π The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
TypeScript agentic-aiagentic-frameworkagentic-workflowagentsai-agentsai-assistantai-skillsautonomous-agentsclaudecodexdsh-pluginharnessmcp-servermulti-agentmulti-agent-systemsnpmskillsswarmswarm-intelligencetypescript
Checkpoint: File edits
rUvdb55dc9f9f3b
diff --git a/v3/@claude-flow/cli/README.md b/v3/@claude-flow/cli/README.mdindex 06245ec673..9e31246506 100644--- a/v3/@claude-flow/cli/README.md+++ b/v3/@claude-flow/cli/README.md@@ -1049,6 +1049,247 @@ All checks passed with some warnings.---+### `route` - Q-Learning Agent Router++Intelligent task-to-agent routing using reinforcement learning.++```bash+claude-flow route <subcommand> [options]+```++#### Subcommands++| Subcommand | Description |+|------------|-------------|+| `task` | Route a task to optimal agent using Q-Learning |+| `list-agents` | List available agent types with capabilities |+| `stats` | Show router statistics and learning metrics |+| `feedback` | Provide routing feedback for continuous learning |+| `reset` | Reset Q-Learning router state |+| `export` | Export Q-table for persistence |+| `import` | Import Q-table from file |++#### Agent Types++| Type | Description |+|------|-------------|+| `coder` | Code implementation and debugging |+| `tester` | Testing and quality assurance |+| `reviewer` | Code review and quality checks |+| `architect` | System design and architecture |+| `researcher` | Research and information gathering |+| `optimizer` | Performance optimization |+| `debugger` | Bug fixing and troubleshooting |+| `documenter` | Documentation and comments |++#### How It Works++1. **State Encoding** - Analyzes task description using hash-based feature extraction+2. **Q-Learning** - Uses Q-table to map (state, action) pairs to expected rewards+3. **Epsilon-Greedy** - Balances exploration (try new agents) vs exploitation (use best known)+4. **Feedback Loop** - Learns from success/failure outcomes to improve routing++#### Examples++```bash+# Route a task to optimal agent+claude-flow route task "Implement user authentication with JWT"++# Route with verbose output showing decision process+claude-flow route task "Fix memory leak in API handler" --verbose++# List available agent types+claude-flow route list-agents++# View router statistics+claude-flow route stats++# Provide feedback (reward: -1 to 1)+claude-flow route feedback --agent coder --reward 0.8 --task "auth implementation"++# Export Q-table for backup+claude-flow route export -o ./q-table.json++# Import Q-table+claude-flow route import -i ./q-table.json++# Reset router (start fresh)+claude-flow route reset --confirm+```++---++### `analyze` - Code Analysis++Comprehensive code analysis with AST parsing, diff classification, complexity metrics, and dependency analysis.++```bash+claude-flow analyze <subcommand> [options]+```++#### Subcommands++| Subcommand | Description | Algorithm |+|------------|-------------|-----------|+| `ast` | AST analysis with symbol extraction | tree-sitter (regex fallback) |+| `complexity` | Cyclomatic and cognitive complexity | McCabe + cognitive metrics |+| `symbols` | Extract functions, classes, types | AST parsing |+| `imports` | Import dependency analysis | Static analysis |+| `diff` | Diff classification and risk assessment | Pattern matching |+| `boundaries` | Code boundaries detection | MinCut algorithm |+| `modules` | Module community detection | Louvain algorithm |+| `dependencies` | Full dependency graph | Graph building |+| `circular` | Circular dependency detection | Tarjan's SCC |+| `deps` | Project dependency analysis | npm/yarn |+| `code` | Static code quality analysis | Multi-metric |++#### Diff Risk Levels++| Risk | Color | Description |+|------|-------|-------------|+| `critical` | Red | Breaking changes, security-sensitive |+| `high` | Orange | Core logic, API changes |+| `medium` | Yellow | Feature additions, significant refactoring |+| `low` | Green | Documentation, tests, formatting |++#### Examples++```bash+# Analyze AST of a file+claude-flow analyze ast ./src/auth/login.ts++# Analyze with symbol extraction+claude-flow analyze symbols ./src/services/ --recursive++# Check cyclomatic complexity+claude-flow analyze complexity ./src --threshold 10++# Classify a git diff+claude-flow analyze diff HEAD~1++# Classify staged changes+claude-flow analyze diff --staged++# Find code boundaries (module splits)+claude-flow analyze boundaries ./src --min-cut++# Detect module communities+claude-flow analyze modules ./src --algorithm louvain++# Find circular dependencies+claude-flow analyze circular ./src++# Full dependency graph+claude-flow analyze dependencies ./src -o ./deps.dot --format dot++# Project dependency analysis+claude-flow analyze deps --outdated+```++---++### `issues` - Collaborative Issue Claims (ADR-016)++Human-agent collaborative issue management with work stealing, load balancing, and handoffs.++```bash+claude-flow issues <subcommand> [options]+```++#### Subcommands++| Subcommand | Description |+|------------|-------------|+| `list` | List all active claims |+| `claim` | Claim an issue for yourself or an agent |+| `release` | Release a claim |+| `handoff` | Request handoff to another agent/user |+| `status` | Update claim status and progress |+| `stealable` | List issues available for work stealing |+| `steal` | Steal an issue from overloaded/stale claimant |+| `load` | View agent load distribution |+| `rebalance` | Rebalance claims across swarm |+| `board` | Visual Kanban-style claim board |++#### Claim Status++| Status | Description |+|--------|-------------|+| `active` | Actively being worked on |+| `paused` | Temporarily paused |+| `blocked` | Blocked by dependency |+| `stealable` | Available for work stealing |+| `completed` | Work completed |+| `handoff-pending` | Awaiting handoff acceptance |+| `review-requested` | Needs review |++#### Work Stealing++Issues can be marked stealable when:+- **Overloaded** - Claimant has too many active claims+- **Stale** - No progress for extended period+- **Blocked** - Blocked beyond timeout threshold+- **Voluntary** - Claimant voluntarily releases++#### Examples++```bash+# List all claims+claude-flow issues list++# List with status filter+claude-flow issues list --status active++# Claim an issue+claude-flow issues claim ISSUE-123++# Claim as agent+claude-flow issues claim ISSUE-123 --as-agent --type coder++# Update progress+claude-flow issues status ISSUE-123 --progress 50++# Mark as blocked+claude-flow issues status ISSUE-123 --set blocked --note "Waiting for API"++# Request handoff to another user+claude-flow issues handoff ISSUE-123 --to user:alice:Alice++# Request handoff to agent+claude-flow issues handoff ISSUE-123 --to agent:coder:coder-001++# View stealable issues+claude-flow issues stealable++# Steal an issue+claude-flow issues steal ISSUE-123 --reason overloaded++# View agent load+claude-flow issues load++# Rebalance swarm+claude-flow issues rebalance --strategy even++# Visual board view+claude-flow issues board+```++#### Board View++```+π Issue Claims Board+ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ++ACTIVE (3) PAUSED (1) BLOCKED (1)+ββββββββββββββββ ββββββββββββββββ ββββββββββββββββ+β ISSUE-123 β β ISSUE-456 β β ISSUE-789 β+β π€ coder β β π€ alice β β π€ tester β+β 75% βββββββ β β 30% βββββββ β β 50% βββββββ β+ββββββββββββββββ ββββββββββββββββ ββββββββββββββββ+```++---+### `completions` - Shell CompletionsGenerate shell completion scripts for tab completion of commands.