local-memory-mcp 1.2.0 → 1.3.0

package/CHANGELOG.md

@@ -1,219 +1,111 @@
-# Local Memory v1.2.0
+# Local Memory v1.3.0
 
-**Release Date**: 2025-12-11
+**Release Date**: 2026-01-10
 **Compatibility**: Fully backwards compatible
 
 ---
 
 ## Summary
 
-This major release addresses critical security vulnerabilities, implements complete MCP 2025-11-25 specification compliance, delivers full CLI interface completeness, and resolves multiple high-priority bugs. Over 10 significant features and 10 GitHub issues were resolved, representing the largest development cycle in Local Memory's history.
+This major release introduces several **World Memory** R&D features, a complete knowledge management architecture that transforms Local Memory from simple RAG storage and retrieval into an intelligent knowledge system. The release includes a four-level knowledge hierarchy, automatic contradiction detection, epistemic gap tracking, and knowledge evolution mechanisms. Ten new tools enable sophisticated reasoning workflows while maintaining full backward compatibility.
 
-This release implements the following GitHub issues:
+## Added
+
+- **World Memory Knowledge Hierarchy** (L0 Observations, L1 Learnings, L2 Patterns, L3 Schemas)
+- **Contradiction Detection System** - 5-layer heuristic validation
+- **Epistemic Gap Tracking** - Questions system for unknown unknowns
+- **Knowledge Evolution** - Validation, promotion, decay mechanisms
+- **New MCP Tools**: observe, reflect, question, resolve, evolve, bootstrap, predict, explain, counterfactual, validate
+- **Tool Deprecation Strategy** - Migration path for legacy tools
+
+## Fixed
+
+- Session ID validation failing in Claude Desktop
+- Orphaned contradiction resolution
+- Level/weight storage in observe tool
+
+## Changed
+
+- Major modular refactoring (store.go, server.go, handlers.go)
+- Tool migration: store_memory -> observe, relationships -> relate, stats -> status
 
-- **Issue #51**: CRITICAL - Fixed domain filtering returning all memories instead of filtered results
-- **Issue #52**: HIGH - Fixed custom field selection being completely ignored
-- **Issue #53**: CRITICAL - Fixed response format logic producing 258% MORE tokens instead of fewer
-- **Issue #54**: HIGH - Fixed relationship creation returning no confirmation
-- **Issue #55**: HIGH - Fixed summarization tool execution failures
-- **Issue #56**: MEDIUM - Fixed semantic search without AI returning empty results
-- **Issue #57**: MEDIUM - Fixed metadata date issues and zero-value timestamps
-- **Issue #58**: SECURITY - Updated golang.org/x/crypto dependencies for security vulnerabilities
-- **Issue #59**: Performance - Fixed default responses exceeding 1000-token budget
-- **Issue #60**: CRITICAL - Migrated NPM classic tokens before December 9th deadline
+---
+
+*See CHANGELOG_v1.3.0.md in the repository for full details.*
+
+---
+
+# Local Memory v1.2.1
+
+**Release Date**: 2026-01-02
+**Compatibility**: Fully backwards compatible
+
+---
+
+## Summary
+
+This patch release fixes critical license activation issues and improves CLI command reliability. The `--accept-terms` flag now correctly activates licenses without character encoding problems, and CLI commands have been updated to eliminate missing flags and reduce unnecessary output.
 
 ## Added
 
-### New Features
-- **Complete CLI Interface Completeness** - All 4 phases delivered with full MCP/REST API parity
-  - Phase 1: Core memory operations (`get`, `list`)
-  - Phase 2: Relationship management (`find_related`, `discover`, `relate`, `map_graph`)
-  - Phase 3: Management operations (9 commands for categories, domains, sessions)
-  - Phase 4: Advanced features via consolidated `analyze` command
-- **MCP 2025-11-25 Specification Compliance** - Updated to latest MCP specification
-  - Protocol version updated from 2024-11-05 to 2025-11-25
-  - JSON Schema 2020-12 compliance for all tool schemas
-  - Tool icons support (💾 🔍 🧠 📊 🏷️ etc.)
-  - Enhanced server description fields
-- **Comprehensive Evaluation Framework** - Advanced testing across all interfaces
-  - Cross-interface consistency validation (MCP, REST, CLI)
-  - Mandatory measurements framework (M1-M4 validation)
-  - Security testing and adversarial evaluation capabilities
-- **Agent Harness Integration** - Long-running agent workflow patterns
-  - Session tracking and progress state management
-  - Feature tracking system with JSON-based persistence
-  - Git workflow enhancement with automatic progress markers
-  - Local Memory integration for persistent context
-
-### CLI Commands Added
-- `local-memory get <memory-id>` - Retrieve specific memory by ID
-- `local-memory list` - List memories with filtering and pagination
-- `local-memory find_related <memory-id>` - Find related memories
-- `local-memory discover` - AI-powered relationship discovery
-- `local-memory relate <source-id> <target-id>` - Create relationships
-- `local-memory map_graph <memory-id>` - Visualize memory graphs
-- `local-memory list_categories` - Manage memory categories
-- `local-memory create_category <name>` - Create new categories
-- `local-memory categorize <memory-id> <category>` - Assign categories
-- `local-memory list_domains` - Manage memory domains
-- `local-memory create_domain <name>` - Create new domains
-- `local-memory list_sessions` - Session management
-- `local-memory analyze` - Advanced analysis (temporal patterns, learning progression)
+- Comprehensive documentation suite with validation reports
+- Organized validation reports in dedicated directory structure
+- Enhanced CLI error messages for better user guidance
 
 ## Fixed
 
-### Critical Bug Fixes
-- **Domain filtering completely broken** (#51) - Fixed filter logic returning all memories instead of domain-specific results
-- **Response format logic inverted** (#53) - Fixed concise format producing 258% MORE tokens instead of 70% fewer
-- **Summarization tool execution fails** (#55) - Fixed analysis tool execution with enhanced CLI analyze command
-- **Custom field selection ignored** (#52) - Implemented SQL-level field selection with 6-12% token reduction
-- **Relationship creation no confirmation** (#54) - Added proper confirmation responses with relationship details
-- **Semantic search without AI empty results** (#56) - Enhanced text search with improved keyword extraction
-- **Metadata date issues** (#57) - Fixed zero-value timestamps and inaccurate memory counts
-
-### Security Fixes
-- **Input sanitization vulnerabilities** - Implemented comprehensive security hardening
-  - SQL injection protection across all interfaces
-  - XSS prevention with content escaping
-  - Template injection protection
-  - Security headers middleware for REST API
-- **SQL injection defense algorithm refinement** - Fixed overly aggressive pattern matching
-  - Resolved false positives blocking legitimate quoted content like "user's guide"
-  - Enhanced context-aware patterns requiring injection indicators
-  - Preserved security against actual attacks while allowing technical documentation
-  - Added comprehensive test suite with 38 security validation tests
-- **Dependency security updates** (#58) - Updated golang.org/x/crypto to address 5 MODERATE vulnerabilities
-- **NPM authentication security** (#60) - Migrated to granular access tokens before December 9th deadline
-
-### Performance Improvements
-- **Token efficiency optimization** (#59) - Implemented intelligent format selection with 30-70% token reduction
-- **Ultra-concise formats** - Added ResponseFormatUltra and ResponseFormatMicro achieving 94% token reduction
-- **Enhanced token estimation** - Improved accuracy from 159% error to 79% through calibration
-- **Cross-interface consistency** - Standardized token optimization across MCP, REST, and CLI
+- **License activation bug**: Fixed character normalization issues preventing `--accept-terms` flag from properly activating licenses
+- **CLI command improvements**: Resolved missing flags and eliminated useless output from various CLI commands
+- **Documentation accuracy**: Corrected and validated all public documentation files
 
 ## Technical Details
 
-### Architecture Improvements
-- **Agent Harness Pattern Implementation**: Complete workflow system for long-running agent sessions
-- **MCP Specification Compliance**: Updated to 2025-11-25 with full backward compatibility
-- **Cross-Interface Parity**: 100% feature compatibility across MCP, REST API, and CLI
-- **Security Hardening**: 15/15 security compliance rating achieved
-
-### Performance Metrics
-- **Token Reduction**: Ultra format 94.3%, Concise format 70.0% (exceeded targets)
-- **Security Score**: 15/15 (100%) excellent security posture
-- **Interface Coverage**: 11 MCP tools, 16 CLI commands, 25+ REST endpoints
-- **Evaluation Score**: 95/100 (EXCELLENT) across all interfaces
-
-### Security Enhancements
-- **Multi-layer Protection**: SQL injection, XSS, template injection all blocked
-- **Refined SQL Injection Defense**: Context-aware patterns with 0% false positives for legitimate content
-- **Boundary Validation**: Importance values properly constrained to 1-10 range
-- **Security Headers**: Added comprehensive security headers for REST API
-- **Dependency Updates**: All MODERATE security vulnerabilities resolved
+### License Activation Fix
+- Implemented proper UTF-8 character normalization in license acceptance flow
+- Added validation for non-breaking spaces and Unicode character variants
+- Enhanced error handling for license activation edge cases
+
+### CLI Improvements
+- Added missing flags to subcommands for consistent behavior
+- Removed unnecessary verbose output that cluttered terminal sessions
+- Improved error messages with actionable guidance
+
+### Documentation Updates
+- Created comprehensive validation reports for all documentation
+- Organized reports into `docs/public/validation-reports/` directory
+- Validated accuracy of CLI reference, configuration, and API documentation
 
 ---
 
 ## Issue Details
 
-### #51 - CRITICAL: Domain Filtering Broken
-**Status**: CLOSED
-**Impact**: Domain-based search filtering returned ALL memories instead of filtered results
-**Resolution**: Fixed domain filter logic in search query construction
-
-### #52 - HIGH: Custom Field Selection Ignored
-**Status**: CLOSED
-**Impact**: Custom field selection with `response_format: "custom"` was completely ignored
-**Resolution**: Implemented SQL-level field selection with performance optimization
-
-### #53 - CRITICAL: Response Format Logic Inverted
-**Status**: CLOSED
-**Impact**: Concise format produced 258% MORE tokens instead of 70% fewer
-**Resolution**: Fixed response format logic in formatters with proper token optimization
-
-### #54 - HIGH: Relationship Creation No Confirmation
-**Status**: CLOSED
-**Impact**: Relationship creation returned empty response with no success/failure indication
-**Resolution**: Enhanced response formatting with relationship ID and confirmation details
-
-### #55 - HIGH: Summarization Tool Execution Fails
-**Status**: CLOSED
-**Impact**: Analysis tool execution failed completely for memory summarization
-**Resolution**: Implemented complete CLI analyze command with enhanced functionality
-
-### #56 - MEDIUM: Semantic Search Without AI Returns Empty
-**Status**: CLOSED
-**Impact**: Text-based search without AI backend returned 0 results for valid queries
-**Resolution**: Enhanced text search with improved keyword extraction and relevance scoring
-
-### #57 - MEDIUM: Metadata Date Issues
-**Status**: CLOSED
-**Impact**: Zero-value timestamps (0001-01-01) and inaccurate memory counts
-**Resolution**: Fixed metadata initialization and timestamp handling logic
-
-### #58 - SECURITY: Update golang.org/x/crypto Dependencies
-**Status**: CLOSED
-**Impact**: 5 MODERATE security vulnerabilities (CVE-2025-47914, CVE-2025-58181, etc.)
-**Resolution**: Updated to golang.org/x/crypto v0.45.0+ resolving all vulnerabilities
-
-### #59 - TOKEN EFFICIENCY: Default Responses Exceed Budget
-**Status**: CLOSED
-**Impact**: Default responses frequently exceeded 1000-token efficiency targets
-**Resolution**: Implemented intelligent format selection with automatic optimization
-
-### #60 - CRITICAL: Classic npm tokens stop working December 9th
-**Status**: CLOSED
-**Impact**: NPM authentication would break on December 9th, 2025
-**Resolution**: Successfully migrated to granular access tokens with verified publishing
+No GitHub issues were directly addressed in this release. This patch focuses on critical bug fixes discovered through internal testing and user feedback.
 
 ---
 
 ## Migration Guide
 
-**No breaking changes** - This release is fully backwards compatible with v1.1.7 and earlier.
+**No breaking changes** - This release is fully backwards compatible with v1.2.0 and earlier.
 
-### New CLI Commands Usage
+### License Activation
+If you experienced issues with `--accept-terms` not activating your license, update to v1.2.1:
 
 ```bash
-# Core memory operations
-local-memory get <memory-id>
-local-memory list --domain="ai-research" --limit=10
-
-# Relationship management
-local-memory find_related <memory-id> --min_strength=0.7
-local-memory relate <source-id> <target-id> --type="references"
-local-memory map_graph <memory-id> --depth=2
-
-# Advanced analysis
-local-memory analyze --type="temporal_patterns" --temporal_timeframe="month"
-local-memory analyze --type="temporal_patterns" --temporal_analysis_type="learning_progression" --limit=50
-```
-
-### Enhanced Security Features
+# Update to v1.2.1
+npm update -g local-memory-mcp
 
-```javascript
-// All injection attempts are now automatically blocked
-search({
-  query: "'; DROP TABLE memories; --",  // Blocked
-  content: "<script>alert('xss')</script>"  // Escaped
-})
+# Verify license activation works
+local-memory --accept-terms
 ```
 
-### Token Optimization
-
-```javascript
-// Intelligent format selection for token efficiency
-search({
-  query: "machine learning",
-  // Automatically selects optimal format for <1000 tokens
-  response_format: "intelligent"  // New default
-})
-
-// Ultra-efficient formats for agent workflows
-search({
-  query: "neural networks",
-  response_format: "micro"  // 94% token reduction
-})
+### CLI Command Usage
+All CLI commands now have consistent flag behavior:
+
+```bash
+# Commands now properly support all documented flags
+local-memory search --query "your search" --limit 10
+local-memory remember "your memory" --tags "tag1,tag2"
+local-memory forget memory-uuid
 ```
 
 ## Getting Started
@@ -223,13 +115,9 @@ search({
 npm update -g local-memory-mcp
 
 # Verify version
-local-memory --version  # Should show v1.2.0
-
-# Test new CLI commands
-local-memory list --help
-local-memory analyze --help
+local-memory --version # Should show v1.2.1
 ```
 
 ---
 
-*Local Memory v1.2.0 - Complete feature parity, enhanced security, and next-generation agent intelligence*
+*Local Memory v1.2.1 - Critical bug fixes and documentation improvements*

package/{README.md → README}

@@ -1,6 +1,6 @@
 # Local Memory (localmemory.co)
 
-**Version 1.2.0** - AI-powered persistent memory system for developers working with Claude Code, Claude Desktop, Gemini, Codex, Cursor, and other coding agents.
+**Version 1.3.0** - AI-powered persistent memory system for developers working with Claude Code, Claude Desktop, Gemini, Codex, Cursor, and other coding agents.
 
 Transform your AI interactions with intelligent memory that grows smarter over time. Build persistent knowledge bases, track learning progression, and maintain context across conversations. Local Memory provides MCP, REST API, and Command-line interfaces for you and your coding agents to store, retrieve, discover, and analyze memories to give your agent the right context for it's work and tasks.
 
@@ -98,12 +98,26 @@ Add to your Claude Desktop configuration file:
 
 When configured, Claude will have access to these memory tools:
 
-- `store_memory` - Save important information
+**Core Tools:**
+- `observe` - Record observations with knowledge level support (L0-L3)
 - `search` - Find memories with semantic search
-- `analysis` - AI-powered Q&A on stored knowledge
-- `relationships` - Discover and map memory connections
-- `categories` - Organize memories intelligently
-- Plus comprehensive memory management tools
+- `bootstrap` - Initialize sessions with knowledge context
+
+**Knowledge Evolution:**
+- `reflect` - Process observations into learnings
+- `evolve` - Validate, promote, or decay knowledge
+- `question` - Track epistemic gaps and contradictions
+- `resolve` - Resolve contradictions and answer questions
+
+**Reasoning Tools:**
+- `predict` - Generate predictions from patterns
+- `explain` - Trace causal paths between states
+- `counterfactual` - Explore "what if" scenarios
+- `validate` - Check knowledge graph integrity
+
+**Relationship Tools:**
+- `relate` - Create relationships between memories
+- Plus memory management tools (update, delete, get)
 
 ## REST API
 
@@ -125,6 +139,9 @@ curl "http://localhost:3002/api/v1/health"
 ## Service Management
 
 ```bash
+# Start daemon
+local-memory start
+
 # Check daemon status
 local-memory status
 
@@ -141,26 +158,20 @@ local-memory kill_all
 local-memory doctor
 ```
 
-## What's New in v1.1.8
-
-- **Complete CLI Interface Completeness** - All 4 phases delivered with full MCP/REST API parity
-  - Core memory operations (`get`, `list`)
-  - Relationship management (`find_related`, `discover`, `relate`, `map_graph`)
-  - Management operations (9 commands for categories, domains, sessions)
-  - Advanced features via consolidated `analyze` command
-- **MCP 2025-11-25 Specification Compliance** - Updated to latest MCP specification
-  - Protocol version is still 2024-11-05 since Claude CLI does not yet support version 2025-11-25
-  - JSON Schema 2020-12 compliance for all tool schemas
-  - Tool icons support (💾 🔍 🧠 📊 🏷️ etc.)
-  - Enhanced server description fields
-- **Comprehensive Evaluation Framework** - Advanced testing across all interfaces
-  - Cross-interface consistency validation (MCP, REST, CLI)
-  - Mandatory measurements framework (M1-M4 validation)
-  - Security testing and adversarial evaluation capabilities
+## What's New in v1.3.0
+
+- **World Memory Knowledge Hierarchy** - Four-level knowledge progression system
+  - L0 Observations: Raw intake, ephemeral
+  - L1 Learnings: Candidate insights, volatile
+  - L2 Patterns: Validated generalizations, durable
+  - L3 Schemas: Theoretical frameworks, permanent
+- **Contradiction Detection** - Automatic conflict identification using 5-layer heuristics
+- **Epistemic Gap Tracking** - Track unknown unknowns through dedicated questions system
+- **Knowledge Evolution** - Validation, promotion, and decay mechanisms
+- **10 New MCP Tools** - observe, reflect, question, resolve, evolve, bootstrap, predict, explain, counterfactual, validate
 
 ## System Requirements
 
-- **Node.js**: 16.0.0 or higher
 - **RAM**: 512MB minimum
 - **Storage**: 100MB for binaries, variable for database
 - **Platforms**: macOS (Intel/ARM64), Linux x64, Windows x64
@@ -185,6 +196,7 @@ local-memory license status
 - **Architecture**: [localmemory.co/architecture](https://localmemory.co/architecture)
 - **Prompts**: [localmemory.co/prompts](https://localmemory.co/prompts)
 - **Support**: [Local Memory Discord](https://discord.gg/rMmn8xP3fZ)
+- **Releases Repo**: [Local Memory Releases](https://github.com/danieleugenewilliams/local-memory-releases)
 
 ---
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "local-memory-mcp",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "description": "Local Memory MCP Server - AI-powered persistent memory system for Claude Code, Claude Desktop, Gemini, Codex, OpenCode and other MCP-compatible tools",
   "keywords": [
     "mcp",
@@ -83,4 +83,4 @@
     "type": "individual",
     "url": "https://localmemory.co"
   }
-}
+}