🚀 Turbo-Flow Claude v1.0.5 Alpha
Advanced Agentic Development Environment
DevPods • GitHub Codespaces • Google Cloud Shell
What's New in v1.0.5
- ✅ Fixed Claude Flow Initialization - Now initializes in correct workspace directory (v9 script fixes)
- ✅ agtrace - AI agent observability with live dashboard and MCP integration
- ✅ Claudish - Multi-model proxy supporting 100+ models via OpenRouter
- ✅ OpenSpec - Spec-driven development from Fission AI
- ✅ Enhanced Playwright MCP - 143+ device emulation, cross-browser testing
- ✅ Chrome DevTools MCP - Full DevTools access with performance tracing
- ✅ Improved Reliability - Synchronous npm cache clean, preserved npx cache, absolute path checks
- ✅ Better Status Reporting - Shows actual initialization status with fix commands if needed
- ✅ Spec-Kit Integration - GitHub's spec-driven development workflow (
/speckit.*commands) - ✅ AI Agent Skills - 38+ installable skills for Claude Code via
ai-agent-skills - ✅ n8n-MCP Server - Build n8n workflows with AI assistance
- ✅ PAL MCP Server - Multi-model AI orchestration (Gemini + GPT + Grok + Ollama)
- ✅ Skills-Based Architecture - Claude Code now uses skills; wrapper scripts removed
- ✅ Dynamic CLAUDE.md - Generated from specs instead of pre-loaded
⚡ Quick Start
DevPod
# Install DevPod
# macOS: brew install loft-sh/devpod/devpod
# Windows: choco install devpod
# Linux: curl -L -o devpod "https://github.com/loft-sh/devpod/releases/latest/download/devpod-linux-amd64" && sudo install devpod /usr/local/bin
# Launch workspace
devpod up https://github.com/marcuspat/turbo-flow-claude --ide vscode
GitHub Codespaces
See github_codespaces_setup.md
Google Cloud Shell
See google_cloud_shell_setup.md
🛠️ What Gets Installed
Core Tools
| Tool | Alias | Description |
|------|-------|-------------|
| Claude Code | claude, dsp | Anthropic's AI coding CLI |
| Claude Flow | cf, cf-swarm, cf-hive | AI orchestration with SPARC methodology |
| Agentic Flow | af, af-run, af-coder | Self-learning agent workflows (66 agents, 213 MCP tools) |
| Agentic QE | aqe, aqe-mcp | AI-powered testing (19 agents, 11 TDD subagents) |
| Agentic Jujutsu | aj | AI-enhanced version control (23x faster than Git) |
| Claude Usage | cu, claude-usage | API usage tracking |
| Claudish | claudish | Multi-model proxy (100+ models via OpenRouter) |
New in v1.0.5
| Tool | Alias | Description |
|------|-------|-------------|
| Spec-Kit | sk, sk-here, sk-check | Spec-driven development from GitHub |
| OpenSpec | os, os-init, os-list | Spec-driven development from Fission AI |
| AI Agent Skills | skills-list, skills-install | 38+ skills for Claude/Cursor/Codex/Amp |
| n8n-MCP | n8n-mcp | n8n workflow automation (543 nodes) |
| PAL MCP | pal, pal-setup | Multi-model AI (Gemini, GPT, Grok, Ollama) |
| agtrace | agt-watch, agt-sessions | AI agent observability with live TUI dashboard |
MCP Servers (Auto-Configured)
| Server | Alias | Description |
|--------|-------|-------------|
| Playwright MCP | mcp-playwright | Browser automation (143+ devices, cross-browser) |
| Chrome DevTools MCP | mcp-chrome | Full DevTools access (24+ tools, performance tracing) |
| n8n-MCP | n8n-mcp | Workflow automation (543 nodes, 2,646 examples) |
| PAL MCP | pal | Multi-model AI orchestration |
| agtrace MCP | agt-mcp | Agent self-reflection and history queries |
Resources
- 610+ AI Agents - Specialized subagents for any task (422 non-coding, 188 coding)
- TypeScript Setup - Pre-configured with
tsconfig.json - Project Structure -
src/,tests/,docs/,scripts/,examples/,config/
🎯 Recommended Workflow
Spec-Driven Development
# 1. Initialize spec-kit in your project
sk-here # or: specify init . --ai claude
# 2. Generate CLAUDE.md (dynamic project context)
./devpods/generate-claude-md.sh # or: generate-claude-md
# 3. Start Claude Code
claude
# 4. Follow the spec-kit workflow
/speckit.constitution # Define project principles
/speckit.specify # Write specifications
/speckit.plan # Create implementation plan
/speckit.tasks # Break down into tasks
/speckit.implement # Build it
# 5. Regenerate CLAUDE.md after spec changes
generate-claude-md # Updates with latest specs
OpenSpec Workflow
# 1. Initialize OpenSpec
os-init
# 2. Create a change proposal
claude
> "/openspec:proposal Add OAuth authentication"
# 3. Review and iterate
os-show add-oauth
os-validate add-oauth
# 4. Apply when ready
> "/openspec:apply add-oauth"
# 5. Archive completed change
os-archive add-oauth
Multi-Model AI with PAL
# Setup PAL (first time only)
pal-setup # Installs dependencies
# Edit PAL config with your API keys
nano ~/.pal-mcp-server/.env
# Start PAL server
pal
# Use PAL for multi-model collaboration in Claude:
# "Use pal to analyze this with gemini pro and o3"
# "Get consensus from multiple models on this approach"
Multi-Model with Claudish
# Set API key
export OPENROUTER_API_KEY='sk-or-v1-...'
# Use different models
claudish --model x-ai/grok-code-fast-1 "fix the bug"
claudish --model google/gemini-2.5-flash "review this code"
claudish --model openai/gpt-5-codex "refactor the API"
# List available models
claudish --models
claudish --top-models
Agent Observability with agtrace
# Initialize in your project
agtrace init
# Launch live dashboard (in separate terminal)
agtrace watch
# Browse session history
agtrace session list
# Search across sessions
agtrace lab grep "error"
# Let agents query their own history (add MCP)
claude mcp add agtrace -- agtrace mcp serve
Agent Discovery
# Browse available skills
skills-list
# Install a skill
skills-install frontend-design
# Search for skills
skills-search "code review"
# Get skill details
skills-info frontend-design
# Install for different agents
skills-install frontend-design --agent cursor
skills-install frontend-design --agent vscode
📋 All Available Aliases
Claude Code
claude # Start Claude Code
claude-hierarchical # claude --dangerously-skip-permissions
dsp # claude --dangerously-skip-permissions (short)
Claude Flow
cf # Claude Flow base command
cf-init # Initialize claude-flow (--force)
cf-swarm # Run swarm mode
cf-hive # Spawn hive-mind agents
cf-spawn # Spawn hive-mind (alias)
cf-status # Check hive-mind status
cf-help # Show help
cf-fix # Fix better-sqlite3 dependency
cf-task "task" # Quick swarm task (function)
Agentic Tools
af # Agentic Flow
af-run # Run with agent
af-coder # Run coder agent
af-help # Agentic Flow help
af-task "agent" "task" # Quick agentic task (function)
aqe # Agentic QE testing
aqe-init # Initialize AQE
aqe-generate # Generate tests
aqe-flaky # Hunt flaky tests
aqe-gate # Quality gate check
aqe-mcp # Agentic QE MCP server
aj # Agentic Jujutsu (git)
aj-status # Jujutsu status
aj-analyze # Analyze repository
cu # Claude usage stats
claude-usage # Claude usage stats (full)
Spec-Kit
sk # Specify CLI
sk-init # Initialize new project
sk-check # Check installed tools
sk-here # Init in current directory with Claude
OpenSpec
os # OpenSpec CLI
os-init # Initialize OpenSpec
os-list # List change proposals
os-show # Show specific change
os-validate # Validate a change
os-archive # Archive completed change
os-update # Update configurations
AI Agent Skills
skills # Base command
skills-list # List all 38+ skills
skills-search "query" # Search skills
skills-install <name> # Install a skill
skills-info <name> # Get skill details
agtrace (Observability)
agt # agtrace base command
agt-init # Initialize workspace
agt-watch # Live TUI dashboard
agt-sessions # List sessions
agt-grep # Search across sessions
agt-mcp # Start MCP server
MCP Servers
n8n-mcp # n8n workflow MCP
pal # Start PAL multi-model MCP server
pal-setup # Setup PAL dependencies (uv sync)
mcp-playwright # Playwright MCP
mcp-chrome # Chrome DevTools MCP
Claudish (Multi-Model)
claudish # Interactive mode
claudish --models # List available models
claudish --top-models # Show recommended models
claudish-grok # Use Gr