claude-flow-novice 2.18.12 → 2.18.14

package/.claude/skills/cfn-session-handoff/SKILL.md

@@ -1,53 +1 @@
----
-name: cfn-session-handoff
-type: skill
-description: Session handoff functionality is provided by the handoff-coordinator agent
-version: 1.0.0
-status: delegated
-triggers:
-  - When: Session transition or context reset is needed
-  - Who: Teams requiring handoff documentation
-usage: See handoff-coordinator agent at .claude/agents/cfn-dev-team/coordinators/handoff-coordinator.md
----
-
-# CFN Session Handoff
-
-## Status: DELEGATED TO AGENT
-
-This skill's functionality is now provided by the **handoff-coordinator agent**.
-
-## Actual Implementation Location
-
-`.claude/agents/cfn-dev-team/coordinators/handoff-coordinator.md`
-
-## Usage
-
-Instead of using this skill, invoke the handoff-coordinator agent directly:
-
-```bash
-npx claude-flow-novice agent-spawn handoff-coordinator \
-  --task-id "handoff-$(date +%s)" \
-  --prompt "Create session handoff for [domain], duration: [hours], focus: [areas]"
-```
-
-## Features Provided by Agent
-
-- Extracts recent context from conversation history
-- Generates comprehensive handoff documentation
-- Creates smoke test scripts for validation
-- Maintains context continuity across sessions
-
-## Migration Note
-
-This skill was deprecated in favor of the agent-based approach which provides:
-- Better integration with CFN Loop
-- More sophisticated context extraction
-- Direct agent-to-agent communication
-- Automated session state management
-
-## Documentation Retained
-
-This SKILL.md is retained for:
-- Historical reference
-- Search visibility
-- Redirecting users to the actual implementation
+Task("handoff-coordinator", prompt with session context)

package/.claude/skills/cfn-skill-management/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: cfn-skill-management
-description: Skill loading, propagation, and building for CFN
+description: "Skill loading, propagation, and building for CFN. Use when dynamically discovering and loading skills, deploying skills across environments, or creating new skill scaffolding."
 version: 1.0.0
 tags: [mega-skill, skills, loader, builder, deployment]
 status: production

package/.claude/skills/cfn-task-decomposition/SKILL.md

@@ -0,0 +1 @@
+./decompose-task.sh --task-id=ts-errors --description="Fix all TypeScript errors in the codebase" --tool-budget=12 --complexity=high

package/.claude/skills/cfn-task-intelligence/SKILL.md

@@ -1,18 +1,39 @@
-task-intelligence/
-├── SKILL.md                          # This file
-├── cfn-task-intelligence.sh          # Main CLI interface
-└── lib/
-    ├── classifier/                   # Task classification
-    │   ├── SKILL.md
-    │   └── classify-task.sh
-    ├── complexity/                   # Complexity estimation
-    │   ├── SKILL.md
-    │   └── estimate-complexity.sh
-    ├── specialist/                   # Specialist recommendation
-    │   ���── SKILL.md
-    │   └── recommend-specialist.sh
-    └── integration/                  # System integration layer
-        ├── SKILL.md
-        ├── pre-execution-hook.sh
-        ├── post-feedback-hook.sh
-        └── learning-loop.sh
+---
+name: cfn-task-intelligence
+description: Classify tasks (type/domain), estimate complexity/iterations, recommend specialists from feedback themes
+version: 1.0.0
+tags: classification, complexity, specialists, cfn-loop
+status: production
+---
+
+# What it does
+Analyzes task descriptions → classifies type (software, content, research, design, infrastructure, data), detects domains (frontend, backend, security, devops, testing, database, docs), estimates complexity + iterations, recommends specialists from recurring feedback.
+
+# When to use
+1. New task → Determine CFN Loop config and agents
+2. Estimate iterations → How many loops before gates pass
+3. Recurring issues → After 3 security feedback, recommend security-specialist
+4. Multi-agent coordination → Identify domains and dependencies
+
+# When NOT to use
+1. Task already classified → Proceed with execution
+2. Simple type check → Use /write-plan instead
+3. Agent failure debugging → Use /cfn-check-errors
+4. Adjust teams mid-execution → Use /cfn-agent-lifecycle
+
+# How to use
+Classify: ./cfn-task-intelligence.sh --task-description "..." --mode classify
+Complexity: --mode complexity (returns iterations 2-7, confidence 0.70-0.80)
+Specialist: --mode specialist --current-loop3 "..." --feedback-themes "security,auth" --recurring-count 3
+Full: --mode all
+
+# Parameters
+--task-description, --mode (classify/complexity/specialist/all), --current-loop3, --feedback-themes, --recurring-count
+
+# Expected output
+Classify: {task_type, domains[], complexity, keyword_counts}
+Complexity: {complexity, estimated_iterations, confidence, factors, reasoning}
+Specialist: {add_specialist, reasoning, new_loop3_agents[]}
+
+# Real-world
+"Microservice with K8s and streaming" → infrastructure, [backend,devops,database], high, 6 iterations → spawn DevOps + backend, enterprise mode

package/.claude/skills/cfn-task-planning/SKILL.md

@@ -1,15 +1,40 @@
-# 1. Classify task
-TASK_JSON=$(./cli/classify-task.sh "$TASK_DESCRIPTION" --format=json)
-TASK_TYPE=$(echo "$TASK_JSON" | jq -r '.task_type')
-COMPLEXITY=$(echo "$TASK_JSON" | jq -r '.complexity')
+---
+name: cfn-task-planning
+description: Classify tasks, initialize structured configs with scope boundaries, decompose complex tasks
+version: 1.0.0
+tags: planning, classification, scope, decomposition, task-mode
+status: production
+---
 
-# 2. Initialize config
-./cli/init-config.sh --task-id "$TASK_ID" --description "$TASK_DESCRIPTION"
+## What it does
+Three-phase planning: (1) classify by type/complexity → agent specializations, (2) initialize config with scope boundaries, deliverables, acceptance criteria, (3) decompose large tasks into subtasks within tool budgets.
 
-# 3. Decompose if complex
-if [ "$COMPLEXITY" = "high" ]; then
-  SUBTASKS=$(./cli/decompose-task.sh --task-id "$TASK_ID" --description "$TASK_DESCRIPTION")
-fi
+## When to use (4 triggers)
+1. Starting CFN Loop Task Mode → Generate scope config before spawning
+2. Analyzing complexity → Select right agents and iteration thresholds
+3. Breaking down epics → Sequential subtasks when one agent can't complete
+4. Scope contracts → Establish in/out-of-scope before implementation
 
-# 4. Store audit data after completion
-./lib/audit/store-task-audit.sh --task-id "$TASK_ID" --result "success" --confidence 0.92
+## When NOT to use (4 anti-patterns)
+1. Well-defined and scoped → Go straight to spawning
+2. Real-time classification during execution → Planning phase only
+3. CLI mode with Redis → CLI stores in Redis, this is for Task Mode configs
+4. Simple single-step → Skip for trivial changes
+
+## How to use
+Step 1 Classify: `./classify-task.sh "Create REST API..." --format=json`
+Step 2 Init: `./init-config.sh --task-id cfn-phase-123 --task-description "..." --mode standard`
+Step 3 Decompose: `./decompose-task.sh --task-id ... --description "..." --complexity high`
+
+## Parameters
+- **classify**: TASK_DESCRIPTION, --format (json/simple)
+- **init**: --task-id, --task-description, --mode (mvp/standard/enterprise)
+- **decompose**: --task-id, --description, --tool-budget, --complexity
+
+## Expected output
+- **Classify**: `{task_type, complexity, keywords_matched, suggested_agents}`
+- **Init**: `.cfn/task-configs/task-{id}.json` with scope, agents, thresholds, acceptance criteria
+- **Decompose**: JSON array of subtasks with deliverables, tool_budget, estimated_effort
+
+## Real-world example
+"Add JWT auth" → classify backend → init config with deliverables `[src/auth/jwt.ts, tests/]` + acceptance criteria → spawn agents with known scope

package/.claude/skills/cfn-transparency-middleware/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: cfn-transparency-middleware
-description: Agent interaction capture, logging, and analysis with memory tracking and security (Rust implementation)
+description: "Agent interaction capture, logging, and analysis with memory tracking and security (Rust implementation). Use when you need to capture and analyze agent interactions, track tool usage and performance metrics, query execution history, or export audit trails for compliance."
 version: 1.0.0
 tags: [middleware, logging, security, transparency, memory, rust]
 status: production

package/.claude/skills/cfn-utilities/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: cfn-utilities
-description: Reusable bash utility functions for CFN Loop - logging, error handling, retry, file operations
+description: "Reusable bash utility functions for CFN Loop - logging, error handling, retry, file operations. Use when you need structured logging, atomic file operations, retry logic with exponential backoff, or standardized error handling in bash scripts."
 version: 1.0.0
 tags: [bash, logging, error-handling, retry, file-operations, utilities]
 status: production

package/.claude/skills/cfn-validation-framework/SKILL.md

@@ -1,60 +1,3 @@
----
-name: cfn-validation-framework
-description: Multi-layer validation for CFN Loop - templates, defense-in-depth, deliverables, instrumentation
-version: 1.0.0
-tags: [mega-skill, validation, quality, testing]
-status: production
----
-
-# Validation Framework Skill (Mega-Skill)
-
-**Version:** 1.0.0
-**Purpose:** Multi-layer validation for CFN Loop
-**Status:** Production
-**Consolidates:** cfn-validation-templates, cfn-defense-in-depth, cfn-deliverable-validation, cfn-validation-runner-instrumentation, json-validation
-
----
-
-## Overview
-
-This mega-skill provides complete validation capabilities:
-- **Templates** - Task-type specific validation templates
-- **Layers** - Defense-in-depth multi-layer validation
-- **Deliverables** - File and content validation
-- **Instrumentation** - Validation runner metrics
-- **JSON** - Schema validation and sanitization
-
----
-
-## Directory Structure
-
-```
-validation-framework/
-├── SKILL.md
-├── lib/
-│   ├── templates/        # From cfn-validation-templates
-│   ├── layers/           # From cfn-defense-in-depth
-│   ├── deliverables/     # From cfn-deliverable-validation
-│   ├── instrumentation/  # From cfn-validation-runner-instrumentation
-│   └── json/             # From json-validation
-└── cli/
-```
-
----
-
-## Migration Paths
-
-| Old Path | New Path |
-|----------|----------|
-| cfn-validation-templates/ | validation-framework/lib/templates/ |
-| cfn-defense-in-depth/ | validation-framework/lib/layers/ |
-| cfn-deliverable-validation/ | validation-framework/lib/deliverables/ |
-| cfn-validation-runner-instrumentation/ | validation-framework/lib/instrumentation/ |
-| json-validation/ | validation-framework/lib/json/ |
-
----
-
-## Version History
-
-### 1.0.0 (2025-12-02)
-- Consolidated 5 validation skills into mega-skill
+./execute.sh validate-json --file deliverables.json --schema schema.json
+./execute.sh calculate-deliverable-confidence --deliverables ./artifacts/
+./execute.sh run-instrumented --command "npm test" --output-dir ./test-results/

package/.claude/skills/cfn-vision-analysis/SKILL.md

@@ -1,20 +1,3 @@
----
-name: cfn-vision-analysis
-description: Image and vision analysis capabilities for screenshot and mockup validation
-version: 1.0.0
-tags: [vision, image, analysis, screenshots]
-status: production
----
-
-# CFN Vision Analysis Skill
-
-## Purpose
-
-Provides image and vision analysis capabilities for validating screenshots, mockups, and visual content in the CFN Loop workflow.
-
-## Use Cases
-
-- Screenshot validation against mockups
-- UI consistency checking
-- Visual regression detection
-- Brand guideline compliance
+do { 
+  implement → screenshot → analyze_image → video → analyze_video → score 
+} while (score < 85%)

package/.claude-flow/metrics/performance.json

@@ -1,7 +1,7 @@
 {
-  "startTime": 1765173092406,
-  "sessionId": "session-1765173092406",
-  "lastActivity": 1765173092406,
+  "startTime": 1765330300775,
+  "sessionId": "session-1765330300775",
+  "lastActivity": 1765330300775,
   "sessionDuration": 0,
   "totalTasks": 1,
   "successfulTasks": 1,

package/.claude-flow/metrics/task-metrics.json

@@ -1,10 +1,10 @@
 [
   {
-    "id": "cmd-hooks-1765173092515",
+    "id": "cmd-hooks-1765330300826",
     "type": "hooks",
     "success": true,
-    "duration": 37.66801499999997,
-    "timestamp": 1765173092553,
+    "duration": 25.54045400000001,
+    "timestamp": 1765330300852,
     "metadata": {}
   }
 ]

package/CLAUDE.md

@@ -46,9 +46,17 @@ Always provide `context_files` when code needs imports from existing files.
 
 ### Context Discovery Priority (fastest to slowest)
 1. **RuVector semantic search** (for "where is X?" queries):
+   - **Centralized index:** `~/.local/share/ruvector/index_v2.db` (shared across all projects)
+   - **Dual storage:** V1 (semantic/fuzzy) + V2 (structural/SQL)
+   - **V1 queries:** "Find similar code" via vector embeddings (cosine similarity)
+   - **V2 queries:** "Find callers/refs" via SQL on AST entities
    ```bash
+   # Semantic search
    /codebase-search "authentication middleware pattern"
-   ./.claude/skills/cfn-ruvector-codebase-index/search.sh "query" --top 5
+   ./.claude/skills/cfn-local-ruvector-accelerator/target/release/local-ruvector query --pattern "auth"
+
+   # Structural SQL query
+   sqlite3 ~/.local/share/ruvector/index_v2.db "SELECT * FROM refs WHERE target_name = 'MyFunction';"
    ```
 2. **Query past errors** before similar work:
    ```bash

package/README.md

@@ -219,6 +219,62 @@ bash tests/cli-mode/test-*.sh
 bash tests/docker/validation/validate-bug6-redis-vars.sh
 ```
 
+## RuVector - Semantic Codebase Search
+
+**Centralized Index:** `~/.local/share/ruvector/index_v2.db`
+
+RuVector provides dual-mode code intelligence with a centralized index shared across all projects:
+
+### Dual Storage Architecture
+
+| Schema | Purpose | Query Type | Use Case |
+|--------|---------|------------|----------|
+| **V1** (embeddings, files) | Semantic similarity | Vector distance | "Find code similar to X" |
+| **V2** (entities, refs, modules) | Structured relationships | SQL joins | "Who calls this function?" |
+
+**V1 - Semantic Search:**
+- Stores text chunks with OpenAI embeddings (text-embedding-3-small, 1536 dims)
+- Queries: Fuzzy semantic similarity via cosine distance
+- Returns: Code that's semantically related regardless of exact syntax
+
+**V2 - Code Intelligence:**
+- Stores AST entities (functions, classes, interfaces) with relationships
+- Queries: Precise SQL on structured code graph
+- Returns: Exact references, callers, type usage, module dependencies
+
+### Usage
+
+```bash
+# Build RuVector
+cd .claude/skills/cfn-local-ruvector-accelerator
+cargo build --release
+
+# Initialize centralized index
+./target/release/local-ruvector init
+
+# Index current project (all file types)
+./target/release/local-ruvector index --path . --types rs,ts,js,json,md,sh
+
+# Semantic search across all indexed projects
+./target/release/local-ruvector query --pattern "authentication middleware"
+
+# SQL queries on code structure
+sqlite3 ~/.local/share/ruvector/index_v2.db "
+  SELECT * FROM refs WHERE target_name = 'MyFunction';
+"
+```
+
+**Supported Files:**
+- AST extraction: Rust (.rs), TypeScript/JavaScript (.ts, .tsx, .js, .jsx)
+- Text indexing: JSON, YAML, Markdown (.md), Shell scripts (.sh)
+- Auto-excluded: `.artifacts/` (logs/reports)
+
+**Benefits:**
+- Cross-project semantic search
+- Shared learnings across codebases
+- Dual query modes (semantic + structural)
+- Full paths preserve project context
+
 ## Support
 
 - **Documentation Issues**: File issue at GitHub repo