-server","tier":"T0","t":["rvf-server","rvf","server","format","tcp","http","stream","rest","database","vector"],"s":["rvf-server","rvf","server","database","http","vector"]},
+{"id":"rvf-wire","family":"rvf-wire","kind":"crate","store":"ruvector","version":null,"description":"RuVector Format wire format reader/writer -- zero-copy segment serialization","keywords":["binary-format","database","rvf","vector","wire-protocol"],"source":"ruvector/crates/rvf/rvf-wire/Cargo.toml","sourceSha256":"7ad03ca139033cabf83e4268b6196bb6d1223db861c5117b90a98df6e841afbd","variants":[],"readme":"Zero-copy wire format reader and writer for RuVector Format (RVF) segments. rvf-wire handles serialization and deserialization of RVF binary segments: - Writer -- append segments with automatic CRC32c and XXH3 checksums - Reader -- stream-parse segments with validation and integrity checks - Zero-copy -- borrows directly from memory-mapped buffers where possible - std (default) -- enable std I/O support MIT OR Apache-2.0","readmeSource":"ruvector/crates/rvf/rvf-wire/README.md","product":"key:ruvector:rvf-wire","canonical":"rvf-wire","tier":"T0","t":["rvf-wire","rvf","wire","format","reader","writer","zero-copy","zero","copy","segment","serialization","binary-format","binary","database","vector","wire-protocol","protocol"],"s":["rvf-wire","rvf","wire","binary-format","binary","format","database","vector","wire-protocol","protocol"]},
+{"id":"rvface-cli","family":"rvface","kind":"crate","store":"rvface","version":null,"description":"rvFACE native CLI: detect faces, compare identities, benchmark","keywords":[],"source":"rvface/crates/rvface-cli/Cargo.toml","sourceSha256":"792312c6742ded9caeeda5c70133d19b0ae309d33580730e393d4e260b8cc8dd","variants":["rvface-core","rvface-wasm"],"product":"repo:rvface","canonical":"rvface-cli","tier":null,"t":["rvface-cli","rvface","detect","face","compare","identity","benchmark"],"s":["rvface-cli","rvface"]},
+{"id":"rvface-models","family":"rvface-models","kind":"crate","store":"rvface","version":null,"description":"Burn ports of the rvFACE networks: slim-320 SSD detector, MobileFaceNet landmarks, IRN-50 embedder","keywords":[],"source":"rvface/crates/rvface-models/Cargo.toml","sourceSha256":"e08545d4d8f2542d9f83d8f095d1282d09f180b6ac20fa894be0dd03b2dad6d2","variants":[],"product":"key:rvface:rvface-models","canonical":"rvface-models","tier":null,"t":["rvface-models","rvface","model","burn","port","network","slim-320","slim","320","ssd","detector","mobilefacenet","landmark","irn-50","irn","embedder"],"s":["rvface-models","rvface","model"]},
+{"id":"rvforge-reader","family":"rvforge-reader","kind":"crate","store":"ruvector","version":null,"description":"RVForge Reader — desktop reader for signed .rvf agent packages (ADR-289)","keywords":[],"source":"ruvector/crates/rvforge-reader/Cargo.toml","sourceSha256":"330bc88a9e4f6779b525799509609560bafda6e235e5f5be0ff2d61d5dd64f0a","variants":[],"readme":"Tauri v2 desktop reader for signed .rvf agent packages — the host described in [ADR-289](../../docs/adr/ADR-289-desktop-host-adapters.md). RVF reading, verification, capability derivation, runtime selection, and encrypted state capsules are real; installing and running an agent are not, because they need rvm-ffi, which does not exist yet. The crate is a standalone workspace (its Cargo.toml carries an empty [workspace] table) and is listed in the repo root's workspace.exclude. A Tauri app pulls a large, app-specific dependency graph — webview bindings, bundler, plugin build scripts — that every cargo build --workspace in the parent repo would otherwise pay for. Tauri is behind an optional des","readmeSource":"ruvector/crates/rvforge-reader/README.md","product":"key:ruvector:rvforge-reader","canonical":"rvforge-reader","tier":"T0","t":["rvforge-reader","rvforge","reader","desktop","sign","rvf","adr-289","adr","289"],"s":["rvforge-reader","rvforge","reader"]},
+{"id":"rvforge-registry","family":"rvforge-registry","kind":"crate","store":"ruvector","version":null,"description":"RVForge file-backed content-addressed release registry: typed objects, publish rules, trust levels, revocation, and an append-only Merkle transparency log (ADR-294)","keywords":[],"source":"ruvector/crates/rvforge-registry/Cargo.toml","sourceSha256":"6df829db393cdf1fe64e4aedfb4f368894d3157495df4f51a31c28e353116569","variants":[],"product":"key:ruvector:rvforge-registry","canonical":"rvforge-registry","tier":"T0","t":["rvforge-registry","rvforge","registry","file-backed","file","back","content-addressed","content","addres","release","typed","object","publish","rule","trust","level","revocation","append-only","append","only","merkle","transparency","log","adr-294","adr","294"],"s":["rvforge-registry","rvforge","registry"]},
+{"id":"rvlite","family":"rvlite","kind":"crate","store":"ruvector","version":null,"description":"Standalone vector database with SQL, SPARQL, and Cypher - powered by RuVector WASM","keywords":[],"source":"ruvector/crates/rvlite/Cargo.toml","sourceSha256":"1c474783d18a2fea75bc889a3a5d37063078ca6647a214e62a41b8d5960f6ea5","variants":[],"readme":"Status: Proof of Concept (v0.1.0) RvLite is a lightweight, standalone vector database that runs entirely in WebAssembly. It provides SQL, SPARQL, and Cypher query interfaces, along with graph neural networks and self-learning capabilities. A complete vector database that runs anywhere JavaScript runs: - ✅ Browsers (Chrome, Firefox, Safari, Edge) - ✅ Node.js - ✅ Deno - ✅ Bun - ✅ Cloudflare Workers - ✅ Vercel Edge Functions RvLite is a thin orchestration layer over battle-tested WASM crates: This is a proof of concept to validate: - ✅ Basic WASM compilation with ruvector-core - ✅ WASM bindings setup (wasm-bindgen) - ⏳ Integration with other WASM crates (pending) - ⏳ Bundle size measurement (pe","readmeSource":"ruvector/crates/rvlite/README.md","product":"key:ruvector:rvlite","canonical":"rvlite","tier":"T0","t":["rvlite","standalone","vector","database","sql","sparql","cypher","power"],"s":["rvlite"]},
+{"id":"rvm-anchor","family":"rvm-anchor","kind":"crate","store":"rvm","version":null,"description":"Cross-repo receipt anchoring: verify a ruflo ADR-322C evaluation receipt and anchor a commitment to it into the RVM witness chain (ADR-156)","keywords":["anchoring","no_std","receipt","verification","witness"],"source":"rvm/crates/rvm-anchor/Cargo.toml","sourceSha256":"9ca03d39ea5c2946393af37d20d409531ed9c1ba8bb7cebcaadcceedade0de6f","variants":[],"product":"key:rvm:rvm-anchor","canonical":"rvm-anchor","tier":"T1","t":["rvm-anchor","rvm","anchor","cross-repo","cross","repo","receipt","verify","ruflo","adr-322c","adr","322c","evaluation","commitment","witness","chain","adr-156","156","std","verification"],"s":["rvm-anchor","rvm","anchor","std","receipt","verification","witness"]},
+{"id":"rvm-boot","family":"rvm-boot","kind":"crate","store":"rvm","version":null,"description":"Boot sequence and initialization for the RVM microhypervisor (ADR-140)","keywords":["boot","hypervisor","initialization","no_std"],"source":"rvm/crates/rvm-boot/Cargo.toml","sourceSha256":"e0ad9270caab3b26a46bcf88b21b8d54e5652c9cfbcfea095582b40701a0be42","variants":[],"readme":"Deterministic 7-phase boot sequence for the RVM microhypervisor. Implements ADR-140: each boot phase is gated by a witness entry and must complete before the next phase begins. Out-of-order phase completion is rejected. The BootTracker provides a simple state machine that enforces the required sequencing. - BootPhase -- enum of 7 phases (HalInit through Handoff) - BootTracker -- state machine enforcing sequential phase completion - new() -- starts at HalInit - complete_phase(phase) -- marks current phase done, advances to next - is_complete() -- true when all 7 phases have completed - current_phase() -- returns the current phase, or None if complete - DC-15: #![no_std], #![forbid(unsafe_code","readmeSource":"rvm/crates/rvm-boot/README.md","product":"key:rvm:rvm-boot","canonical":"rvm-boot","tier":"T1","t":["rvm-boot","rvm","boot","sequence","initialization","microhypervisor","adr-140","adr","140","hypervisor","std"],"s":["rvm-boot","rvm","boot","hypervisor","initialization","std"]},
+{"id":"rvm-cap","family":"rvm-cap","kind":"crate","store":"rvm","version":null,"description":"Capability system for RVM with P1/P2 proof verification (ADR-135)","keywords":["capability","hypervisor","no_std","security"],"source":"rvm/crates/rvm-cap/Cargo.toml","sourceSha256":"5cc844d56cbbb0bc18d2c5131856e6747493c882a71ab2492b9a641ef3624d23","variants":[],"readme":"Capability-based access control with derivation trees and tiered proof verification. Implements the three-layer proof system from ADR-135. Capabilities are unforgeable kernel-managed tokens with a rights bitmap. Derivation trees enforce monotonic attenuation: a partition can only grant capabilities it holds, and granted rights must be equal or fewer. Delegation depth is bounded at 8 levels. - CapabilityManager -- central manager: issue, derive, revoke, verify - CapabilityTable -- per-partition capability slot table (default 256 slots) - DerivationTree, DerivationNode -- parent-child derivation tracking - GrantPolicy -- grant policy with GRANT_ONCE non-transitive delegation - RevokeResult --","readmeSource":"rvm/crates/rvm-cap/README.md","product":"key:rvm:rvm-cap","canonical":"rvm-cap","tier":"T1","t":["rvm-cap","rvm","cap","capability","proof","verification","adr-135","adr","135","hypervisor","std","security"],"s":["rvm-cap","rvm","cap","capability","hypervisor","std","security"]},
+{"id":"rvm-checkpoint","family":"rvm-checkpoint","kind":"crate","store":"ruvector","version":null,"description":"C2SP tlog-checkpoint export for RVM sealed witness segments (host-side, std)","keywords":["checkpoint","ed25519","signed-note","transparency-log","witness"],"source":"ruvector/crates/rvm/crates/rvm-checkpoint/Cargo.toml","sourceSha256":"1c961693d39f0aecf6776ced4fdb66fbff2ee944a89d37cb8625188d653eb0cd","variants":[],"product":"key:ruvector:rvm-checkpoint","canonical":"rvm-checkpoint","tier":"T0","t":["rvm-checkpoint","rvm","checkpoint","c2sp","tlog-checkpoint","tlog","export","seal","witness","segment","host-side","host","side","std","ed25519","signed-note","sign","note","transparency-log","transparency","log"],"s":["rvm-checkpoint","rvm","checkpoint","ed25519","signed-note","sign","note","transparency-log","transparency","log","witness"]},
+{"id":"rvm-coherence","family":"rvm-coherence","kind":"crate","store":"rvm","version":null,"description":"Coherence monitoring and Phi computation for the RVM microhypervisor (ADR-139)","keywords":["coherence","hypervisor","iit","no_std","phi"],"source":"rvm/crates/rvm-coherence/Cargo.toml","sourceSha256":"811ffb7444de25f759760e12c622c062e0b8c911afb03250bd4fc58e5b64dbd9","variants":[],"readme":"Real-time coherence scoring and Phi computation for the RVM microhypervisor. Coherence is the first-class resource-allocation signal: partitions with higher coherence receive more CPU time and memory grants. Raw Phi values from IIT (Integrated Information Theory) sensors are fed through an EMA (Exponential Moving Average) filter using fixed-point arithmetic to produce smoothed coherence scores. The optional sched feature enables direct feedback to the coherence-weighted scheduler. - EmaFilter -- fixed-point EMA filter (basis points, no floating-point) - SensorReading -- raw Phi reading with partition ID and timestamp - phi_to_coherence_bp(phi) -- convert raw Phi to basis-point coherence (stu","readmeSource":"rvm/crates/rvm-coherence/README.md","product":"key:rvm:rvm-coherence","canonical":"rvm-coherence","tier":"T1","t":["rvm-coherence","rvm","coherence","monitor","phi","computation","microhypervisor","adr-139","adr","139","hypervisor","iit","std"],"s":["rvm-coherence","rvm","coherence","hypervisor","iit","std","phi"]},
+{"id":"rvm-context","family":"rvm-context","kind":"crate","store":"rvm","version":null,"description":"Capability-gated ruv context namespace for RVM","keywords":["capability","context","no_std","rvf","rvm","uri","wasm"],"source":"rvm/crates/rvm-context/Cargo.toml","sourceSha256":"b406b0498344431e0c596c5fa1b9dcbff7bd5282de571644892198b6ffaa26a3","variants":["rvm-context-wasm"],"product":"key:rvm:rvm-context","canonical":"rvm-context","tier":"T1","t":["rvm-context","rvm","context","capability-gated","capability","gated","namespace","std","rvf","uri"],"s":["rvm-context","rvm","context","capability","std","rvf","uri"]},
+{"id":"rvm-context-service","family":"rvm-context-service","kind":"crate","store":"rvm","version":null,"description":"Durable encrypted RuVector-backed ruv context service","keywords":["context","encryption","redb","ruvector"],"source":"rvm/crates/rvm-context-service/Cargo.toml","sourceSha256":"3390d0ee26191fe9e9d1002a4eb0cd30f80dd547ddbba487ac15f668c03896cd","variants":[],"product":"key:rvm:rvm-context-service","canonical":"rvm-context-service","tier":"T1","t":["rvm-context-service","rvm","context","service","durable","encrypt","ruvector-backed","back","encryption","redb"],"s":["rvm-context-service","rvm","context","service","encryption","redb"]},
+{"id":"rvm-docs-mcp","family":"rvm-docs-mcp","kind":"npm","store":"rvm","version":"1.0.0","description":"MCP server and CLI for RVM documentation search, navigation, and cross-reference","keywords":[],"source":"rvm/userguide/mcp/package.json","sourceSha256":"f75fbd6c897dffe0c021bbae7efd5759ace22cb527829acc82ba90a9dbf60561","variants":[],"readme":"MCP server and CLI tool for searching, navigating, and querying the RVM user guide documentation. Provides six tools for fast, structured access to the RVM documentation set -- from keyword search to task-oriented reading paths. Register the MCP server so Claude Code can use the documentation tools: Once registered, the six docs_* tools are available in any Claude Code session. The CLI provides the same functionality as the MCP tools, with colored terminal output. Search across all documentation files by keyword or phrase. Input: Output: Matching paragraphs with file path, line number, and surrounding context. Results are ranked: exact phrase match > all terms present > any term present. Get","readmeSource":"rvm/userguide/mcp/README.md","product":"key:rvm:rvm-docs-mcp","canonical":"rvm-docs-mcp","tier":"T1","t":["rvm-docs-mcp","rvm","docs","mcp","server","documentation","search","navigation","cross-reference","cross","reference"],"s":["rvm-docs-mcp","rvm","docs","mcp"]},
+{"id":"rvm-gpu","family":"rvm-gpu","kind":"crate","store":"rvm","version":null,"description":"GPU compute subsystem for the RVM microhypervisor","keywords":["compute","gpu","hypervisor","no_std"],"source":"rvm/crates/rvm-gpu/Cargo.toml","sourceSha256":"3e5d51011bfad55fbbc75e73df3d849f54e1cf696b51a7b9ee182385bb1ac077","variants":[],"product":"key:rvm:rvm-gpu","canonical":"rvm-gpu","tier":"T1","t":["rvm-gpu","rvm","gpu","compute","subsystem","microhypervisor","hypervisor","std"],"s":["rvm-gpu","rvm","gpu","compute","hypervisor","std"]},
+{"id":"rvm-hal","family":"rvm-hal","kind":"crate","store":"rvm","version":null,"description":"Hardware abstraction layer for the RVM microhypervisor (ADR-133)","keywords":["embedded","hal","hypervisor","no_std"],"source":"rvm/crates/rvm-hal/Cargo.toml","sourceSha256":"4efd40b12fd08f6bef7eaf64cdb5a12b484b482794fcd8d44605f99b78ee8480","variants":[],"readme":"Platform-agnostic hardware abstraction traits for the RVM microhypervisor. Defines the trait interfaces that concrete platform implementations (AArch64, RISC-V, x86-64) must satisfy. All trait methods return RvmResult and pass borrowed slices rather than owned buffers. The trait definitions themselves contain no unsafe code. - Platform -- CPU discovery, total memory query, halt - MmuOps -- stage-2 page table management (map, unmap, translate, TLB flush) - TimerOps -- monotonic nanosecond timer with one-shot deadline support - InterruptOps -- interrupt enable/disable, acknowledge, end-of-interrupt - DC-15: #![no_std], #![forbid(unsafe_code)], #![deny(missing_docs)] - ADR-133: all trait method","readmeSource":"rvm/crates/rvm-hal/README.md","product":"key:rvm:rvm-hal","canonical":"rvm-hal","tier":"T1","t":["rvm-hal","rvm","hal","hardware","abstraction","layer","microhypervisor","adr-133","adr","133","embed","hypervisor","std"],"s":["rvm-hal","rvm","hal","embed","hypervisor","std"]},
+{"id":"rvm-host","family":"rvm-host","kind":"crate","store":"rvm","version":null,"description":"Host adapters that place a verified RVF into an isolation context and report the isolation it actually obtained (ADR-289, ADR-285)","keywords":["hypervisor","isolation","no_std","rvf","sandbox"],"source":"rvm/crates/rvm-host/Cargo.toml","sourceSha256":"0c47ff482572f4ef988567399cd77e7ecdd137523705ce248f7903d54e9bb1f6","variants":[],"product":"key:rvm:rvm-host","canonical":"rvm-host","tier":"T1","t":["rvm-host","rvm","host","adapter","place","verify","rvf","isolation","context","report","actually","obtain","adr-289","adr","289","adr-285","285","hypervisor","std","sandbox"],"s":["rvm-host","rvm","host","hypervisor","isolation","std","rvf","sandbox"]},
+{"id":"rvm-kernel","family":"rvm-kernel","kind":"crate","store":"rvm","version":null,"description":"Top-level integration kernel for the RVM microhypervisor","keywords":["coherence","hypervisor","kernel","no_std"],"source":"rvm/crates/rvm-kernel/Cargo.toml","sourceSha256":"4160e4fd544c746a5e0659895b607fc08eda31d9d588ad5d27275ec378d5702b","variants":[],"readme":"Top-level integration crate for the RVM coherence-native microhypervisor. Wires together all 12 subsystem crates (HAL, capabilities, witness, proof, partitions, scheduler, memory, coherence, boot, Wasm, and security) into a single API surface. Each subsystem is re-exported as a named module. This crate contains no logic of its own beyond the re-exports and version constants. - kernel::types -- rvm-types - kernel::hal -- rvm-hal - kernel::cap -- rvm-cap - kernel::witness -- rvm-witness - kernel::proof -- rvm-proof - kernel::partition -- rvm-partition - kernel::sched -- rvm-sched - kernel::memory -- rvm-memory - kernel::coherence -- rvm-coherence - kernel::boot -- rvm-boot - kernel::wasm -- rv","readmeSource":"rvm/crates/rvm-kernel/README.md","product":"key:rvm:rvm-kernel","canonical":"rvm-kernel","tier":"T1","t":["rvm-kernel","rvm","kernel","top-level","top","level","integration","microhypervisor","coherence","hypervisor","std"],"s":["rvm-kernel","rvm","kernel","coherence","hypervisor","std"]},
+{"id":"rvm-launch","family":"rvm-launch","kind":"crate","store":"rvm","version":null,"description":"Instance lifecycle for verified RVF agents: create, start, suspend, resume, checkpoint, terminate (ADR-289, ADR-288)","keywords":["checkpoint","hypervisor","lifecycle","no_std","rvf"],"source":"rvm/crates/rvm-launch/Cargo.toml","sourceSha256":"75a045ec5aa17c98c27954d083fb82f592086d734d4350f6b29b307e4b87cab3","variants":[],"product":"key:rvm:rvm-launch","canonical":"rvm-launch","tier":"T1","t":["rvm-launch","rvm","launch","instance","lifecycle","verify","rvf","create","start","suspend","resume","checkpoint","terminate","adr-289","adr","289","adr-288","288","hypervisor","std"],"s":["rvm-launch","rvm","launch","checkpoint","hypervisor","lifecycle","std","rvf"]},
+{"id":"rvm-memory","family":"rvm-memory","kind":"crate","store":"rvm","version":null,"description":"Memory management and guest physical address space for the RVM microhypervisor (ADR-136, ADR-138)","keywords":["hypervisor","memory","mmu","no_std"],"source":"rvm/crates/rvm-memory/Cargo.toml","sourceSha256":"752eac9bd07b68803689fb2f091d60eda39331f88f32161491aaf09129b4d2ff","variants":[],"readme":"Guest physical address space management for the RVM microhypervisor. Provides safe abstractions over stage-2 page table mappings with capability-gated access. Each partition has an independent guest physical address space. All mapping operations are recorded in the witness trail. Memory regions can be shared between partitions with explicit grants. - MemoryRegion -- descriptor: guest base, host base, page count, permissions, owner - MemoryPermissions -- RWX permission flags with constants (READ_ONLY, READ_WRITE, READ_EXECUTE) - validate_region(region) -- checks alignment, page count, and permission validity - regions_overlap(a, b) -- detects overlapping regions within the same partition - PA","readmeSource":"rvm/crates/rvm-memory/README.md","product":"key:rvm:rvm-memory","canonical":"rvm-memory","tier":"T1","t":["rvm-memory","rvm","memory","management","guest","physical","address","space","microhypervisor","adr-136","adr","136","adr-138","138","hypervisor","mmu","std"],"s":["rvm-memory","rvm","memory","hypervisor","mmu","std"]},
+{"id":"rvm-partition","family":"rvm-partition","kind":"crate","store":"rvm","version":null,"description":"Partition object model for RVM coherence domains (ADR-133)","keywords":["coherence","hypervisor","no_std","partition"],"source":"rvm/crates/rvm-partition/Cargo.toml","sourceSha256":"f4579039f503336dc7951e00a47c330157ef3c26e4165b71696e11e9457679d7","variants":[],"readme":"Partition lifecycle, isolation, and coherence domain management. A partition is not a VM. It has no emulated hardware, no guest BIOS, and no virtual device model. A partition is a container for a scoped capability table, communication edges to other partitions, coherence metrics, and CPU affinity. Partitions are the unit of scheduling, isolation, migration, and fault containment. Every lifecycle transition emits a witness record. - Partition -- the partition object: state, type, coherence, capability table, edges - PartitionManager -- create, destroy, lookup partitions (max 256 per instance) - PartitionState -- lifecycle states (e.g., Created, Running, Suspended) - PartitionType -- classific","readmeSource":"rvm/crates/rvm-partition/README.md","product":"key:rvm:rvm-partition","canonical":"rvm-partition","tier":"T1","t":["rvm-partition","rvm","partition","object","model","coherence","domain","adr-133","adr","133","hypervisor","std"],"s":["rvm-partition","rvm","partition","coherence","hypervisor","std"]},
+{"id":"rvm-proof","family":"rvm-proof","kind":"crate","store":"rvm","version":null,"description":"Proof-gated state transitions for the RVM microhypervisor (ADR-135)","keywords":["attestation","hypervisor","no_std","proof"],"source":"rvm/crates/rvm-proof/Cargo.toml","sourceSha256":"524fcbabaaa1426c1521aafa5ecca1eb41e6cedc7ef9b67d5b2437a89000ef0f","variants":[],"readme":"Proof-gated state transitions for the RVM microhypervisor. Every mutation to partition state requires a valid proof recorded in the witness trail. This crate defines the proof tiers, the Proof payload structure, and verification functions. Currently ships with a stub Hash-tier verifier; Witness-tier and ZK-tier verification are accepted but not yet fully implemented. - ProofTier -- enum: Hash, Witness, Zk - Proof -- proof payload with tier, commitment hash, and up to 64 bytes of data - verify(proof, commitment) -- verify a proof against an expected commitment - verify_with_cap(proof, commitment, token) -- verify with capability gate - DC-15: #![no_std], #![forbid(unsafe_code)], #![deny(missi","readmeSource":"rvm/crates/rvm-proof/README.md","product":"key:rvm:rvm-proof","canonical":"rvm-proof","tier":"T1","t":["rvm-proof","rvm","proof","proof-gated","gated","state","transition","microhypervisor","adr-135","adr","135","attestation","hypervisor","std"],"s":["rvm-proof","rvm","proof","attestation","hypervisor","std"]},
+{"id":"rvm-rvf","family":"rvm-rvf","kind":"crate","store":"rvm","version":null,"description":"RVF execution contract for RVM: verify-before-load, capability mapping, witnessed results (ADR-284, ADR-286)","keywords":["capability","hypervisor","no_std","rvf","verification"],"source":"rvm/crates/rvm-rvf/Cargo.toml","sourceSha256":"d6c55ef83eb14ba3c51033caebfeb8a39fa5859d475b676569fcd86c3852eed1","variants":[],"product":"key:rvm:rvm-rvf","canonical":"rvm-rvf","tier":"T1","t":["rvm-rvf","rvm","rvf","execution","contract","verify-before-load","verify","before","load","capability","mapp","witnes","result","adr-284","adr","284","adr-286","286","hypervisor","std","verification"],"s":["rvm-rvf","rvm","rvf","capability","hypervisor","std","verification"]},
+{"id":"rvm-sched","family":"rvm-sched","kind":"crate","store":"rvm","version":null,"description":"Coherence-aware scheduler for the RVM microhypervisor (ADR-132 DC-4)","keywords":["coherence","hypervisor","no_std","scheduler"],"source":"rvm/crates/rvm-sched/Cargo.toml","sourceSha256":"7e21318c034270c5b7e60eff88c590c9c8a63f74f5511d429861f213353e0813","variants":[],"readme":"Coherence-weighted 2-signal scheduler for the RVM microhypervisor. Combines deadline urgency and cut-pressure boost into a single priority signal: priority = deadline_urgency + cut_pressure_boost. The scheduler operates in three modes (Reflex, Flow, Recovery) and degrades gracefully when the coherence engine is unavailable, falling back to deadline-only scheduling. Partition switches are the hot path -- no allocation, no graph work, no policy evaluation during a switch. - Scheduler -- top-level scheduler managing all per-CPU schedulers - PerCpuScheduler -- per-CPU run queue and priority computation - SchedulerMode -- Reflex (hard RT), Flow (normal), Recovery (stabilization) - EpochTracker, E","readmeSource":"rvm/crates/rvm-sched/README.md","product":"key:rvm:rvm-sched","canonical":"rvm-sched","tier":"T1","t":["rvm-sched","rvm","sched","coherence-aware","coherence","aware","scheduler","microhypervisor","adr-132","adr","132","dc-4","hypervisor","std"],"s":["rvm-sched","rvm","sched","coherence","hypervisor","std","scheduler"]},
+{"id":"rvm-security","family":"rvm-security","kind":"crate","store":"rvm","version":null,"description":"Security policy enforcement for the RVM microhypervisor","keywords":["capability","hypervisor","no_std","security"],"source":"rvm/crates/rvm-security/Cargo.toml","sourceSha256":"36146abb4b4a3239a6dc53fa3a6387964eacbbbb64e2fa7d098efe0144c1562c","variants":[],"readme":"Unified security gate for the RVM microhypervisor. Provides the policy decision point that every hypercall passes through. The gate combines three stages: capability type and rights checking, proof commitment presence verification, and witness logging. Only after all stages pass does the hypercall proceed. Actual proof verification is delegated to rvm-proof; this crate handles the policy decision. 1. Capability check -- does the caller hold the required type and rights? 2. Proof verification -- is the proof commitment present and non-zero? 3. Witness logging -- record the decision (caller responsibility) - GateRequest -- bundles token, required type, required rights, optional proof commitmen","readmeSource":"rvm/crates/rvm-security/README.md","product":"key:rvm:rvm-security","canonical":"rvm-security","tier":"T1","t":["rvm-security","rvm","security","policy","enforcement","microhypervisor","capability","hypervisor","std"],"s":["rvm-security","rvm","security","capability","hypervisor","std"]},
+{"id":"rvm-wasm","family":"rvm","kind":"crate","store":"rvm","version":null,"description":"Optional WebAssembly guest runtime for the RVM microhypervisor","keywords":["coherence","hypervisor","no_std","types","wasm","webassembly"],"source":"rvm/crates/rvm-wasm/Cargo.toml","sourceSha256":"1449d7b843bd5ba4aac01ea3ce8cb1197bf0dba59a92ca0b120560e6bf10c488","variants":["rvm-types"],"readme":"Optional WebAssembly guest runtime for RVM partitions. When enabled, partitions can host Wasm modules as an alternative to native AArch64/RISC-V/x86-64 guests. Wasm modules execute in a sandboxed interpreter, host functions are exposed through the capability system, and all state transitions are witness-logged. This crate is compile-time optional; disabling it removes all Wasm code from the final binary. - WasmModuleState -- lifecycle: Loaded, Validated, Running, Terminated - WasmModuleInfo -- module metadata: partition, state, size, export/import counts - validate_header(bytes) -- checks the 8-byte Wasm preamble (magic + version 1) - MAX_MODULE_SIZE -- maximum module size (1 MiB default) -","readmeSource":"rvm/crates/rvm-wasm/README.md","product":"repo:rvm","canonical":"rvm-wasm","tier":"T1","t":["rvm-wasm","rvm","optional","guest","runtime","microhypervisor","coherence","hypervisor","std","type"],"s":["rvm-wasm","rvm","coherence","hypervisor","std","type"]},
+{"id":"rvm-witness","family":"rvm-witness","kind":"crate","store":"rvm","version":null,"description":"Witness logging subsystem for RVM audit trail (ADR-134)","keywords":["audit","hypervisor","no_std","witness"],"source":"rvm/crates/rvm-witness/Cargo.toml","sourceSha256":"29a338cf0fe77a4dd72283a9df3d11ee9f36081a376caf4b3d17d287cecd060e","variants":[],"readme":"Append-only witness trail with FNV-1a hash-chain integrity. Implements ADR-134: every privileged action emits a 64-byte witness record before the mutation is committed. If emission fails, the mutation does not proceed (\"no witness, no mutation\"). Records are stored in a fixed-capacity ring buffer and linked by a hash chain for tamper-evident auditing. - WitnessLog<N> -- generic ring buffer of capacity N records - WitnessEmitter -- builds records with auto-incrementing sequence and hash chain - WitnessRecord, ActionKind -- the 64-byte record and action discriminant - WitnessSigner, NullSigner -- pluggable record signing trait - verify_chain -- verify hash-chain integrity of a record slice - C","readmeSource":"rvm/crates/rvm-witness/README.md","product":"key:rvm:rvm-witness","canonical":"rvm-witness","tier":"T1","t":["rvm-witness","rvm","witness","logg","subsystem","audit","trail","adr-134","adr","134","hypervisor","std"],"s":["rvm-witness","rvm","witness","audit","hypervisor","std"]},
+{"id":"sky-monitor","family":"sky-monitor","kind":"crate","store":"skygraph","version":null,"description":"SkyGraph appliance core (ADR-199): ADS-B -> observer-relative sky coordinates -> track stitching -> SkyGraph -> embeddings -> anomaly scoring -> daily sky brief","keywords":[],"source":"skygraph/core/Cargo.toml","sourceSha256":"7da9dfff91b2ce008fd14c2d1617e88e2186197961b8d2f2d4ced709491e39da","variants":["sky-monitor-wasm"],"readme":"A local sky-monitoring appliance core that observes, projects, records, and explains activity above a fixed observer (the reference *Oakville node*, 43.4675 N / −79.6877 E / 100 m). The sky is treated as a continuously changing spatial graph, not a dashboard. Everything runs on a deterministic synthetic ADS-B + weather scenario — no network, no SDR hardware — while a dump1090 aircraft.json parser keeps the door open for real RTL-SDR data. Vectors live in [ruvector-core](../../crates/ruvector-core) (VectorDB), the SkyGraph in [ruvector-graph](../../crates/ruvector-graph) (GraphDB). Browser-facing bindings for the ADR-199 presentation plane (\"dashboard first\"), wrapping the pure subset with wa","readmeSource":"skygraph/core/README.md","product":"key:skygraph:sky-monitor","canonical":"sky-monitor","tier":"T2","t":["sky-monitor","sky","monitor","skygraph","appliance","core","adr-199","adr","199","ads-b","ads","observer-relative","observer","relative","coordinate","track","stitch","embedding","anomaly","scor","daily","brief"],"s":["sky-monitor","sky","monitor"]},
+{"id":"sonic_ct","family":"sonic_ct","kind":"crate","store":"ruvector","version":null,"description":"Ultrasound Computed Tomography (USCT) simulator: synthetic phantoms, ring acquisition, time-of-flight SART reconstruction, tissue segmentation, and self-learning acoustic memory.","keywords":["medical-imaging","reconstruction","tomography","ultrasound","usct"],"source":"ruvector/crates/sonic-ct/Cargo.toml","sourceSha256":"bb4fa8d8bfcb465934e19321bc22f302cf570f68c6b6e570866521922243773d","variants":[],"readme":"A dependency-free, browser-ready ultrasound CT (USCT) simulator and reconstruction toolkit. Synthetic phantoms · ring acquisition · time-of-flight SART reconstruction · tissue segmentation · self-learning acoustic memory — all in pure Rust, compiled to WebAssembly, visualised with React Three Fiber. sonic_ct models a ring ultrasound CT scanner — the kind of architecture behind next-generation full-body scanners that immerse a subject in a water bath and surround them with a dense ring of ultrasound transducers. Every element transmits and receives, so sound passes *through* the body from hundreds of angles. From those measurements the system reconstructs maps of acoustic properties (speed of","readmeSource":"ruvector/crates/sonic-ct/README.md","product":"key:ruvector:sonic_ct","canonical":"sonic_ct","tier":"T0","t":["sonic","ultrasound","comput","tomography","usct","simulator","synthetic","phantom","ring","acquisition","time-of-flight","flight","sart","reconstruction","tissue","segmentation","self-learning","self","learn","acoustic","memory","medical-imaging","medical","imag"],"s":["sonic","medical-imaging","medical","imag","reconstruction","tomography","ultrasound","usct"]},
+{"id":"sonic-ct-wasm","family":"sonic-ct","kind":"crate","store":"ruvector","version":null,"description":"WebAssembly (raw C-ABI) wrapper for the sonic_ct USCT simulator.","keywords":[],"source":"ruvector/crates/sonic-ct-wasm/Cargo.toml","sourceSha256":"ee5bac648e4beb6fbe54f3860e7516a659fed0ddea14322d55c62656bf6bfadb","variants":[],"product":"key:ruvector:sonic-ct","canonical":"sonic-ct-wasm","tier":"T0","t":["sonic-ct-wasm","sonic","sonic-ct","raw","c-abi","abi","wrapper","usct","simulator"],"s":["sonic-ct-wasm","sonic","sonic-ct"]},
+{"id":"strange-loop","family":"strange-loop","kind":"npm","store":"sublinear-time-solver","version":"0.2.0","description":"Hyper-optimized strange loops with temporal consciousness and quantum-classical hybrid computing. NPX: npx strange-loops","keywords":["consciousness","optimization","quantum","strange-loop","temporal"],"source":"sublinear-time-solver/crates/strange-loop/wasm-real/package.json","sourceSha256":"8d5037a2f5dca6f4197c9703cdf2ebdaa187b6d126e3c39289c3981b50770e29","variants":[],"readme":"A framework where thousands of tiny agents collaborate in real-time, each operating within nanosecond budgets, forming emergent intelligence through temporal feedback loops and quantum-classical hybrid computing. Experience the framework instantly with our JavaScript/WebAssembly NPX package: The NPX package provides: - 🎪 Interactive demos - nano-agents, quantum computing, temporal prediction - 📊 Performance benchmarks - validated 575,600+ ticks/second throughput - 🏗️ JavaScript SDK - full WASM integration for web and Node.js - 📦 Project templates - quick-start templates for different use cases NPM Package: [strange-loops](https://www.npmjs.com/package/strange-loops) - 🔧 Nano-Agent Frame","readmeSource":"sublinear-time-solver/crates/strange-loop/wasm-real/README.md","product":"key:sublinear-time-solver:strange-loop","canonical":"strange-loop","tier":"T1","t":["strange-loop","strange","loop","hyper-optimized","hyper","temporal","consciousness","quantum-classical","quantum","classical","hybrid","comput","npx","strange-loops"],"s":["strange-loop","strange","loop","consciousness","quantum","temporal"]},
+{"id":"strange-loops","family":"strange-loops","kind":"npm","store":"sublinear-time-solver","version":"0.5.0","description":"A framework where thousands of tiny agents collaborate in real-time, each operating within nanosecond budgets, forming emergent intelligence through temporal consciousness and quantum-classical hybrid computing","keywords":["consciousness","emergent-intelligence","hybrid-computing","nano-agents","quantum","real-time","strange-loops","temporal","wasm"],"source":"sublinear-time-solver/crates/strange-loop/package.json","sourceSha256":"e826050f470d525f9401b5d094fa21c49c13cc462306d71292ef61c372f86bb8","variants":[],"readme":"A framework where thousands of tiny agents collaborate in real-time, each operating within nanosecond budgets, forming emergent intelligence through temporal feedback loops and quantum-classical hybrid computing. - 🔧 Nano-Agent Framework - Thousands of lightweight agents executing in nanosecond time budgets - 🌀 Quantum-Classical Hybrid - Bridge quantum superposition with classical computation - ⏰ Temporal Prediction - Computing solutions before data arrives with sub-microsecond timing - 🧬 Self-Modifying Behavior - AI agents that evolve their own algorithms - 🌪️ Strange Attractor Dynamics - Chaos theory and non-linear temporal flows - ⏪ Retrocausal Feedback - Future state influences past","readmeSource":"sublinear-time-solver/crates/strange-loop/README.md","product":"key:sublinear-time-solver:strange-loops","canonical":"strange-loops","tier":"T1","t":["strange-loops","strange","loop","thousand","tiny","collaborate","real","operat","within","nanosecond","budget","form","emergent","intelligence","through","temporal","consciousness","quantum-classical","quantum","classical","hybrid","comput","emergent-intelligence","hybrid-computing","nano-agents","nano"],"s":["strange-loops","strange","loop","consciousness","emergent-intelligence","emergent","intelligence","hybrid-computing","hybrid","comput","nano-agents","nano","quantum","real","temporal"]},
+{"id":"stream-reformat","family":"stream-reformat","kind":"crate","store":"ai-text-watermark","version":null,"description":"Inflight LLM token-stream reformatting (separate thinking, collapse repetition, normalize) for Google Vertex / OpenRouter / meta-llm, built on ruvnet/midstream temporal-compare. Rust core + optional WASM.","keywords":["llm","midstream","reformat","streaming","wasm"],"source":"ai-text-watermark/crates/stream-reformat/Cargo.toml","sourceSha256":"d56737b0d58b0240b6338113f423d6ad2088afed3a66d27b9b6c539ab557311b","variants":[],"product":"key:ai-text-watermark:stream-reformat","canonical":"stream-reformat","tier":null,"t":["stream-reformat","stream","reformat","inflight","token-stream","token","reformatt","separate","think","collapse","repetition","normalize","google","vertex","openrouter","meta-llm","meta","built","midstream","temporal-compare","temporal","compare","core","optional"],"s":["stream-reformat","stream","reformat","midstream"]},
+{"id":"sublinear","family":"sublinear","kind":"crate","store":"sublinear-time-solver","version":null,"description":"High-performance sublinear-time solver for asymmetric diagonally dominant systems","keywords":["linear-algebra","solver","sparse-matrix","sublinear","wasm"],"source":"sublinear-time-solver/Cargo.toml","sourceSha256":"21144f95445392ff877d6ec3e5080218ea6e0caa8467f94ba87821c368cbcbb7","variants":[],"readme":"Every public solver, sampler, and analyser in this crate declares its worst-case complexity class at the type level (Complexity trait, compile-time const CLASS) and at the MCP wire level (x-complexity JSON Schema extension on every tool). Callers with a J/decision budget — Cognitum reflex loops, RuView change detection, Ruflo agentic inner loops — can refuse anything over budget at *tool-list time*, not after the call returns. The 12-tier taxonomy (Logarithmic → DoubleExponential, plus Adaptive { default, worst } for solvers that degrade on hard inputs) lives in src/complexity.rs. The decision rationale and full 6-item roadmap is in [docs/adr/ADR-001-complexity-as-architecture.md](docs/adr/A","readmeSource":"sublinear-time-solver/README.md","product":"repo:sublinear-time-solver","canonical":"sublinear-time-solver","tier":"T1","t":["sublinear","high","sublinear-time","solver","asymmetric","diagonally","dominant","linear-algebra","linear","algebra","sparse-matrix","sparse","matrix"],"s":["sublinear","linear-algebra","linear","algebra","solver","sparse-matrix","sparse","matrix"]},
+{"id":"sublinear-time-solver","family":"sublinear-time-solver","kind":"npm","store":"sublinear-time-solver","version":"1.7.2","description":"The Ultimate Mathematical & AI Toolkit: Sublinear algorithms, consciousness exploration, psycho-symbolic reasoning, chaos analysis, and temporal prediction in one unified MCP interface. WASM-accelerated with Lyapunov exponents and attractor dynamics.","keywords":["asymmetric-matrices","cli","consciousness","cross-tool-sharing","diagonally-dominant","dimension-reduction","dynamic-responses","emergence","emergent-behavior","enhanced-ai","feedback-loops","graph-algorithms","information-theory","johnson-lindenstrauss","linear-algebra","mcp","model-context-protocol","neumann-series","o-log-n","pagerank","persistent-learning","random-walk","self-modification","solver","spectral-sparsification","stochastic-exploration","sublinear-time","true-sublinear","typescript"],"source":"sublinear-time-solver/package.json","sourceSha256":"3331695d65f42df081f89392809bf0b8a924b1a2453ae19f2c51311d904d2b8f","variants":[],"readme":"Every public solver, sampler, and analyser in this crate declares its worst-case complexity class at the type level (Complexity trait, compile-time const CLASS) and at the MCP wire level (x-complexity JSON Schema extension on every tool). Callers with a J/decision budget — Cognitum reflex loops, RuView change detection, Ruflo agentic inner loops — can refuse anything over budget at *tool-list time*, not after the call returns. The 12-tier taxonomy (Logarithmic → DoubleExponential, plus Adaptive { default, worst } for solvers that degrade on hard inputs) lives in src/complexity.rs. The decision rationale and full 6-item roadmap is in [docs/adr/ADR-001-complexity-as-architecture.md](docs/adr/A","readmeSource":"sublinear-time-solver/README.md","product":"repo:sublinear-time-solver","canonical":"sublinear-time-solver","tier":"T1","t":["sublinear-time-solver","sublinear","solver","ultimate","mathematical","toolkit","algorithm","consciousness","exploration","psycho-symbolic","psycho","symbolic","reason","chao","analysis","temporal","prediction","unify","mcp","interface","wasm-accelerated","accelerat","lyapunov","exponent","attractor","dynamic","asymmetric-matrices","asymmetric","matrice","cross-tool-sharing","cross","shar","diagonally-dominant","diagonally","dominant","dimension-reduction","dimension","reduction","dynamic-responses","respons","emergence","emergent-behavior","emergent","behavior","enhanced-ai","enhanc","feedback-loops","feedback","loop","graph-algorithms","graph","information-theory","information","theory","johnson-lindenstrauss","johnson","lindenstrauss","linear-algebra","linear","algebra","model-context-protocol","model","context","protocol","neumann-series","neumann","sery","o-log-n","log","pagerank","persistent-learning","persistent","learn","random-walk","random","walk","self-modification","self","modification","spectral-sparsification","spectral","sparsification","stochastic-exploration","stochastic","sublinear-time","true-sublinear","true"],"s":["sublinear-time-solver","sublinear","solver","asymmetric-matrices","asymmetric","matrice","consciousness","cross-tool-sharing","cross","shar","diagonally-dominant","diagonally","dominant","dimension-reduction","dimension","reduction","dynamic-responses","dynamic","respons","emergence","emergent-behavior","emergent","behavior","enhanced-ai","enhanc","feedback-loops","feedback","loop","graph-algorithms","graph","algorithm","information-theory","information","theory","johnson-lindenstrauss","johnson","lindenstrauss","linear-algebra","linear","algebra","mcp","model-context-protocol","model","context","protocol","neumann-series","neumann","sery","o-log-n","log","pagerank","persistent-learning","persistent","learn","random-walk","random","walk","self-modification","self","modification","spectral-sparsification","spectral","sparsification","stochastic-exploration","stochastic","exploration","sublinear-time","true-sublinear","true"]},
+{"id":"swe-bench-adapter","family":"swe-bench-adapter","kind":"crate","store":"ruv-fann","version":null,"description":"SWE-Bench adapter for ruv-swarm orchestration system","keywords":["benchmarking","evaluation","orchestration","swarm","swe-bench"],"source":"ruv-fann/ruv-swarm/crates/swe-bench-adapter/Cargo.toml","sourceSha256":"34b84f144a138b68c534f1a054876b03e262da9c81c845b96ff48a5b4b3f96be","variants":[],"readme":"A high-performance adapter for integrating SWE-Bench evaluation with the ruv-swarm orchestration system and Claude Code CLI. - Instance Loading: Load and categorize SWE-Bench instances with automatic difficulty assessment - Prompt Generation: Generate optimized prompts for Claude Code CLI based on instance complexity - Patch Evaluation: Evaluate generated patches against test requirements in sandboxed environments - Performance Benchmarking: Comprehensive performance metrics and comparative analysis - Stream Parsing: Real-time metrics collection from Claude Code CLI output streams The adapter consists of five main components: - Loads SWE-Bench instances from local storage or remote repositor","readmeSource":"ruv-fann/ruv-swarm/crates/swe-bench-adapter/README.md","product":"key:ruv-fann:swe-bench-adapter","canonical":"swe-bench-adapter","tier":"T1","t":["swe-bench-adapter","swe","bench","adapter","swe-bench","ruv-swarm","swarm","orchestration","benchmark","evaluation"],"s":["swe-bench-adapter","swe","bench","adapter","benchmark","evaluation","orchestration","swarm","swe-bench"]},
+{"id":"symbolic-scribe-harness","family":"symbolic-scribe-harness","kind":"npm","store":"symbolic-scribe","version":"0.1.1","description":"Repo-aware agent harness for Symbolic Scribe — maintainer/benchmarker/release/security agents with doctor, repo-triage, and release-check commands.","keywords":["agent-harness","ai-agent","claude-code","mcp","metaharness","symbolic-scribe"],"source":"symbolic-scribe/harness/package.json","sourceSha256":"791f250855f44212cbac5efac08f7ab6e6a6fdbffa2a5742764ba1c6f3ccf58e","variants":[],"product":"key:symbolic-scribe:symbolic-scribe-harness","canonical":"symbolic-scribe-harness","tier":"T2","t":["symbolic-scribe-harness","symbolic","scribe","harness","repo-aware","repo","aware","maintainer","benchmarker","release","security","doctor","repo-triage","triage","release-check","check","command","agent-harness","ai-agent","claude-code","claude","code","mcp","metaharness","symbolic-scribe"],"s":["symbolic-scribe-harness","symbolic","scribe","harness","agent-harness","ai-agent","claude-code","claude","code","mcp","metaharness","symbolic-scribe"]},
+{"id":"synaptic-daa-swarm","family":"synaptic-daa-swarm","kind":"crate","store":"synaptic-mesh","version":null,"description":"Distributed Autonomous Agent swarm intelligence framework for Synaptic Neural Mesh","keywords":["ai","autonomous","daa","distributed","swarm"],"source":"synaptic-mesh/standalone-crates/synaptic-daa-swarm/Cargo.toml","sourceSha256":"77372efa292944486d4f2165b908ba1264def463c9e2dcca2ee913965bafd957","variants":[],"readme":"Distributed Autonomous Agent swarm intelligence framework for the Synaptic Neural Mesh project. - Swarm Intelligence: Emergent behaviors from distributed agents - Self-Organization: Autonomous coordination and adaptation - Evolutionary Algorithms: Adaptive optimization strategies - Scalable: Designed for thousands of agents MIT OR Apache-2.0","readmeSource":"synaptic-mesh/standalone-crates/synaptic-daa-swarm/README.md","product":"key:synaptic-mesh:synaptic-daa-swarm","canonical":"synaptic-daa-swarm","tier":"T1","t":["synaptic-daa-swarm","synaptic","daa","swarm","distribut","autonomous","intelligence","neural","mesh"],"s":["synaptic-daa-swarm","synaptic","daa","swarm","autonomous","distribut"]},
+{"id":"synaptic-mesh","family":"synaptic-mesh","kind":"npm","store":"synaptic-mesh","version":"0.1.0","description":"Distributed neural mesh with decentralized Claude-Max marketplace","keywords":["claude","cli","cognition","distributed","distributed-ai","marketplace","mesh","mesh-network","neural","neural-network","p2p","quantum-resistant"],"source":"synaptic-mesh/npx-wrapper/package.json","sourceSha256":"8570967522fde3eeb0eb332d5c878c915b84f546a432c4314efdb00fca0b3c67","variants":["synaptic-mesh-cli"],"product":"repo:synaptic-mesh","canonical":"synaptic-mesh","tier":"T1","t":["synaptic-mesh","synaptic","mesh","distribut","neural","decentraliz","claude-max","claude","max","marketplace","cognition","distributed-ai","mesh-network","network","neural-network","p2p","quantum-resistant","quantum","resistant"],"s":["synaptic-mesh","synaptic","mesh","claude","cognition","distribut","distributed-ai","marketplace","mesh-network","network","neural","neural-network","p2p","quantum-resistant","quantum","resistant"]},
+{"id":"synaptic-mesh-p2p","family":"synaptic-mesh-p2p","kind":"crate","store":"synaptic-mesh","version":null,"description":"P2P networking layer for Synaptic Neural Mesh with QuDAG integration","keywords":[],"source":"synaptic-mesh/src/rs/synaptic-mesh-p2p/Cargo.toml","sourceSha256":"edb4a6756a6d550b63a272c419ab8c2c9c16017a62a8b9026bdb620c2f4c1dc9","variants":[],"product":"key:synaptic-mesh:synaptic-mesh-p2p","canonical":"synaptic-mesh-p2p","tier":"T1","t":["synaptic-mesh-p2p","synaptic","mesh","p2p","network","layer","neural","qudag","integration"],"s":["synaptic-mesh-p2p","synaptic","mesh","p2p"]},
+{"id":"synaptic-neural-mesh","family":"synaptic-neural-mesh","kind":"crate","store":"synaptic-mesh","version":null,"description":"Neural mesh coordination layer for distributed AI agents in Synaptic Neural Mesh","keywords":["ai","coordination","distributed","mesh","neural"],"source":"synaptic-mesh/standalone-crates/synaptic-neural-mesh/Cargo.toml","sourceSha256":"9b6a6bd09d00591ab0bac7bbea3b0e668ff433690476be0af14d55fb935403a6","variants":[],"readme":"Neural mesh coordination layer for distributed AI agents in the Synaptic Neural Mesh project. - Agent Coordination: Efficient coordination of distributed neural agents - Mesh Topology: Self-organizing mesh network for AI communication - Async Operations: Built on Tokio for high-performance async I/O - Integration: Seamless integration with QuDAG core networking MIT OR Apache-2.0","readmeSource":"synaptic-mesh/standalone-crates/synaptic-neural-mesh/README.md","product":"key:synaptic-mesh:synaptic-neural-mesh","canonical":"synaptic-neural-mesh","tier":"T1","t":["synaptic-neural-mesh","synaptic","neural","mesh","coordination","layer","distribut"],"s":["synaptic-neural-mesh","synaptic","neural","mesh","coordination","distribut"]},
+{"id":"synaptic-neural-wasm","family":"synaptic-neural","kind":"crate","store":"synaptic-mesh","version":null,"description":"WASM-optimized neural network engine for Synaptic Neural Mesh with SIMD acceleration","keywords":["ai","mesh","neural","simd","wasm"],"source":"synaptic-mesh/standalone-crates/synaptic-neural-wasm/Cargo.toml","sourceSha256":"6e6752e75e119f49aec818b5e116cefb4495b5f15f49a530424b33a96622af1c","variants":[],"readme":"WASM-optimized neural network engine for the Synaptic Neural Mesh project. - SIMD Acceleration: Leverages WebAssembly SIMD for fast computations - Lightweight: Minimal dependencies for small WASM bundle size - Browser & Node.js: Works in any WASM environment - Neural Operations: Matrix operations, activation functions, and more MIT OR Apache-2.0","readmeSource":"synaptic-mesh/standalone-crates/synaptic-neural-wasm/README.md","product":"key:synaptic-mesh:synaptic-neural","canonical":"synaptic-neural-wasm","tier":"T1","t":["synaptic-neural-wasm","synaptic","neural","synaptic-neural","wasm-optimized","network","mesh","acceleration"],"s":["synaptic-neural-wasm","synaptic","neural","synaptic-neural","mesh"]},
+{"id":"synaptic-qudag-core","family":"synaptic-qudag","kind":"crate","store":"synaptic-mesh","version":null,"description":"QuDAG core networking for Synaptic Neural Mesh - DAG-based consensus and P2P networking","keywords":["consensus","dag","mesh","network","p2p"],"source":"synaptic-mesh/publish-ready/synaptic-qudag-core/Cargo.toml","sourceSha256":"a0ae2cd76425b23ecd725c0608919cf73e336c847703c78b3d832d208263cca9","variants":[],"readme":"Core DAG networking and consensus library for the Synaptic Neural Mesh project. - DAG-based consensus: Efficient directed acyclic graph structure - P2P networking: Built on modern cryptographic primitives - Async/await: Fully asynchronous implementation - Production-ready: Designed for distributed neural mesh networks MIT OR Apache-2.0","readmeSource":"synaptic-mesh/publish-ready/synaptic-qudag-core/README.md","product":"key:synaptic-mesh:synaptic-qudag","canonical":"synaptic-qudag-core","tier":"T1","t":["synaptic-qudag-core","synaptic","qudag","core","synaptic-qudag","network","neural","mesh","dag-based","dag","consensus","p2p"],"s":["synaptic-qudag-core","synaptic","qudag","core","synaptic-qudag","consensus","dag","mesh","network","p2p"]},
+{"id":"temporal-compare","family":"temporal-compare","kind":"crate","store":"sublinear-time-solver","version":null,"description":"High-performance framework for benchmarking temporal prediction algorithms inspired by Time-R1","keywords":["benchmark","machine-learning","prediction","temporal","time-series"],"source":"sublinear-time-solver/crates/temporal-compare/Cargo.toml","sourceSha256":"8a65fb214f3f792ef30d54024146c134da6dc242e83d8b27a2f95173799f5b07","variants":[],"readme":"Imagine trying to predict the next word you'll type, the next stock price movement, or the next frame in a video. These are temporal prediction tasks - predicting future states from historical sequences. Temporal-Compare provides a testing ground to compare different approaches to this fundamental problem. This crate implements a clean, extensible framework for comparing: - 15+ ML backends from basic MLPs to ensemble methods - INT8 quantization (3.69x model compression, 0.42% accuracy loss) - SIMD acceleration (AVX2/AVX-512 intrinsics for 6x speedup) - Production-ready optimizations with real benchmarks, no overfitting - 🚀 INT8 Quantization: 3.69x model compression (9.7KB → 2.6KB) - ⚡ AVX2/","readmeSource":"sublinear-time-solver/crates/temporal-compare/README.md","product":"key:sublinear-time-solver:temporal-compare","canonical":"temporal-compare","tier":"T1","t":["temporal-compare","temporal","compare","high","benchmark","prediction","algorithm","inspir","time-r1","machine-learning","machine","learn","time-series","sery"],"s":["temporal-compare","temporal","compare","benchmark","machine-learning","machine","learn","prediction","time-series","sery"]},
+{"id":"temporal-lead-solver","family":"temporal-lead-solver","kind":"crate","store":"sublinear-time-solver","version":null,"description":"Temporal computational lead via sublinear local solvers for diagonally dominant systems","keywords":["diagonal-dominance","prediction","solver","sublinear","temporal"],"source":"sublinear-time-solver/crates/temporal-lead-solver/Cargo.toml","sourceSha256":"3bbd3aefa328a10fdeb9e6149dcc7d5ede3005cb71a6a3c9eb8c53b3dbeb8e73","variants":[],"readme":"Achieve temporal computational lead through sublinear-time algorithms for diagonally dominant systems. Created by rUv - [github.com/ruvnet](https://github.com/ruvnet) - Temporal Computational Lead: Predict solutions before network messages arrive - O(poly(1/ε, 1/δ)) query complexity - Model-based inference (NOT faster-than-light signaling) - Scientifically rigorous implementation - Light travel time: 36.3 ms - Computation time: 0.996 ms - Temporal advantage: 35.3 ms - Effective velocity: 36× speed of light 1. Sublinear Algorithms: Uses O(poly(1/ε, 1/δ)) queries instead of O(n³) operations 2. Local Computation: All queries access locally available data 3. Model-Based Inference: Exploits diago","readmeSource":"sublinear-time-solver/crates/temporal-lead-solver/README.md","product":"key:sublinear-time-solver:temporal-lead-solver","canonical":"temporal-lead-solver","tier":"T1","t":["temporal-lead-solver","temporal","lead","solver","computational","via","sublinear","local","diagonally","dominant","diagonal-dominance","diagonal","dominance","prediction"],"s":["temporal-lead-solver","temporal","lead","solver","diagonal-dominance","diagonal","dominance","prediction","sublinear"]},
+{"id":"temporal-neural-solver","family":"temporal-neural-solver","kind":"npm","store":"sublinear-time-solver","version":"1.0.0","description":"Ultra-fast temporal neural network solver with <40ns P99.9 latency","keywords":["avx2","high-performance","int8","neural-network","rust","sublinear","temporal-solver","wasm"],"source":"sublinear-time-solver/crates/neural-network-implementation/real-implementation/package.json","sourceSha256":"f2b5d50b5f37f7807286c7fdaf286fb224fcdbf132a0fe30bebdc2d9f5413331","variants":["temporal-neural-solver-wasm"],"product":"key:sublinear-time-solver:temporal-neural-solver","canonical":"temporal-neural-solver","tier":"T1","t":["temporal-neural-solver","temporal","neural","solver","ultra-fast","ultra","network","40ns","p99.9","latency","avx2","high","int8","neural-network","sublinear","temporal-solver"],"s":["temporal-neural-solver","temporal","neural","solver","avx2","high","int8","neural-network","network","sublinear","temporal-solver"]},
+{"id":"thermal-brain","family":"thermal-brain","kind":"crate","store":"cognitum-cogs","version":null,"description":"Neuromorphic sparse spiking neural system with thermal self-regulation","keywords":["edge-ai","embedded","neuromorphic","spiking-neural-network","wasm"],"source":"cognitum-cogs/crates/thermal-brain/Cargo.toml","sourceSha256":"b452d3c9abe4765fe0fa1083dc85a294dbaca61d65057efece252e237a3f37c8","variants":[],"readme":"A brain-inspired embedded system that thinks like neurons and stays cool under pressure. ThermalBrain brings the efficiency of biological neural networks to embedded devices. Just like your brain processes information through electrical spikes while regulating its temperature, ThermalBrain uses spiking neural networks with built-in thermal management to run AI inference on resource-constrained hardware. Imagine running pattern recognition on a tiny microcontroller that: - Uses 90% less memory than traditional neural networks - Automatically slows down when it gets too hot (like your laptop, but smarter) - Speeds up temporarily when it needs burst performance - Learns new patterns on-device w","readmeSource":"cognitum-cogs/crates/thermal-brain/README.md","product":"key:cognitum-cogs:thermal-brain","canonical":"thermal-brain","tier":null,"t":["thermal-brain","thermal","brain","neuromorphic","sparse","spik","neural","self-regulation","self","regulation","edge-ai","edge","embed","spiking-neural-network","network"],"s":["thermal-brain","thermal","brain","edge-ai","edge","embed","neuromorphic","spiking-neural-network","spik","neural","network"]},
+{"id":"thermorust","family":"thermorust","kind":"crate","store":"ruvector","version":null,"description":"Thermodynamic neural motif engine: energy-driven state transitions with Landauer dissipation and Langevin noise","keywords":["ising","langevin","neural","physics","thermodynamics"],"source":"ruvector/crates/thermorust/Cargo.toml","sourceSha256":"d0409bd15ef4aa02a2cd0d7bd036666d5bd30a7f602a96fb21a06f3f00bda226","variants":[],"readme":"A minimal thermodynamic neural-motif engine for Rust. Treats computation as energy-driven state transitions with Landauer-style dissipation tracking and Langevin/Metropolis noise baked in. - Ising and soft-spin Hamiltonians with configurable coupling matrices and local fields. - Metropolis-Hastings (discrete) and overdamped Langevin (continuous) dynamics. - Landauer dissipation accounting -- every accepted irreversible transition charges kT ln 2 of heat, giving a physical energy audit of your computation. - Langevin and Poisson spike noise sources satisfying the fluctuation-dissipation theorem. - Thermodynamic observables -- magnetisation, pattern overlap, binary entropy, free energy, and ru","readmeSource":"ruvector/crates/thermorust/README.md","product":"key:ruvector:thermorust","canonical":"thermorust","tier":"T0","t":["thermorust","thermodynamic","neural","motif","energy-driven","energy","driven","state","transition","landauer","dissipation","langevin","noise","ising","physic"],"s":["thermorust","ising","langevin","neural","physic","thermodynamic"]},
+{"id":"timesfm","family":"timesfm","kind":"crate","store":"ruvector","version":null,"description":"TimesFM 1.0 200M decoder-only patched time-series Transformer inference in Rust/candle (ADR-189/191)","keywords":["anomaly","candle","forecasting","ruvector","time-series","timesfm","transformer"],"source":"ruvector/crates/timesfm/Cargo.toml","sourceSha256":"6067bfd4e0563b25fbedea84eb76e9e17f993ea106d69f476c2f94b8a90b829e","variants":["ruvector-timesfm"],"readme":"Native Rust / [candle](https://github.com/huggingface/candle) port of Google's TimesFM 1.0 200M patched time-series Transformer — zero-shot forecasting inside RuVector, no Python microservice. TimesFM (Time Series Foundation Model) is Google's decoder-only Transformer for zero-shot time-series forecasting, released at ICML 2024. The 200M variant uses: - Zero-shot forecasting — no fine-tuning required for new time series - RevIN normalisation — per-series instance norm (bug-fixed vs. naive implementation) - Autoregressive decode — arbitrary horizon via multi-step patch decoding - Quantile outputs — [B, N, 128, 10] tensor with 10 calibrated quantile levels - CPU + GPU — accelerated via --featu","readmeSource":"ruvector/crates/timesfm/README.md","product":"key:ruvector:timesfm","canonical":"timesfm","tier":"T0","t":["timesfm","1.0","200m","decoder-only","decoder","only","patch","time-series","sery","transformer","inference","candle","adr-189","adr","189","191","anomaly","forecast"],"s":["timesfm","anomaly","candle","forecast","time-series","sery","transformer"]},
+{"id":"timesfm-harness","family":"timesfm-harness","kind":"npm","store":"ruvector","version":"0.1.0","description":"TimesFM 1.0 200M decoder-only time-series forecasting inference crate (Rust/candle) — engineering pod harness","keywords":[],"source":"ruvector/harnesses/timesfm-harness/package.json","sourceSha256":"c9a1fc806f1dc51b9ba97c98f8139fd784a8c7685feb500b844514c03d583a87","variants":[],"readme":"TimesFM 1.0 200M decoder-only time-series forecasting inference crate (Rust/candle) — engineering pod harness This harness ships with the claude-code adapter. MIT","readmeSource":"ruvector/harnesses/timesfm-harness/README.md","product":"key:ruvector:timesfm-harness","canonical":"timesfm-harness","tier":"T0","t":["timesfm-harness","timesfm","harness","1.0","200m","decoder-only","decoder","only","time-series","sery","forecast","inference","crate","candle","engineer","pod"],"s":["timesfm-harness","timesfm","harness"]},
+{"id":"ultrasonic-agent","family":"ultrasonic-agent","kind":"npm","store":"ultrasonic","version":"0.1.0","description":"Maintain and validate ultrasonic","keywords":[],"source":"ultrasonic/.harness/generated/package.json","sourceSha256":"5e30e32cc3917ba40145fc4d07d15d2c8132d63756f4aa5c961706b789ea6d3b","variants":[],"readme":"Repository specific agent harness generated with metaharness@0.4.16 and reviewed for local operation. It connects the repo CLI and MCP to Codex, Claude Code, and Copilot. From the repository root: Install domain dependencies as described in the root README before domain tools. The last command prints an absolute local MCP configuration for hosts with arbitrary working directories. Checked in host profiles assume repository root as working directory. No unpublished npm package is invoked. Run npm run bench:generate --prefix .harness/generated followed by npm run bench:verify --prefix .harness/generated. Generated suites use repository tests and are not independent hidden evaluation. The real","readmeSource":"ultrasonic/.harness/generated/README.md","product":"key:ultrasonic:ultrasonic-agent","canonical":"ultrasonic-agent","tier":"T3","t":["ultrasonic-agent","ultrasonic","maintain","validate"],"s":["ultrasonic-agent","ultrasonic"]},
+{"id":"verifier","family":"verifier","kind":"crate","store":"autogenous","version":null,"description":"Autogenous AGL verifier: admission verdicts composing constitution pinning, typed mutation admission, prohibited effects, and hard fitness gates (ADR-392 Phase 3).","keywords":[],"source":"autogenous/crates/verifier/Cargo.toml","sourceSha256":"6b40d1dc1a2a87bc76ee9c9802f3899ee6b37111c24c1d1c19e60bce2253c00e","variants":[],"readme":"verifier is the admission gate of [Autogenous](../), a self-evolving AI-defense runtime in Rust. It composes every structural check into a single verdict: constitution pinning, typed-mutation admission (authority, invariants, rollback, expiry), prohibited effects, and the hard fitness gates. It returns the complete list of violations, not just a pass/fail — a failure you can act on, not a mystery. - verify(...) — the one call. Give it the constitution, parent genome, proposed mutation, prohibited-effect list, and optional fitness; get a Verdict. - Verdict — admissible / promotable plus the full violations list. The verifier holds no mutable state and takes everything as input (ADR-392 §3.4).","readmeSource":"autogenous/crates/verifier/README.md","product":"key:autogenous:verifier","canonical":"verifier","tier":null,"t":["verifier","autogenous","agl","admission","verdict","compos","constitution","pinn","typed","mutation","prohibit","effect","hard","fitness","gate","adr-392","adr","392","phase"],"s":["verifier"]},
+{"id":"vscode-remote-mcp","family":"vscode-remote-mcp","kind":"npm","store":"vsc-remote-mcp","version":"1.0.0","description":"Enhanced MCP server for VSCode Remote integration","keywords":["docker","mcp","remote","vscode"],"source":"vsc-remote-mcp/vscode-remote-mcp/package.json","sourceSha256":"ef7914866ab69a9d367ee60f897d45d6c35c57e105bba746c930e35b5ba6a5e0","variants":[],"readme":"The VSCode Remote MCP Server is an enhanced implementation of the Model Context Protocol (MCP) designed specifically for VSCode Remote integration. It provides a robust set of tools for code analysis, modification, searching, and VSCode instance management through Docker containers. This server enables AI assistants and other tools to interact with your codebase and manage development environments programmatically. The VSCode Remote MCP Server follows a client-server architecture based on the Model Context Protocol (MCP) specification. The server exposes a set of tools and resources that clients can use to perform various operations on code files and manage VSCode instances. The server imple","readmeSource":"vsc-remote-mcp/vscode-remote-mcp/README.md","product":"key:vsc-remote-mcp:vscode-remote-mcp","canonical":"vscode-remote-mcp","tier":"T3","t":["vscode-remote-mcp","vscode","remote","mcp","enhanc","server","integration","docker"],"s":["vscode-remote-mcp","vscode","remote","mcp","docker"]},
+{"id":"wifi-densepose-aether","family":"wifi-densepose-aether","kind":"crate","store":"ruview","version":null,"description":"AETHER pure-compute stack (ADR-024): contrastive CSI embedding, CSI-to-pose transformer, SONA drift/LoRA, and quantization — std-only, no async/server deps so the Python `[aether]` wheel stays lean (ADR-185 §3.2)","keywords":[],"source":"ruview/v2/crates/wifi-densepose-aether/Cargo.toml","sourceSha256":"effc09d38e94096cc3b272cb5a910de18e0bee4de9752ade2e3f0c2b16134d45","variants":[],"product":"key:ruview:wifi-densepose-aether","canonical":"wifi-densepose-aether","tier":"T0","t":["wifi-densepose-aether","wifi","densepose","aether","pure-compute","pure","compute","stack","adr-024","adr","024","contrastive","csi","embedd","csi-to-pose","pose","transformer","sona","drift","lora","quantization","std-only","std","only","async","server","deps","python","wheel","stay","lean","adr-185","185","3.2"],"s":["wifi-densepose-aether","wifi","densepose","aether"]},
+{"id":"wifi-densepose-bfld","family":"wifi-densepose-bfld","kind":"crate","store":"ruview","version":null,"description":"BFLD — Beamforming Feedback Layer for Detection. Privacy-gated WiFi BFI sensing primitives. See ADR-118.","keywords":[],"source":"ruview/v2/crates/wifi-densepose-bfld/Cargo.toml","sourceSha256":"f30214e8f401078898d50a108c69774be227a36046e5b823bf3503768c2a48cf","variants":[],"readme":"BFLD — Beamforming Feedback Layer for Detection. Privacy-gated WiFi sensing primitives derived from 802.11ac/ax Beamforming Feedback Information (BFI). See [ADR-118](../../../docs/adr/ADR-118-bfld-beamforming-feedback-layer-for-detection.md) for the umbrella architecture decision and [docs/research/BFLD/](../../../docs/research/BFLD/) for the full design dossier. The crate enforces three privacy invariants structurally (via the type system + memory hygiene), not by policy text: Minimal in-process consumer (see examples/bfld_minimal.rs): Production worker-thread + HA-DISCO publishing (see examples/bfld_handle.rs): Stripping to --no-default-features keeps the no_std-compatible core (BfldFrameH","readmeSource":"ruview/v2/crates/wifi-densepose-bfld/README.md","product":"key:ruview:wifi-densepose-bfld","canonical":"wifi-densepose-bfld","tier":"T0","t":["wifi-densepose-bfld","wifi","densepose","bfld","beamform","feedback","layer","detection","privacy-gated","privacy","gated","bfi","sens","primitive","see","adr-118","adr","118"],"s":["wifi-densepose-bfld","wifi","densepose","bfld"]},
+{"id":"wifi-densepose-calibration","family":"wifi-densepose-calibration","kind":"crate","store":"ruview","version":null,"description":"ADR-151 per-room calibration & specialized model training (baseline → enroll → extract → train)","keywords":[],"source":"ruview/v2/crates/wifi-densepose-calibration/Cargo.toml","sourceSha256":"bfb33f1105875694259d5de3acfca6f5e01cb576ba69e01af550c41dd6af6f98","variants":[],"product":"key:ruview:wifi-densepose-calibration","canonical":"wifi-densepose-calibration","tier":"T0","t":["wifi-densepose-calibration","wifi","densepose","calibration","adr-151","adr","151","per-room","per","room","specializ","model","train","baseline","enroll","extract"],"s":["wifi-densepose-calibration","wifi","densepose","calibration"]},
+{"id":"wifi-densepose-core","family":"wifi-densepose","kind":"crate","store":"ruview","version":null,"description":"Core types, traits, and utilities for WiFi-DensePose pose estimation system","keywords":["browser","densepose","wasm","webassembly","wifi"],"source":"ruview/v2/crates/wifi-densepose-core/Cargo.toml","sourceSha256":"6ae5af4c02a48966b4993d357be6afc66b773cd8dca0246ac8ef94895ed92fea","variants":["wifi-densepose-wasm"],"readme":"Core types, traits, and utilities for the WiFi-DensePose pose estimation system. wifi-densepose-core is the foundation crate for the WiFi-DensePose workspace. It defines the shared data structures, error types, and trait contracts used by every other crate in the ecosystem. The crate is no_std-compatible (with the std feature disabled) and forbids all unsafe code. - Core data types -- CsiFrame, ProcessedSignal, PoseEstimate, PersonPose, Keypoint, KeypointType, BoundingBox, Confidence, Timestamp, and more. - Trait abstractions -- SignalProcessor, NeuralInference, and DataStore define the contracts for signal processing, neural network inference, and data persistence respectively. - Error hier","readmeSource":"ruview/v2/crates/wifi-densepose-core/README.md","product":"key:ruview:wifi-densepose","canonical":"wifi-densepose-core","tier":"T0","t":["wifi-densepose-core","wifi","densepose","core","wifi-densepose","type","trait","utility","pose","estimation","browser"],"s":["wifi-densepose-core","wifi","densepose","core","wifi-densepose","browser"]},
+{"id":"wifi-densepose-desktop","family":"wifi-densepose-desktop","kind":"crate","store":"ruview","version":null,"description":"Tauri v2 desktop frontend for RuView WiFi DensePose","keywords":[],"source":"ruview/v2/crates/wifi-densepose-desktop/Cargo.toml","sourceSha256":"9a2bc3ece913a9fec6a4a193fa03fbdb7dccd2155f9fbd03b2535e41b1b6dcee","variants":[],"readme":"Cross-platform desktop application for managing ESP32 WiFi sensing networks. Built with Tauri v2 (Rust backend) and React + TypeScript (frontend), following the [ADR-053 design system](../../docs/adr/ADR-053-ui-design-system.md). RuView Desktop provides a unified interface for node discovery, firmware management, over-the-air updates, WASM edge module deployment, real-time sensing data visualization, and mesh network topology monitoring — all from a single native application. The UI follows a dark professional theme with the following design tokens:","readmeSource":"ruview/v2/crates/wifi-densepose-desktop/README.md","product":"key:ruview:wifi-densepose-desktop","canonical":"wifi-densepose-desktop","tier":"T0","t":["wifi-densepose-desktop","wifi","densepose","desktop","tauri","frontend","ruview"],"s":["wifi-densepose-desktop","wifi","densepose","desktop"]},
+{"id":"wifi-densepose-engine","family":"wifi-densepose-engine","kind":"crate","store":"ruview","version":null,"description":"RuView streaming-engine integration layer — composes the ADR-135..146 building blocks into one trust-traceable pipeline cycle","keywords":[],"source":"ruview/v2/crates/wifi-densepose-engine/Cargo.toml","sourceSha256":"18023aa1b8b1c280ff1b318352c06e48cf9610891bed88b02c04d328cfd90323","variants":[],"product":"key:ruview:wifi-densepose-engine","canonical":"wifi-densepose-engine","tier":"T0","t":["wifi-densepose-engine","wifi","densepose","ruview","streaming-engine","stream","integration","layer","compos","adr-135..146","adr","135..146","block","trust-traceable","trust","traceable","pipeline","cycle"],"s":["wifi-densepose-engine","wifi","densepose"]},
+{"id":"wifi-densepose-geo","family":"wifi-densepose-geo","kind":"crate","store":"worldgraph","version":null,"description":"Geospatial satellite integration — free satellite tiles, DEM, OSM, temporal tracking","keywords":[],"source":"worldgraph/wifi-densepose-geo/Cargo.toml","sourceSha256":"dd31eece3fe30d73637ec1fe92b9cdd1ca9f142d0f9b52ef3fd6f78bae4c32e5","variants":[],"readme":"Free satellite imagery, terrain elevation, and map data for RuView spatial sensing. No API keys required. Integrates your local sensor data (camera + WiFi CSI point cloud) with geographic context: - Satellite tiles — 10m Sentinel-2 cloudless imagery for your location - Elevation — SRTM 30m DEM for terrain modeling - Buildings + roads — OpenStreetMap data via Overpass API - Weather — Open Meteo current conditions + forecast - Geo-registration — maps local sensor coordinates to WGS84 - Temporal tracking — detects changes over time (construction, vegetation, weather) - Brain integration — stores geospatial context as ruOS brain memories Geospatial context is stored as brain memories: The ruOS a","readmeSource":"worldgraph/wifi-densepose-geo/README.md","product":"key:worldgraph:wifi-densepose-geo","canonical":"wifi-densepose-geo","tier":"T2","t":["wifi-densepose-geo","wifi","densepose","geo","geospatial","satellite","integration","free","tile","dem","osm","temporal","track"],"s":["wifi-densepose-geo","wifi","densepose","geo"]},
+{"id":"wifi-densepose-hardware","family":"wifi-densepose-hardware","kind":"crate","store":"ruview","version":null,"description":"Hardware interface abstractions for WiFi CSI sensors (ESP32, Intel 5300, Atheros)","keywords":["csi","esp32","hardware","sensor","wifi"],"source":"ruview/v2/crates/wifi-densepose-hardware/Cargo.toml","sourceSha256":"b6458752de889b50469c464850660bf59045c96b76eb8a61ce736383ffc154fa","variants":[],"readme":"Hardware interface abstractions for WiFi CSI sensors (ESP32, Intel 5300, Atheros). wifi-densepose-hardware provides platform-agnostic parsers for WiFi CSI data from multiple hardware sources. All parsing operates on byte buffers with no C FFI or hardware dependencies at compile time, making the crate fully portable and deterministic -- the same bytes in always produce the same parsed output. Until Realtek hardware and the radar report SDK arrive, the Rust-only simulator exercises the same versioned CFR/Range-FFT wire codec used by the future device adapter. Every frame is marked SYNTHETIC. Add --udp 127.0.0.1:5005 --realtime to stream one ADR-264 frame per UDP datagram. Replay files contain","readmeSource":"ruview/v2/crates/wifi-densepose-hardware/README.md","product":"key:ruview:wifi-densepose-hardware","canonical":"wifi-densepose-hardware","tier":"T0","t":["wifi-densepose-hardware","wifi","densepose","hardware","interface","abstraction","csi","sensor","esp32","intel","5300","athero"],"s":["wifi-densepose-hardware","wifi","densepose","hardware","csi","esp32","sensor"]},
+{"id":"wifi-densepose-mat","family":"wifi-densepose-mat","kind":"crate","store":"ruview","version":null,"description":"Mass Casualty Assessment Tool - WiFi-based disaster survivor detection","keywords":["detection","disaster","rescue","vital-signs","wifi"],"source":"ruview/v2/crates/wifi-densepose-mat/Cargo.toml","sourceSha256":"15fae51d2db0092cc8e0cec5db51b241ed710855fb2369625fc104b656dd3973","variants":[],"readme":"Mass Casualty Assessment Tool for WiFi-based disaster survivor detection and localization. wifi-densepose-mat uses WiFi Channel State Information (CSI) to detect and locate survivors trapped in rubble, debris, or collapsed structures. The crate follows Domain-Driven Design (DDD) with event sourcing, organized into three bounded contexts -- detection, localization, and alerting -- plus a machine learning layer for debris penetration modeling and vital signs classification. Use cases include earthquake search and rescue, building collapse response, avalanche victim location, flood rescue operations, and mine collapse detection. - Vital signs detection -- Breathing patterns, heartbeat signature","readmeSource":"ruview/v2/crates/wifi-densepose-mat/README.md","product":"key:ruview:wifi-densepose-mat","canonical":"wifi-densepose-mat","tier":"T0","t":["wifi-densepose-mat","wifi","densepose","mat","mass","casualty","assessment","wifi-based","disaster","survivor","detection","rescue","vital-signs","vital","sign"],"s":["wifi-densepose-mat","wifi","densepose","mat","detection","disaster","rescue","vital-signs","vital","sign"]},
+{"id":"wifi-densepose-nn","family":"wifi-densepose-nn","kind":"crate","store":"ruview","version":null,"description":"Neural network inference for WiFi-DensePose pose estimation","keywords":["deep-learning","densepose","inference","neural-network","onnx"],"source":"ruview/v2/crates/wifi-densepose-nn/Cargo.toml","sourceSha256":"91b39d79c06d31aaff1a75bcd62df1160cf09f032bf91e963eeb899b88bae326","variants":[],"readme":"Multi-backend neural network inference for WiFi-based DensePose estimation. wifi-densepose-nn provides the inference engine that maps processed WiFi CSI features to DensePose body surface predictions. It supports three backends -- ONNX Runtime (default), PyTorch via tch-rs, and Candle -- so models can run on CPU, CUDA GPU, or TensorRT depending on the deployment target. The crate implements two key neural components: - DensePose Head -- Predicts 24 body part segmentation masks and per-part UV coordinate regression. - Modality Translator -- Translates CSI feature embeddings into visual feature space, bridging the domain gap between WiFi signals and image-based pose estimation. - ONNX Runtime","readmeSource":"ruview/v2/crates/wifi-densepose-nn/README.md","product":"key:ruview:wifi-densepose-nn","canonical":"wifi-densepose-nn","tier":"T0","t":["wifi-densepose-nn","wifi","densepose","neural","network","inference","wifi-densepose","pose","estimation","deep-learning","deep","learn","neural-network","onnx"],"s":["wifi-densepose-nn","wifi","densepose","deep-learning","deep","learn","inference","neural-network","neural","network","onnx"]},
+{"id":"wifi-densepose-occworld-candle","family":"wifi-densepose-occworld-candle","kind":"crate","store":"ruview","version":null,"description":"ADR-147 — OccWorld TransVQVAE inference ported to Candle (Rust-native, no Python IPC)","keywords":[],"source":"ruview/v2/crates/wifi-densepose-occworld-candle/Cargo.toml","sourceSha256":"4056231d715b44c57bb355618d7400449f81c7cf270c414a28032672b3542518","variants":[],"product":"key:ruview:wifi-densepose-occworld-candle","canonical":"wifi-densepose-occworld-candle","tier":"T0","t":["wifi-densepose-occworld-candle","wifi","densepose","occworld","candle","adr-147","adr","147","transvqvae","inference","port","rust-native","python","ipc"],"s":["wifi-densepose-occworld-candle","wifi","densepose","occworld","candle"]},
+{"id":"wifi-densepose-physics","family":"wifi-densepose-physics","kind":"crate","store":"ruview","version":null,"description":"Bounded, provenance-preserving pose physics assessment for RuView","keywords":[],"source":"ruview/v2/crates/wifi-densepose-physics/Cargo.toml","sourceSha256":"bb46023f828cf24305fa26d9a326b4f906824105b4d893434a3cac9af8ad260d","variants":[],"product":"key:ruview:wifi-densepose-physics","canonical":"wifi-densepose-physics","tier":"T0","t":["wifi-densepose-physics","wifi","densepose","physic","bound","provenance-preserving","provenance","preserv","pose","assessment","ruview"],"s":["wifi-densepose-physics","wifi","densepose","physic"]},
+{"id":"wifi-densepose-pointcloud","family":"wifi-densepose-pointcloud","kind":"crate","store":"ruview","version":null,"description":"Real-time dense point cloud from camera depth + WiFi CSI tomography","keywords":[],"source":"ruview/v2/crates/wifi-densepose-pointcloud/Cargo.toml","sourceSha256":"ad4595abbca2059e482ab5b5c29f4d81101dc8edcedc09b7c1d24616474817b5","variants":[],"product":"key:ruview:wifi-densepose-pointcloud","canonical":"wifi-densepose-pointcloud","tier":"T0","t":["wifi-densepose-pointcloud","wifi","densepose","pointcloud","real","dense","point","cloud","camera","depth","csi","tomography"],"s":["wifi-densepose-pointcloud","wifi","densepose","pointcloud"]},
+{"id":"wifi-densepose-privshield","family":"wifi-densepose-privshield","kind":"crate","store":"ruview","version":null,"description":"WiFi Veil privacy shield (ADR-288): compliant-waveform countermeasure against unauthorized WiFi sensing. Deterministic attacker-vs-protector experiment that drives beamforming-feedback identity inference toward chance while preserving link throughput. Std-only pure-compute leaf, no async/server/RF-hardware deps; SYNTHETIC data only.","keywords":[],"source":"ruview/v2/crates/wifi-densepose-privshield/Cargo.toml","sourceSha256":"095a080ddf502cd6036b731a8252fcb527fdd5e4dcf1cca1460e087b0bf5a42b","variants":[],"readme":"WiFi Veil (codename VEIL — Verifiable Emission-shaping for Identity-Leakage prevention) is the compliant-waveform countermeasure counterpart to [BFLD](../wifi-densepose-bfld) (ADR-118/121). BFLD *detects* when beamforming feedback becomes identifying; WiFi Veil *acts* — it shapes a node's own outgoing beamforming feedback so that an unauthorized passive sniffer cannot re-identify people or infer activity, while a legitimate receiver (which shares the per-session key) sees an essentially unchanged link. This crate is a deterministic, dependency-free, WASM-ready reference and experiment — not a radio driver. It never emits RF. Every number it prints is SYNTHETIC, reproduced by cargo test -p wi","readmeSource":"ruview/v2/crates/wifi-densepose-privshield/README.md","product":"key:ruview:wifi-densepose-privshield","canonical":"wifi-densepose-privshield","tier":"T0","t":["wifi-densepose-privshield","wifi","densepose","privshield","veil","privacy","shield","adr-288","adr","288","compliant-waveform","compliant","waveform","countermeasure","against","unauthoriz","sens","deterministic","attacker-vs-protector","attacker","protector","experiment","drive","beamforming-feedback","beamform","feedback","identity","inference","toward","chance","while","preserv","link","throughput","std-only","std","only","pure-compute","pure","compute","leaf","async","server","rf-hardware","hardware","deps","synthetic"],"s":["wifi-densepose-privshield","wifi","densepose","privshield"]},
+{"id":"wifi-densepose-privshield-harness","family":"wifi-densepose-privshield-harness","kind":"npm","store":"ruview","version":"0.1.0","description":"Harness for wifi-densepose-privshield (WiFi Veil — compliant-waveform WiFi-sensing privacy shield, ADR-288/289)","keywords":[],"source":"ruview/harness/wifi-densepose-privshield/package.json","sourceSha256":"5c68eba5d00dbd392ccc002440e6d19b62cdc141e34f57adf1781ba08636b05d","variants":[],"readme":"A metaharness (contributor harness) for [wifi-densepose-privshield](../../v2/crates/wifi-densepose-privshield) — WiFi Veil, the compliant-waveform WiFi-sensing privacy shield (ADR-288). Defined by ADR-289. Or run without installing: guidance topics: overview, threat, countermeasure, compliance, optimization, experiment. It needs no dependencies or build step, so it works offline and in CI before npm install. WiFi Veil shapes a node's own beamforming feedback with keyed Givens rotations so a third-party passive sniffer cannot re-identify people, while a keyed receiver sees an essentially unchanged link. Compliant waveform controls only — never jamming. Reference results are SYNTHETIC / eviden","readmeSource":"ruview/harness/wifi-densepose-privshield/README.md","product":"key:ruview:wifi-densepose-privshield-harness","canonical":"wifi-densepose-privshield-harness","tier":"T0","t":["wifi-densepose-privshield-harness","wifi","densepose","privshield","harness","wifi-densepose-privshield","veil","compliant-waveform","compliant","waveform","wifi-sensing","sens","privacy","shield","adr-288","adr","288","289"],"s":["wifi-densepose-privshield-harness","wifi","densepose","privshield","harness"]},
+{"id":"wifi-densepose-py","family":"wifi-densepose-py","kind":"crate","store":"ruview","version":null,"description":"PyO3 bindings for the WiFi-DensePose Rust core — ships as the `wifi-densepose` PyPI wheel (ADR-117)","keywords":[],"source":"ruview/python/Cargo.toml","sourceSha256":"0659d7e92236b3922632709609ed310325ea7bd4f14ff4ea115c686bf2f3cd7e","variants":[],"readme":"Detect human presence, count people, read breathing and heart rate, and estimate skeletal pose — using only the WiFi signal already in your home. No cameras. No wearables. Works through walls and in the dark. wifi-densepose is the Python binding for the [RuView](https://github.com/ruvnet/RuView) sensing stack: a Rust core that turns the Channel State Information (CSI) emitted by ordinary WiFi chips into ambient-intelligence signals. The wheel ships compiled DSP for fast offline analysis, plus an opt-in Python client for talking to a live RuView sensing-server over WebSocket or MQTT. - 17-keypoint pose — full-body skeletal estimate from WiFi CSI, no camera - Vital signs — respiratory rate (6–","readmeSource":"ruview/python/README.md","product":"key:ruview:wifi-densepose-py","canonical":"wifi-densepose-py","tier":"T0","t":["wifi-densepose-py","wifi","densepose","pyo3","wifi-densepose","core","ship","pypi","wheel","adr-117","adr","117"],"s":["wifi-densepose-py","wifi","densepose"]},
+{"id":"wifi-densepose-rufield","family":"wifi-densepose-rufield","kind":"crate","store":"ruview","version":null,"description":"ADR-262 anti-corruption bridge: converts RuView WiFi-CSI sensing output into signed RuField FieldEvents (P0–P5 privacy mapping + ed25519 provenance)","keywords":[],"source":"ruview/v2/crates/wifi-densepose-rufield/Cargo.toml","sourceSha256":"f631d94d124178c4f6769293414178118c930f2033108fbf459423ff59b9d51a","variants":[],"product":"key:ruview:wifi-densepose-rufield","canonical":"wifi-densepose-rufield","tier":"T0","t":["wifi-densepose-rufield","wifi","densepose","rufield","adr-262","adr","262","anti-corruption","anti","corruption","bridge","convert","ruview","wifi-csi","csi","sens","output","sign","fieldevent","privacy","mapp","ed25519","provenance"],"s":["wifi-densepose-rufield","wifi","densepose","rufield"]},
+{"id":"wifi-densepose-ruvector","family":"wifi-densepose-ruvector","kind":"crate","store":"ruview","version":null,"description":"RuVector v2.0.4 integration layer — ADR-017 signal processing and MAT ruvector integrations","keywords":["csi","disaster-detection","ruvector","signal-processing","wifi"],"source":"ruview/v2/crates/wifi-densepose-ruvector/Cargo.toml","sourceSha256":"e0aec63f507c242a5d2f57c886ddea0381950fe4bee0422125b385b3e18cb7dc","variants":[],"readme":"RuVector v2.0.4 integration layer for WiFi-DensePose — ADR-017. This crate implements all 7 ADR-017 ruvector integration points for the signal-processing pipeline and the Multi-AP Triage (MAT) disaster-detection module. Add to your Cargo.toml (workspace member or direct dependency): Breathing buffer for 56 subcarriers × 60 s × 100 Hz:","readmeSource":"ruview/v2/crates/wifi-densepose-ruvector/README.md","product":"key:ruview:wifi-densepose-ruvector","canonical":"wifi-densepose-ruvector","tier":"T0","t":["wifi-densepose-ruvector","wifi","densepose","v2.0.4","integration","layer","adr-017","adr","017","signal","process","mat","csi","disaster-detection","disaster","detection","signal-processing"],"s":["wifi-densepose-ruvector","wifi","densepose","csi","disaster-detection","disaster","detection","signal-processing","signal","process"]},
+{"id":"wifi-densepose-sar","family":"wifi-densepose-sar","kind":"crate","store":"ruview","version":null,"description":"Coherent wideband RF tomography research crate (ADR-287): synthetic stepped-frequency multi-position measurement simulation + delay-and-sum backprojection reconstruction. SYNTHETIC/L0 only -- no wideband RF hardware backs this crate.","keywords":["backprojection","radar","rf-tomography","sar","simulation"],"source":"ruview/v2/crates/wifi-densepose-sar/Cargo.toml","sourceSha256":"6ca29133c80da74dfef3d7f047ba15dd9869a22ffd6e4806635b89c53dd4c675","variants":[],"readme":"Coherent wideband RF tomography research crate (ADR-287): synthetic stepped-frequency multi-position measurement simulation + delay-and-sum backprojection reconstruction of a 3D reflectivity field. This is not a hardware capability. It is the reconstruction primitive a handheld through-wall RF imaging device would need, validated against its own synthetic ground truth. Every number this crate produces is SYNTHETIC / evidence level L0 (ADR-282) until real wideband RF hardware (a VNA, SDR, or purpose-built radar front end) exists to feed it real measurements. See the crate-level doc comment in src/lib.rs for the full honesty boundary, and the tutorial at docs/tutorials/coherent-rf-tomography-b","readmeSource":"ruview/v2/crates/wifi-densepose-sar/README.md","product":"key:ruview:wifi-densepose-sar","canonical":"wifi-densepose-sar","tier":"T0","t":["wifi-densepose-sar","wifi","densepose","sar","coherent","wideband","tomography","research","crate","adr-287","adr","287","synthetic","stepped-frequency","step","frequency","multi-position","multi","position","measurement","simulation","delay-and-sum","delay","sum","backprojection","reconstruction","only","hardware","back","radar","rf-tomography"],"s":["wifi-densepose-sar","wifi","densepose","sar","backprojection","radar","rf-tomography","tomography","simulation"]},
+{"id":"wifi-densepose-sar-harness","family":"wifi-densepose-sar-harness","kind":"npm","store":"ruview","version":"0.1.0","description":"Harness for wifi-densepose-sar","keywords":[],"source":"ruview/harness/wifi-densepose-sar/package.json","sourceSha256":"61e0931ab75666c720e3d5529077a8e89c9d974d23dde0154910f82f743bc10d","variants":[],"readme":"Harness for wifi-densepose-sar This harness ships with the claude-code adapter. - npm run evolve / evolve:dry — Darwin Mode self-mutation of the harness's own config (@metaharness/darwin). - npm run route -- <e0> <e1> <e2> <e3> (after npm run build) — cost-optimal model routing via @metaharness/router. - npm run flywheel:dry — the SYNTHETIC @metaharness/flywheel self-improvement demo (propose → evaluate → gate → promote, signed + independently replayable). See CLAUDE.md's \"Darwin, router, flywheel\" section and the comments at the top of src/router.ts / src/flywheel.ts for what's real wiring vs. illustrative/synthetic data. - .harness/manifest.json / manifest.sha256 reflect the initial metaha","readmeSource":"ruview/harness/wifi-densepose-sar/README.md","product":"key:ruview:wifi-densepose-sar-harness","canonical":"wifi-densepose-sar-harness","tier":"T0","t":["wifi-densepose-sar-harness","wifi","densepose","sar","harness","wifi-densepose-sar"],"s":["wifi-densepose-sar-harness","wifi","densepose","sar","harness"]},
+{"id":"wifi-densepose-sensing-server","family":"wifi-densepose-sensing-server","kind":"crate","store":"ruview","version":null,"description":"Lightweight Axum server for WiFi sensing UI with RuVector signal processing","keywords":["csi","sensing","server","websocket","wifi"],"source":"ruview/v2/crates/wifi-densepose-sensing-server/Cargo.toml","sourceSha256":"e44e96378e00f966ecc6091a4aa7fd56c64a4c1ea59a0809bf6d522bd32703e8","variants":[],"readme":"Lightweight Axum server for real-time WiFi sensing with RuVector signal processing. wifi-densepose-sensing-server is the operational backend for WiFi-DensePose. It receives raw CSI frames from ESP32 hardware over UDP, runs them through the RuVector-powered signal processing pipeline, and broadcasts processed sensing updates to browser clients via WebSocket. A built-in static file server hosts the sensing UI on the same port. The crate ships both a library (wifi_densepose_sensing_server) exposing the training and inference modules, and a binary (sensing-server) that starts the full server stack. Integrates [wifi-densepose-wifiscan](../wifi-densepose-wifiscan) for multi-BSSID WiFi scanning per","readmeSource":"ruview/v2/crates/wifi-densepose-sensing-server/README.md","product":"key:ruview:wifi-densepose-sensing-server","canonical":"wifi-densepose-sensing-server","tier":"T0","t":["wifi-densepose-sensing-server","wifi","densepose","sens","server","lightweight","axum","signal","process","csi","websocket"],"s":["wifi-densepose-sensing-server","wifi","densepose","sens","server","csi","websocket"]},
+{"id":"wifi-densepose-signal","family":"wifi-densepose-signal","kind":"crate","store":"ruview","version":null,"description":"WiFi CSI signal processing for DensePose estimation","keywords":["csi","densepose","rust","signal-processing","wifi"],"source":"ruview/v2/crates/wifi-densepose-signal/Cargo.toml","sourceSha256":"c8dcde52ededde53e6f565cb427eb0460ff0a5abb00ff66888afd393b86d2ffc","variants":[],"readme":"State-of-the-art WiFi CSI signal processing for human pose estimation. wifi-densepose-signal implements six peer-reviewed signal processing algorithms that extract human motion features from raw WiFi Channel State Information (CSI). Each algorithm is traced back to its original publication and integrated with the [ruvector](https://crates.io/crates/ruvector-mincut) family of crates for high-performance graph and attention operations. - CSI preprocessing -- Noise removal, windowing, normalization via CsiProcessor. - Phase sanitization -- Unwrapping, outlier removal, and smoothing via PhaseSanitizer. - Feature extraction -- Amplitude, phase, correlation, Doppler, and PSD features. - Motion det","readmeSource":"ruview/v2/crates/wifi-densepose-signal/README.md","product":"key:ruview:wifi-densepose-signal","canonical":"wifi-densepose-signal","tier":"T0","t":["wifi-densepose-signal","wifi","densepose","signal","csi","process","estimation","signal-processing"],"s":["wifi-densepose-signal","wifi","densepose","signal","csi","signal-processing","process"]},
+{"id":"wifi-densepose-train","family":"wifi-densepose-train","kind":"crate","store":"ruview","version":null,"description":"Training pipeline for WiFi-DensePose pose estimation","keywords":["deep-learning","pose-estimation","training","wifi"],"source":"ruview/v2/crates/wifi-densepose-train/Cargo.toml","sourceSha256":"64f9d6615740d1848b39742420da011407c2492db17a4c0c1b1e7f0e3f94f843","variants":[],"readme":"Complete training pipeline for WiFi-DensePose, integrated with all five ruvector crates. wifi-densepose-train provides everything needed to train the WiFi-to-DensePose model: dataset loading, subcarrier interpolation, loss functions, evaluation metrics, and the training loop orchestrator. It supports both the MM-Fi dataset (NeurIPS 2023) and deterministic synthetic data for reproducible experiments. Without the tch-backend feature the crate still provides the dataset, configuration, and subcarrier interpolation APIs needed for data preprocessing and proof verification. - MM-Fi dataset loader -- Reads the MM-Fi multimodal dataset (NeurIPS 2023) from disk with memory-mapped .npy files. - Synth","readmeSource":"ruview/v2/crates/wifi-densepose-train/README.md","product":"key:ruview:wifi-densepose-train","canonical":"wifi-densepose-train","tier":"T0","t":["wifi-densepose-train","wifi","densepose","train","pipeline","wifi-densepose","pose","estimation","deep-learning","deep","learn","pose-estimation"],"s":["wifi-densepose-train","wifi","densepose","train","deep-learning","deep","learn","pose-estimation","pose","estimation"]},
+{"id":"wifi-densepose-vitals","family":"wifi-densepose-vitals","kind":"crate","store":"ruview","version":null,"description":"ESP32 CSI-grade vital sign extraction (ADR-021): heart rate and respiratory rate from WiFi Channel State Information","keywords":["breathing","csi","heart-rate","vital-signs","wifi"],"source":"ruview/v2/crates/wifi-densepose-vitals/Cargo.toml","sourceSha256":"cad41c1b1e57b15e9abd01a47048fcb8dc3b1d4459c89e8df0d8f624db57339a","variants":[],"readme":"ESP32 CSI-grade vital sign extraction: heart rate and respiratory rate from WiFi Channel State Information (ADR-021). wifi-densepose-vitals implements a four-stage pipeline that extracts respiratory rate and heart rate from multi-subcarrier CSI amplitude and phase data. The crate has zero external dependencies beyond tracing (and optional serde), uses #[forbid(unsafe_code)], and is designed for resource-constrained edge deployments alongside ESP32 hardware. 1. Preprocessing (CsiVitalPreprocessor) -- EMA-based static component suppression, producing per-subcarrier residuals that isolate body-induced signal variation. 2. Breathing extraction (BreathingExtractor) -- Bandpass filtering at 0.1--0","readmeSource":"ruview/v2/crates/wifi-densepose-vitals/README.md","product":"key:ruview:wifi-densepose-vitals","canonical":"wifi-densepose-vitals","tier":"T0","t":["wifi-densepose-vitals","wifi","densepose","vital","esp32","csi-grade","csi","grade","sign","extraction","adr-021","adr","021","heart","rate","respiratory","channel","state","information","breath","heart-rate","vital-signs"],"s":["wifi-densepose-vitals","wifi","densepose","vital","breath","csi","heart-rate","heart","rate","vital-signs","sign"]},
+{"id":"wifi-densepose-wasm-edge","family":"wifi-densepose-wasm-edge","kind":"crate","store":"ruview","version":null,"description":"WASM-compilable sensing algorithms for ESP32 edge deployment (ADR-040)","keywords":["dsp","esp32","sensing","wasm","wifi"],"source":"ruview/v2/crates/wifi-densepose-wasm-edge/Cargo.toml","sourceSha256":"1b8ef8294126676c66520e13b86a4c91b4d475aca34dc1e780d68296b3b29edd","variants":[],"product":"key:ruview:wifi-densepose-wasm-edge","canonical":"wifi-densepose-wasm-edge","tier":"T0","t":["wifi-densepose-wasm-edge","wifi","densepose","edge","wasm-compilable","compilable","sens","algorithm","esp32","deployment","adr-040","adr","040","dsp"],"s":["wifi-densepose-wasm-edge","wifi","densepose","edge","dsp","esp32","sens"]},
+{"id":"wifi-densepose-wifiscan","family":"wifi-densepose-wifiscan","kind":"crate","store":"ruview","version":null,"description":"Multi-BSSID WiFi scanning domain layer for enhanced Windows WiFi DensePose sensing (ADR-022)","keywords":["bssid","scanning","sensing","wifi","windows"],"source":"ruview/v2/crates/wifi-densepose-wifiscan/Cargo.toml","sourceSha256":"1e92c321ef364c1141796fa43e1657904798097f11737feff84f53763feafb2d","variants":[],"readme":"Multi-BSSID WiFi scanning for Windows-enhanced DensePose sensing (ADR-022). wifi-densepose-wifiscan implements the BSSID Acquisition bounded context for the WiFi-DensePose system. It discovers and tracks nearby WiFi access points, parses platform-specific scan output, and feeds multi-AP signal data into a sensing pipeline that performs motion detection, breathing estimation, attention weighting, and fingerprint matching. The crate uses #[forbid(unsafe_code)] and is designed as a pure-Rust domain layer with pluggable platform adapters. - BSSID registry -- Tracks observed access points with running RSSI statistics, band/radio type classification, and metadata. Types: BssidId, BssidObservation,","readmeSource":"ruview/v2/crates/wifi-densepose-wifiscan/README.md","product":"key:ruview:wifi-densepose-wifiscan","canonical":"wifi-densepose-wifiscan","tier":"T0","t":["wifi-densepose-wifiscan","wifi","densepose","wifiscan","multi-bssid","multi","bssid","scann","domain","layer","enhanc","window","sens","adr-022","adr","022"],"s":["wifi-densepose-wifiscan","wifi","densepose","wifiscan","bssid","scann","sens","window"]},
+{"id":"wifi-densepose-worldgraph","family":"wifi-densepose-worldgraph","kind":"crate","store":"worldgraph","version":null,"description":"ADR-139 — WorldGraph environmental digital twin (typed petgraph) for RuView","keywords":[],"source":"worldgraph/wifi-densepose-worldgraph/Cargo.toml","sourceSha256":"210e240f175997dd6c31b4589cb42b40f5038da7cef8415431fff2ae98ce09b9","variants":[],"readme":"The environmental digital twin for RF sensing — a typed, evidence-tracked graph of a building and the people in it. Part of the [RuView / WiFi-DensePose](https://github.com/ruvnet/RuView) project. Implements ADR-139. When you sense a space with WiFi/RF (people, motion, vital signs), you get a firehose of *frames*. What you actually want is a living map: which rooms exist, where the walls and doorways are, which sensors watch which zones, where each person is right now, and *why the system believes that* — with enough structure to reason over and enough provenance to trust. wifi-densepose-worldgraph is that map. It's a typed graph (built on [petgraph](https://crates.io/crates/petgraph)): - No","readmeSource":"worldgraph/wifi-densepose-worldgraph/README.md","product":"key:worldgraph:wifi-densepose-worldgraph","canonical":"wifi-densepose-worldgraph","tier":"T2","t":["wifi-densepose-worldgraph","wifi","densepose","worldgraph","adr-139","adr","139","environmental","digital","twin","typed","petgraph","ruview"],"s":["wifi-densepose-worldgraph","wifi","densepose","worldgraph"]},
+{"id":"wifi-densepose-worldmodel","family":"wifi-densepose-worldmodel","kind":"crate","store":"worldgraph","version":null,"description":"ADR-147 — OccWorld thin-client bridge: WorldGraph PersonTrack history → OccWorld Python subprocess → TrajectoryPrior","keywords":[],"source":"worldgraph/wifi-densepose-worldmodel/Cargo.toml","sourceSha256":"42f3ae18a1c9aaad28a5f33b85cdc7680375a98de74982e8cc44252a73160c56","variants":[],"readme":"Forward prediction for RF sensing — turn where people *were* into where they'll *be*, as occupancy + trajectory priors. Part of the [RuView / WiFi-DensePose](https://github.com/ruvnet/RuView) project. Implements ADR-147. [wifi-densepose-worldgraph](https://crates.io/crates/wifi-densepose-worldgraph) tells you what the room is *now* (who's where, the walls, the doorways). This crate answers the next question: what happens *next*? It's a thin, async client to an *occupancy world model* (OccWorld). You give it a short history of where people have been (their PersonTrack positions); it rasterizes that into 3-D occupancy grids, ships them to an OccWorld inference process, and gets back: - predict","readmeSource":"worldgraph/wifi-densepose-worldmodel/README.md","product":"key:worldgraph:wifi-densepose-worldmodel","canonical":"wifi-densepose-worldmodel","tier":"T2","t":["wifi-densepose-worldmodel","wifi","densepose","worldmodel","adr-147","adr","147","occworld","thin-client","thin","client","bridge","worldgraph","persontrack","history","python","subprocess","trajectoryprior"],"s":["wifi-densepose-worldmodel","wifi","densepose","worldmodel"]},
+{"id":"wifi-veil","family":"wifi-veil","kind":"crate","store":"wifi-veil","version":null,"description":"WiFi Veil (codename VEIL): compliant-waveform countermeasure against unauthorized WiFi sensing. Deterministic attacker-vs-protector experiment that drives beamforming-feedback identity inference toward chance while preserving link throughput. Std-only pure-compute leaf, no async/server/RF-hardware deps; SYNTHETIC data only.","keywords":["beamforming","privacy","security","sensing","wifi"],"source":"wifi-veil/Cargo.toml","sourceSha256":"2c248a6636387e68ffaa27d584414344d5bb4211abdf9ab8a60e3482dca61959","variants":[],"readme":"[▶ Open the live WiFi Veil Console](https://ruvnet.github.io/wifi-veil/) — no install, runs entirely in your browser (single self-contained HTML file, no network calls). A privacy firewall against unauthorized WiFi sensing — compliant waveform controls only, never jamming. WiFi Veil (codename VEIL — Verifiable Emission-shaping for Identity-Leakage prevention) shapes a node's own outgoing WiFi beamforming feedback so that an unauthorized passive sniffer cannot re-identify people or infer activity, while a legitimate receiver — which shares a per-session key — sees an essentially unchanged link. The threat — silent, device-free identification. Since WiFi 5, your device tells the router how to","readmeSource":"wifi-veil/README.md","product":"repo:wifi-veil","canonical":"wifi-veil","tier":null,"t":["wifi-veil","wifi","veil","codename","compliant-waveform","compliant","waveform","countermeasure","against","unauthoriz","sens","deterministic","attacker-vs-protector","attacker","protector","experiment","drive","beamforming-feedback","beamform","feedback","identity","inference","toward","chance","while","preserv","link","throughput","std-only","std","only","pure-compute","pure","compute","leaf","async","server","rf-hardware","hardware","deps","synthetic","privacy","security"],"s":["wifi-veil","wifi","veil","beamform","privacy","security","sens"]},
+{"id":"wifi-veil-harness","family":"wifi-veil-harness","kind":"npm","store":"wifi-veil","version":"0.1.0","description":"Harness for wifi-veil (WiFi Veil — compliant-waveform WiFi-sensing privacy shield, ADR-288/289)","keywords":[],"source":"wifi-veil/harness/package.json","sourceSha256":"4fdc0022863b5aa8cc85f067d0e3db2cac90bc88a26aa1c9ccfe6be046ebc983","variants":[],"readme":"A metaharness (contributor harness) for [wifi-veil](..) — WiFi Veil, the compliant-waveform WiFi-sensing privacy shield (ADR-288). Defined by ADR-289. Or run without installing: guidance topics: overview, threat, countermeasure, compliance, optimization, experiment. It needs no dependencies or build step, so it works offline and in CI before npm install. WiFi Veil shapes a node's own beamforming feedback with keyed Givens rotations so a third-party passive sniffer cannot re-identify people, while a keyed receiver sees an essentially unchanged link. Compliant waveform controls only — never jamming. Reference results are SYNTHETIC / evidence level L0 (reproduced by cargo test), never MEASURED","readmeSource":"wifi-veil/harness/README.md","product":"key:wifi-veil:wifi-veil-harness","canonical":"wifi-veil-harness","tier":null,"t":["wifi-veil-harness","wifi","veil","harness","wifi-veil","compliant-waveform","compliant","waveform","wifi-sensing","sens","privacy","shield","adr-288","adr","288","289"],"s":["wifi-veil-harness","wifi","veil","harness"]},
+{"id":"witness","family":"witness","kind":"crate","store":"autogenous","version":null,"description":"Autogenous witness plane: content-addressed hashing (SHA-256 over canonical JSON) + ed25519 signing/verification + append-only witness records with separate signing authorities (ADR-392 §13, gap 3).","keywords":[],"source":"autogenous/crates/witness/Cargo.toml","sourceSha256":"e40c7ab5d5baf8e791f21f43cede383f773888dd370899466247466e662abb07","variants":[],"readme":"witness is the cryptography plane of [Autogenous](../), a governed, self-evolving AI-defense runtime. It gives every piece of evidence a content hash, a real ed25519 signature, and a place in an append-only, tamper-evident chain — so provenance is verifiable, not asserted. If another crate ever held a placeholder signature: \"sig\" string, this is the crate that replaced it with actual cryptography. - content_hash<T: Serialize> — SHA-256 over the *canonical JSON* of any serializable value. Same bytes on every machine → the same hash → lineage that reconciles across deployments. - SigningAuthority — an ed25519 keypair. Generator, evaluator, and promotion controller each hold a separate authorit","readmeSource":"autogenous/crates/witness/README.md","product":"key:autogenous:witness","canonical":"witness","tier":null,"t":["witness","autogenous","plane","content-addressed","content","addres","hash","sha-256","sha","256","canonical","json","ed25519","sign","verification","append-only","append","only","record","separate","authority","adr-392","adr","392","gap"],"s":["witness"]},
+{"id":"worldgraph-stream","family":"worldgraph-stream","kind":"crate","store":"worldgraph","version":null,"description":"Versioned, authorized realtime delta streaming for WorldGraph","keywords":[],"source":"worldgraph/worldgraph-stream/Cargo.toml","sourceSha256":"09af97e8a1be453bd475f8930aca231082170d970522ac8eef019c3262dd1519","variants":[],"product":"key:worldgraph:worldgraph-stream","canonical":"worldgraph-stream","tier":"T2","t":["worldgraph-stream","worldgraph","stream","version","authoriz","realtime","delta"],"s":["worldgraph-stream","worldgraph","stream"]},
+{"id":"worldgraph-wasm","family":"worldgraph","kind":"crate","store":"worldgraph","version":null,"description":"ADR-200 — WebAssembly bridge exposing the WorldGraph digital twin to the SuperSplat (PlayCanvas/WebGPU) browser visualizer","keywords":[],"source":"worldgraph/worldgraph-wasm/Cargo.toml","sourceSha256":"a310f962fc1bc14f67261b349958c861bd14abd87b2fd684310049e9df1f6625","variants":[],"product":"repo:worldgraph","canonical":"worldgraphs","tier":"T2","t":["worldgraph-wasm","worldgraph","adr-200","adr","200","bridge","expos","digital","twin","supersplat","playcanva","webgpu","browser","visualizer"],"s":["worldgraph-wasm","worldgraph"]},
+{"id":"worldgraphs","family":"worldgraphs","kind":"npm","store":"worldgraph","version":"0.1.3","description":"AI coding agent for building digital twins with worldgraph — architect, implement, review, and test the Rust spatial/sensor-graph crates","keywords":["agent-harness","ai-agent","ambient-intelligence","coding-agent","digital-twin","environmental-monitoring","knowledge-graph","occupancy","provenance","rust","sensor-fusion","spatial-computing","wifi-sensing","world-model"],"source":"worldgraph/package.json","sourceSha256":"6fe049ad5bdb120e6dd639c0dff9893060d339e79364f9ddcea446c87eca37b5","variants":[],"readme":"A privacy-aware environmental digital twin — and an AI agent that helps you build one. worldgraph is two things in one project: 1. A Rust library that models a physical space as a typed, provenance-tracked graph — rooms, zones, sensors, people, and *beliefs* about what's happening — geospatially grounded and able to forecast occupancy. 2. An AI coding agent (npx worldgraphs) — architect → implement → review → test — that helps you build digital twins and spatial/sensor applications on top of it. [Open RuLab](https://ruvnet.github.io/worldgraph/) · [Implementation and controls](rulab/README.md) · [Architecture decision](docs/adr/205-rulab-temporal-gaussian-world.md) · [Review evidence](docs/m","readmeSource":"worldgraph/README.md","product":"repo:worldgraph","canonical":"worldgraphs","tier":"T2","t":["worldgraph","coding","digital","twin","architect","implement","review","test","spatial","sensor-graph","sensor","graph","crate","agent-harness","harness","ai-agent","ambient-intelligence","ambient","intelligence","coding-agent","digital-twin","environmental-monitoring","environmental","monitor","knowledge-graph","knowledge","occupancy","provenance","sensor-fusion","fusion","spatial-computing","comput","wifi-sensing","wifi","sens","world-model","world","model"],"s":["worldgraph","agent-harness","harness","ai-agent","ambient-intelligence","ambient","intelligence","coding-agent","coding","digital-twin","digital","twin","environmental-monitoring","environmental","monitor","knowledge-graph","knowledge","graph","occupancy","provenance","sensor-fusion","sensor","fusion","spatial-computing","spatial","comput","wifi-sensing","wifi","sens","world-model","world","model"]},
+{"id":"xtask","family":"xtask","kind":"crate","store":"midstream","version":null,"description":"midstream task runner (cargo xtask <command>) — replaces the legacy publish_*.sh / install.sh / setup.sh scripts. See ADR-0037.","keywords":[],"source":"midstream/xtask/Cargo.toml","sourceSha256":"a7250cfab7a13cd98f5b55d0027bd78472b9dbb78491f7bded21093c204810e0","variants":[],"product":"key:midstream:xtask","canonical":"xtask","tier":"T1","t":["xtask","midstream","task","runner","cargo","command","replace","legacy","publish","install.sh","setup.sh","script","see","adr-0037","adr","0037"],"s":["xtask"]}
+]}