tasq/CLAUDE.md

9.2 KiB

Claude Code Configuration - RuFlo V3

Behavioral Rules (Always Enforced)

  • Do what has been asked; nothing more, nothing less
  • NEVER create files unless they're absolutely necessary for achieving your goal
  • ALWAYS prefer editing an existing file to creating a new one
  • NEVER proactively create documentation files (*.md) or README files unless explicitly requested
  • NEVER save working files, text/mds, or tests to the root folder
  • Never continuously check status after spawning a swarm — wait for results
  • ALWAYS read a file before editing it
  • NEVER commit secrets, credentials, or .env files

File Organization

  • NEVER save to root folder — use the directories below
  • Use /src for source code files
  • Use /tests for test files
  • Use /docs for documentation and markdown files
  • Use /config for configuration files
  • Use /scripts for utility scripts
  • Use /examples for example code

Project Architecture

  • Follow Domain-Driven Design with bounded contexts
  • Keep files under 500 lines
  • Use typed interfaces for all public APIs
  • Prefer TDD London School (mock-first) for new code
  • Use event sourcing for state changes
  • Ensure input validation at system boundaries

Project Config

  • Topology: hierarchical-mesh
  • Max Agents: 15
  • Memory: hybrid
  • HNSW: Enabled
  • Neural: Enabled

Build & Test

# Build
npm run build

# Test
npm test

# Lint
npm run lint
  • ALWAYS run tests after making code changes
  • ALWAYS verify build succeeds before committing

Security Rules

  • NEVER hardcode API keys, secrets, or credentials in source files
  • NEVER commit .env files or any file containing secrets
  • Always validate user input at system boundaries
  • Always sanitize file paths to prevent directory traversal
  • Run npx @claude-flow/cli@latest security scan after security-related changes
  • All operations MUST be concurrent/parallel in a single message
  • Use Claude Code's Agent tool for spawning agents, not just MCP
  • ALWAYS spawn ALL agents in ONE message with full instructions via Agent tool
  • ALWAYS batch ALL file reads/writes/edits in ONE message
  • ALWAYS batch ALL Bash commands in ONE message

Swarm Orchestration

  • MUST initialize the swarm using CLI tools when starting complex tasks
  • MUST spawn concurrent agents using Claude Code's Agent tool
  • Never use CLI tools alone for execution — Agent tool agents do the actual work
  • MUST call CLI tools AND Agent tool in ONE message for complex work

3-Tier Model Routing (ADR-026)

Tier Handler Latency Cost Use Cases
1 Agent Booster (WASM) <1ms $0 Simple transforms (var→const, add types) — Skip LLM
2 Haiku ~500ms $0.0002 Simple tasks, low complexity (<30%)
3 Sonnet/Opus 2-5s $0.003-0.015 Complex reasoning, architecture, security (>30%)
  • For Tier 1 simple transforms, use Edit tool directly — no LLM agent needed

Swarm Configuration & Anti-Drift

  • ALWAYS use hierarchical topology for coding swarms
  • Keep maxAgents at 6-8 for tight coordination
  • Use specialized strategy for clear role boundaries
  • Use raft consensus for hive-mind (leader maintains authoritative state)
  • Run frequent checkpoints via post-task hooks
  • Keep shared memory namespace for all agents
npx @claude-flow/cli@latest swarm init --topology hierarchical --max-agents 8 --strategy specialized

Swarm Execution Rules

  • ALWAYS use run_in_background: true for all Agent tool calls
  • ALWAYS put ALL Agent calls in ONE message for parallel execution
  • After spawning, STOP — do NOT add more tool calls or check status
  • Never poll agent status repeatedly — trust agents to return
  • When agent results arrive, review ALL results before proceeding

V3 CLI Commands

Core Commands

Command Subcommands Description
init 4 Project initialization
agent 8 Agent lifecycle management
swarm 6 Multi-agent swarm coordination
memory 11 AgentDB memory with HNSW search
task 6 Task creation and lifecycle
session 7 Session state management
hooks 17 Self-learning hooks + 12 workers
hive-mind 6 Byzantine fault-tolerant consensus

Quick CLI Examples

npx @claude-flow/cli@latest init --wizard
npx @claude-flow/cli@latest agent spawn -t coder --name my-coder
npx @claude-flow/cli@latest swarm init --v3-mode
npx @claude-flow/cli@latest memory search --query "authentication patterns"
npx @claude-flow/cli@latest doctor --fix

Available Agents (16 Roles + Custom)

Core Development

coder, reviewer, tester, planner, researcher

Specialized

security-architect, security-auditor, memory-specialist, performance-engineer

Coordination

hierarchical-coordinator, mesh-coordinator, adaptive-coordinator

GitHub & Repository

pr-manager, code-review-swarm, issue-tracker, release-manager

Any string can be used as a custom agent type — these are the typed roles with specialized behavior.

MCP Tools (use via ToolSearch to discover)

Tool Description
memory_store Store value with ONNX 384-dim vector embedding
memory_search Semantic vector search by query
memory_retrieve Get entry by key
memory_list List entries in namespace
memory_delete Delete entry
memory_import_claude Import Claude Code memories into AgentDB (allProjects=true for all)
memory_search_unified Search across ALL namespaces (Claude + AgentDB + patterns)
memory_bridge_status Show bridge health, vectors, SONA, intelligence

CLI Commands

# Store with vector embedding
npx @claude-flow/cli@latest memory store --key "pattern-auth" --value "JWT with refresh" --namespace patterns

# Semantic search
npx @claude-flow/cli@latest memory search --query "authentication patterns"

# Import all Claude Code memories into AgentDB
node .claude/helpers/auto-memory-hook.mjs import-all

Claude Code ↔ AgentDB Bridge

Claude Code auto-memory files (~/.claude/projects/*/memory/*.md) are automatically imported into AgentDB with ONNX vector embeddings on session start. Use memory_search_unified to search across both stores.

Key MCP Tools (314 available — use ToolSearch to discover)

Most Used Tools

Category Tools What They Do
Memory memory_store, memory_search, memory_search_unified Store/search with ONNX vector embeddings
Claude Bridge memory_import_claude, memory_bridge_status Import Claude memories into AgentDB
Swarm swarm_init, swarm_status, swarm_health Multi-agent coordination
Agents agent_spawn, agent_list, agent_status Agent lifecycle
Hive-Mind hive-mind_init, hive-mind_spawn, hive-mind_consensus Byzantine/Raft consensus
Hooks hooks_route, hooks_session-start, hooks_post-task Task routing + learning
Workers hooks_worker-list, hooks_worker-dispatch 12 background workers
Security aidefence_scan, aidefence_is_safe Prompt injection detection
Intelligence hooks_intelligence, neural_status Pattern learning + SONA

Swarm Capabilities

  • Topologies: hierarchical (anti-drift), mesh, ring, star, adaptive
  • Consensus: Raft (leader-based), Byzantine (PBFT), Gossip (eventual)
  • Hive-Mind: Queen-led coordination with spawn, broadcast, consensus voting, shared memory
  • 12 Background Workers: audit, optimize, testgaps, map, deepdive, document, refactor, benchmark, ultralearn, consolidate, predict, preload

Memory Capabilities

  • ONNX Embeddings: all-MiniLM-L6-v2, 384 dimensions — real neural vectors
  • DiskANN: SSD-friendly vector search (8,000x faster insert than HNSW, perfect recall at 1K)
  • sql.js: Cross-platform SQLite (WASM, no native compilation)
  • Claude Code Bridge: Auto-imports MEMORY.md files into AgentDB on session start
  • Unified Search: memory_search_unified searches Claude memories + AgentDB + patterns
  • SONA Learning: Trajectory recording → pattern extraction → file persistence

How to Discover Tools

Use ToolSearch to find specific tools:

ToolSearch("memory search")     → memory_store, memory_search, memory_search_unified
ToolSearch("swarm")             → swarm_init, swarm_status, swarm_health, swarm_shutdown
ToolSearch("hive consensus")    → hive-mind_consensus, hive-mind_status
ToolSearch("+aidefence")        → aidefence_scan, aidefence_is_safe, aidefence_has_pii

Quick Setup

claude mcp add claude-flow -- npx -y @claude-flow/cli@latest
npx @claude-flow/cli@latest daemon start
npx @claude-flow/cli@latest doctor --fix

Claude Code vs MCP Tools

  • Claude Code Agent tool handles execution: agents, file ops, code generation, git
  • MCP tools (via ToolSearch) handle coordination: swarm, memory, hooks, routing, hive-mind
  • CLI commands (via Bash) are the same tools with terminal output
  • Use ToolSearch("keyword") to discover available MCP tools

Support