claude-flow-novice 2.18.6 → 2.18.7

package/claude-assets/commands/cfn-loop/cfn-optimize-agents.md

@@ -1,103 +0,0 @@
----
-name: cfn-optimize-agents
-description: |
-  Optimize agent profiles for CFN loops, CLI commands, and swarm coordination using cli-agent-optimizer agents.
-  Launches multiple cli-agent-optimizer agents in parallel to review and optimize existing agent profiles.
-  Triggers CFN coordinators for autonomous optimization execution.
-
-parameters:
-  - name: scope
-    type: string
-    default: all
-    description: Optimization scope (all, core, specialized)
-    required: false
-  - name: agents
-    type: string
-    description: Specific agent types to optimize (comma-separated)
-    required: false
-  - name: parallel
-    type: number
-    default: 3
-    description: Number of parallel optimizer agents to run
-    required: false
-
-examples:
-  - /cfn-optimize-agents
-  - /cfn-optimize-agents --scope=all
-  - /cfn-optimize-agents --scope=core --parallel=3
-  - /cfn-optimize-agents --agents=coder,architect,tester --parallel=5
-  - /cfn-optimize-agents --scope=specialized --parallel=5
-
-category: cfn-loop
-subcategory: optimization
----
-
-# CFN Optimize Agents
-
-Optimize agent profiles for CFN loops, CLI commands, and swarm coordination using multiple cli-agent-optimizer agents.
-
-## Usage
-
-```bash
-/cfn-optimize-agents [options]
-```
-
-## Options
-
-- `--scope=<scope>` - Optimization scope (all, core, specialized) [default: all]
-- `--agents=<types>` - Specific agent types to optimize (comma-separated)
-- `--parallel=<count>` - Number of parallel optimizer agents [default: 3]
-
-## Examples
-
-Basic optimization:
-```bash
-/cfn-optimize-agents
-```
-
-Core agents only with 5 parallel optimizers:
-```bash
-/cfn-optimize-agents --scope=core --parallel=5
-```
-
-Specific agents with more parallel workers:
-```bash
-/cfn-optimize-agents --agents=coder,architect,tester --parallel=5
-```
-
-## What It Does
-
-1. **Launches** a coordinator agent via Task tool to manage the optimization process
-2. **Discovers** all agent profiles in .claude/agents/cfn-dev-team/ directory structure
-3. **Distributes** profiles evenly among parallel optimizers (zero overlap)
-4. **Coordinates** parallel work via Redis pub/sub messaging
-5. **Spawns** multiple cli-agent-optimizer agents via CLI with specific assignments
-6. **Monitors** progress and ensures complete coverage of all profiles
-7. **Aggregates** results and validates all optimizations are completed
-
-## Coordinator-Managed Integration
-
-This command triggers autonomous optimization execution with:
-- **Coordinator Agent**: Manages distribution, monitoring, and aggregation via Task tool
-- **Profile Distribution**: Intelligent assignment to prevent overlap (15-20 profiles per optimizer)
-- **Parallel Optimization**: 1-10 cli-agent-optimizer agents with non-overlapping work
-- **Redis Coordination**: Real-time progress tracking via pub/sub channels
-- **Quality Assurance**: Verification of complete coverage and confidence scores
-- **Overlap Prevention**: Ensures zero duplication across optimizer assignments
-
-## Output
-
-- Optimized agent profiles with improved structure
-- Enhanced CLI commands with better error handling
-- Improved Redis coordination patterns
-- Performance metrics and improvement measurements
-- Optimization recommendations and best practices
-
-## Coordinator Pattern Benefits
-
-- **Zero Overlap**: Intelligent distribution prevents duplicate work across optimizers
-- **Complete Coverage**: Coordinator ensures all discovered profiles are assigned and optimized
-- **Real-time Monitoring**: Redis pub/sub coordination provides live progress tracking
-- **Quality Assurance**: Confidence score monitoring and validation of all optimizations
-- **Scalable**: Supports 1-10 parallel optimizers with automatic workload balancing
-- **Fault Tolerant**: Coordinator handles failed assignments and redistributes work

package/claude-assets/commands/cfn-loop/list-agents-rebuild.md

@@ -1,205 +0,0 @@
----
-name: list-agents-rebuild
-description: "Regenerate available agents list from .claude/agents/ folder discovery"
-argument-hint: "[--output=path] [--format=markdown]"
-allowed-tools: ["Bash", "Read", "Write"]
----
-
-# Rebuild Available Agents List - Dynamic Discovery
-
-Regenerate `src/cli/hybrid-routing/AVAILABLE-AGENTS.md` from live agent discovery system.
-
-## Architecture Clarification
-
-**Source of Truth**: `.claude/agents/` folder (50+ agent .md files)
-**Live Discovery**: `HybridWorkerSpawner.loadAgentDefinitions()` method
-**Documentation Snapshot**: `src/cli/hybrid-routing/AVAILABLE-AGENTS.md` (generated file)
-
-**When coordinators spawn agents**:
-1. Load agents dynamically from `.claude/agents/` folder
-2. Parse YAML frontmatter from each .md file
-3. Extract keywords, system prompts, categories
-4. In-memory cache for performance
-5. Select agents via keyword matching or coordinator override
-
-**AVAILABLE-AGENTS.md purpose**:
-- Human-readable reference
-- Documentation for users/coordinators
-- Snapshot of available agents at generation time
-- **NOT** used by the spawning system (system reads .claude/agents/ directly)
-
----
-
-## Execution
-
-```bash
-# Regenerate agent list
-node src/cli/hybrid-routing/spawn-workers.cjs --agents-by-category > src/cli/hybrid-routing/AVAILABLE-AGENTS.md.tmp
-
-# Parse and format into markdown document
-node << 'EOF'
-const fs = require('fs');
-const path = require('path');
-
-// Read raw output
-const rawOutput = fs.readFileSync('src/cli/hybrid-routing/AVAILABLE-AGENTS.md.tmp', 'utf8');
-
-// Parse output (extract agent data)
-const lines = rawOutput.split('\n');
-let markdown = '';
-
-// Generate markdown header
-markdown += '# Available Specialized Agents\n\n';
-markdown += '**Hybrid Routing System - Dynamic Agent Discovery**\n\n';
-markdown += '**Generated**: ' + new Date().toISOString().split('T')[0] + '\n';
-markdown += '**Source**: `.claude/agents/` folder (live discovery)\n';
-markdown += '**Purpose**: Documentation snapshot - coordinators read from `.claude/agents/` directly\n\n';
-
-// Add architecture note
-markdown += '## Architecture\n\n';
-markdown += '**Source of Truth**: `.claude/agents/` folder\n';
-markdown += '- Coordinators use `HybridWorkerSpawner.loadAgentDefinitions()`\n';
-markdown += '- Recursive scanning with YAML frontmatter parsing\n';
-markdown += '- In-memory caching after first load\n';
-markdown += '- This file is documentation only (not used by spawning system)\n\n';
-
-// Parse discovery statistics
-const discoveredMatch = rawOutput.match(/Discovered (\d+) agent files/);
-const loadedMatch = rawOutput.match(/Loaded (\d+) agents/);
-const skippedMatch = rawOutput.match(/\((\d+) skipped\)/);
-const categoriesMatch = rawOutput.match(/(\d+) categories/);
-
-if (discoveredMatch && loadedMatch) {
-  markdown += '## Discovery Statistics\n\n';
-  markdown += '```\n';
-  markdown += `🔍 Discovered ${discoveredMatch[1]} agent files in .claude/agents/\n`;
-  markdown += `✅ Loaded ${loadedMatch[1]} agents (${skippedMatch ? skippedMatch[1] : '?'} skipped)\n`;
-  markdown += `📋 ${categoriesMatch ? categoriesMatch[1] : '?'} categories\n`;
-  markdown += '```\n\n';
-}
-
-// Parse agent categories and create tables
-markdown += '## Agents by Category\n\n';
-
-let currentCategory = null;
-let agents = [];
-
-for (let i = 0; i < lines.length; i++) {
-  const line = lines[i];
-
-  // Detect category header
-  const categoryMatch = line.match(/^📁 ([A-Z-]+) \((\d+) agents?\)/);
-  if (categoryMatch) {
-    // Write previous category
-    if (currentCategory && agents.length > 0) {
-      markdown += createCategorySection(currentCategory, agents);
-      agents = [];
-    }
-
-    currentCategory = {
-      name: categoryMatch[1],
-      count: categoryMatch[2]
-    };
-    continue;
-  }
-
-  // Detect agent entry
-  const agentMatch = line.match(/^\s+•\s+([a-z-]+)/);
-  if (agentMatch) {
-    const agentType = agentMatch[1];
-    const nextLine = lines[i + 1] || '';
-    const keywords = nextLine.trim();
-
-    agents.push({ type: agentType, keywords });
-    i++; // Skip next line (already processed)
-  }
-}
-
-// Write last category
-if (currentCategory && agents.length > 0) {
-  markdown += createCategorySection(currentCategory, agents);
-}
-
-// Add usage section
-markdown += '\n---\n\n';
-markdown += '## Usage\n\n';
-markdown += '### CLI Commands\n\n';
-markdown += '```bash\n';
-markdown += '# List all agents (flat view)\n';
-markdown += 'node src/cli/hybrid-routing/spawn-workers.cjs --list-agents\n\n';
-markdown += '# List agents by category\n';
-markdown += 'node src/cli/hybrid-routing/spawn-workers.cjs --agents-by-category\n\n';
-markdown += '# Regenerate this documentation file\n';
-markdown += '/list-agents-rebuild\n';
-markdown += '```\n\n';
-
-markdown += '### Coordinator Usage\n\n';
-markdown += '```bash\n';
-markdown += '# Automatic selection (keyword-based)\n';
-markdown += 'node src/cli/hybrid-routing/spawn-workers.cjs "Build auth" --max-agents=3\n\n';
-markdown += '# Coordinator override (manual agent types)\n';
-markdown += 'node src/cli/hybrid-routing/spawn-workers.cjs "Task" \\\n';
-markdown += '  --agents=architect,coder,tester\n\n';
-markdown += '# Full override (custom agents + subtasks)\n';
-markdown += 'node src/cli/hybrid-routing/spawn-workers.cjs "Task" \\\n';
-markdown += '  --agents=coder,security-specialist \\\n';
-markdown += '  --subtasks="Subtask 1|Subtask 2"\n';
-markdown += '```\n\n';
-
-// Add notes
-markdown += '---\n\n';
-markdown += '## Notes\n\n';
-markdown += '- **Live Discovery**: Coordinators read from `.claude/agents/` folder directly\n';
-markdown += '- **This File**: Documentation snapshot for human reference\n';
-markdown += '- **Regenerate**: Run `/list-agents-rebuild` to update this documentation\n';
-markdown += '- **Agent Files**: Add/modify agents in `.claude/agents/` folder (auto-discovered)\n';
-markdown += '- **Caching**: Agent definitions cached after first load (lazy loading)\n';
-markdown += '- **Missing Keywords**: Some agents without keywords can be used via coordinator override\n\n';
-
-// Write final markdown
-fs.writeFileSync('src/cli/hybrid-routing/AVAILABLE-AGENTS.md', markdown, 'utf8');
-
-console.log('✅ AVAILABLE-AGENTS.md regenerated successfully');
-
-// Helper function
-function createCategorySection(category, agents) {
-  let section = `### 📁 ${category.name} (${category.count} agents)\n\n`;
-  section += '| Agent Type | Keywords |\n';
-  section += '|------------|----------|\n';
-
-  for (const agent of agents) {
-    const keywords = agent.keywords || '*(No keywords available)*';
-    section += `| **${agent.type}** | ${keywords} |\n`;
-  }
-
-  section += '\n---\n\n';
-  return section;
-}
-EOF
-
-# Cleanup temp file
-rm src/cli/hybrid-routing/AVAILABLE-AGENTS.md.tmp
-
-echo ""
-echo "📋 Agent list regenerated: src/cli/hybrid-routing/AVAILABLE-AGENTS.md"
-echo ""
-echo "Architecture:"
-echo "  • Source: .claude/agents/ folder (50+ agent .md files)"
-echo "  • Coordinators: Load agents dynamically via HybridWorkerSpawner"
-echo "  • This file: Documentation snapshot for human reference"
-```
-
----
-
-## Output
-
-```
-✅ AVAILABLE-AGENTS.md regenerated successfully
-
-📋 Agent list regenerated: src/cli/hybrid-routing/AVAILABLE-AGENTS.md
-
-Architecture:
-  • Source: .claude/agents/ folder (50+ agent .md files)
-  • Coordinators: Load agents dynamically via HybridWorkerSpawner
-  • This file: Documentation snapshot for human reference
-```

package/claude-assets/commands/cfn-ruvector/codebase-reindex.md

@@ -1,37 +0,0 @@
----
-description: Update codebase index using git-detected changes (auto mode, incremental)
----
-
-# Codebase Reindex Command
-
-**Default mode: Auto-detection** - Indexes only changed files detected from git status (fast, incremental).
-
-**Use when:**
-- After committing code changes (recommended workflow)
-- Regular updates after development work
-- Need to refresh index for recently modified files
-
-**Process (--auto mode):**
-1. Detects staged and modified files via `git diff`
-2. Filters for indexable extensions (TypeScript, Python, etc.)
-3. Incrementally updates RuVector database
-4. Only re-indexes changed files
-
-**Estimated time:** 5-30 seconds (depends on number of changed files)
-
----
-
-**For full rebuild (only when needed):**
-Add `--full` flag if you need to rebuild the entire index from scratch:
-- First-time setup
-- Major codebase restructuring
-- Index corrupted or missing
-- After config changes (extensions, ignore patterns)
-
-```bash
-# Default: Auto-detect changed files (fast)
-./.claude/skills/cfn-ruvector-codebase-index/index.sh --auto
-
-# Full rebuild (only when explicitly needed)
-./.claude/skills/cfn-ruvector-codebase-index/index.sh --full
-```

package/claude-assets/commands/cfn-ruvector/codebase-search.md

@@ -1,38 +0,0 @@
----
-description: Search codebase using natural language queries via RuVector semantic index
-arguments:
-  query:
-    description: Natural language search query
-    required: true
-  --top:
-    description: Number of results to return (default 5)
-    required: false
----
-
-# Codebase Search Command
-
-Search your indexed codebase using natural language queries. Uses RuVector's semantic search to find relevant files based on meaning, not just keywords.
-
-**Examples:**
-- `/codebase-search authentication logic`
-- `/codebase-search React components for user profile --top 10`
-- `/codebase-search database migration utilities`
-- `/codebase-search error handling patterns`
-
-**Returns:**
-- File paths ranked by relevance
-- File purpose and exports
-- Code metrics (lines, complexity)
-- Relevance scores
-
-**Prerequisites:**
-- Codebase must be indexed first (run `/codebase-reindex`)
-- OPENAI_API_KEY or ZAI_API_KEY must be set
-
----
-
-Execute the search:
-
-```bash
-./.claude/skills/cfn-ruvector-codebase-index/search.sh "{{query}}" {{#if top}}--top {{top}}{{/if}}
-```

package/claude-assets/commands/cfn-ruvector/detect-stale-docs.md

@@ -1,41 +0,0 @@
----
-description: Detect stale and legacy documentation using RuVector semantic analysis
----
-
-# Detect Stale Documentation
-
-Analyzes all `.md` files in the codebase to detect legacy/outdated documentation by:
-
-- **Cross-referencing code**: Checks if files/functions mentioned in docs still exist
-- **Age analysis**: Older docs more likely to be stale
-- **Orphan detection**: Finds docs with no code references
-- **Deprecated keywords**: Flags docs mentioning "legacy", "deprecated", "obsolete"
-
-**Output:**
-- Staleness score (0-100+)
-- Age in days
-- Missing file/code references
-- Specific findings for each doc
-
-**Use cases:**
-- Identify documentation to archive
-- Find docs needing updates
-- Clean up legacy documentation
-- Maintain documentation hygiene
-
-**Scoring:**
-- Score >= 10: **STALE** (strong candidate for archival)
-- Score 5-9: **LIKELY STALE** (needs review)
-- Score 2-4: **POSSIBLY STALE** (minor issues)
-
-**Prerequisites:**
-- Codebase must be indexed first (`/codebase-reindex`)
-- ZAI_API_KEY must be set (for semantic search)
-
----
-
-Run the stale documentation detector:
-
-```bash
-./.claude/skills/cfn-ruvector-codebase-index/detect-stale-docs.sh
-```

package/claude-assets/commands/cost-routing/switch-api.md

@@ -1,109 +0,0 @@
----
-description: Switch Main Chat and Task tool between Z.ai, Kimi, OpenRouter, and Anthropic providers
-tags: [config, api, cost-optimization]
----
-
-Switch Main Chat and Task() tool API provider between multiple AI providers.
-
-**Important:** CLI agents support custom routing when enabled. This command affects Main Chat and Task() spawned agents.
-
-**Usage:**
-- `/switch-api` - Show current status
-- `/switch-api [zai|kimi|openrouter|max]` - Switch provider
-  - `zai` - Z.ai ($0.50/1M tokens)
-  - `kimi` - Moonshot Kimi ($2/1M tokens)
-  - `openrouter` - OpenRouter (varies by model)
-  - `max` or `claude` - Anthropic ($15/1M tokens, requires re-login)
-
-**Arguments:**
-- `status` - Show current routing configuration (default)
-- `zai` - Route Main Chat + Task tool to Z.ai for cost savings
-- `kimi` - Route Main Chat + Task tool to Moonshot Kimi
-- `openrouter` - Route Main Chat + Task tool to OpenRouter
-- `max` or `claude` - Route Main Chat + Task tool to Anthropic for quality
-
-**What This Does:**
-
-`/switch-api zai`:
-- Adds Z.ai env vars to `.claude/settings.local.json` (or settings.json)
-- Main Chat + Task() agents use Z.ai
-- Cost: $0.50/1M tokens (97% savings)
-- No login required
-
-`/switch-api kimi`:
-- Adds Kimi env vars to `.claude/settings.local.json` (or settings.json)
-- Main Chat + Task() agents use Moonshot Kimi
-- Cost: ~$2/1M tokens
-- Requires `KIMI_API_KEY` in root `.env`
-
-`/switch-api openrouter`:
-- Adds OpenRouter env vars to `.claude/settings.local.json` (or settings.json)
-- Main Chat + Task() agents use OpenRouter
-- Cost: Varies by model
-- Requires `OPENROUTER_API_KEY` in root `.env`
-
-`/switch-api max`:
-- Removes env vars from `.claude/settings.local.json` (or settings.json)
-- Main Chat + Task() agents use Anthropic
-- Cost: $15/1M tokens (or $0 with unlimited plan)
-- **Requires** running `claude login`
-
-**Note:** Script automatically detects and uses `.claude/settings.local.json` if it exists, otherwise uses `.claude/settings.json`.
-
-**Combined Architecture:**
-```
-Main Chat (Anthropic/Z.ai/Kimi/OpenRouter - your choice)
-  ↓
-Task() → Coordinator (uses Main Chat provider)
-  ↓
-CLI spawn → Workers (custom routing when enabled, see agent profiles)
-```
-
-**Execute:**
-```bash
-# Find script in project root or node_modules
-SCRIPT_PATH=""
-if [ -f "$PWD/scripts/switch-api.sh" ]; then
-    SCRIPT_PATH="$PWD/scripts/switch-api.sh"
-else
-    # Check parent directories (for nested project structures)
-    SEARCH_DIR="$PWD"
-    while [ "$SEARCH_DIR" != "/" ]; do
-        if [ -f "$SEARCH_DIR/scripts/switch-api.sh" ]; then
-            SCRIPT_PATH="$SEARCH_DIR/scripts/switch-api.sh"
-            break
-        fi
-        SEARCH_DIR="$(dirname "$SEARCH_DIR")"
-    done
-fi
-
-# Fallback to node_modules (for npm-installed package)
-if [ -z "$SCRIPT_PATH" ]; then
-    SEARCH_DIR="$PWD"
-    while [ "$SEARCH_DIR" != "/" ]; do
-        if [ -f "$SEARCH_DIR/node_modules/claude-flow-novice/scripts/switch-api.sh" ]; then
-            SCRIPT_PATH="$SEARCH_DIR/node_modules/claude-flow-novice/scripts/switch-api.sh"
-            break
-        fi
-        SEARCH_DIR="$(dirname "$SEARCH_DIR")"
-    done
-fi
-
-if [ -z "$SCRIPT_PATH" ]; then
-    echo "Error: switch-api.sh not found. Please ensure you're in a project with CFN installed."
-    exit 1
-fi
-
-bash "$SCRIPT_PATH" {{args}}
-```
-
-**Note:** Script automatically detects and uses `.claude/settings.local.json` if it exists, otherwise uses `.claude/settings.json`.
-
-**Examples:**
-```bash
-/switch-api             # Show current routing
-/switch-api zai         # Cost-optimize with Z.ai
-/switch-api kimi        # Use Moonshot Kimi
-/switch-api openrouter  # Use OpenRouter (access 400+ models)
-/switch-api max         # Quality-optimize with Anthropic (requires re-login)
-```

package/claude-assets/commands/deprecated/README.md

@@ -1,24 +0,0 @@
-# Deprecated Commands
-
-This folder contains commands that have been deprecated and should no longer be used.
-
-## Moved to Skills
-
-Several commands were better suited as skills and have been converted:
-- CFN_EXPERT_UPDATE → cfn-expert-update skill
-- parse-epic → cfn-epic-parser skill
-- github-commit → cfn-github-workflow skill
-- suggest-improvements → cfn-project-analysis skill
-- create-handoff → cfn-session-handoff skill
-
-## Non-Functional
-
-The following commands reference non-existent tools or files:
-- github.md - References 8 non-existent MCP tools
-- hello-world-tests.md - All tests moved to archive
-- run-tests.md - Missing cfn-test-runner skill
-
-## Superseded by Other Features
-
-- context-curate.md - Replaced by RuVector learning system
-- context-*.md - Replaced by RuVector semantic search and pattern management