🚀 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 Grok model
claudish-gemini # Use Gemini model
claudish-gpt # Use GPT model
Tmux
t # tmux
tn / tns # new / new-session -s
ta / tat # attach / attach -t
tl / tls # list sessions
tks # kill-session -t
tsh / tsv # split horizontal / vertical
Helper Functions
cf-task "task" # Quick swarm task
af-task "agent" "task" # Quick agentic task with streaming
generate-claude-md # Generate CLAUDE.md from project analysis
turbo-init # Initialize all tools in new project
turbo-help # Show quick reference
📁 Project Structure
/workspaces/turbo-flow-claude/
├── agents/ # 610+ AI agent definitions
├── src/ # Source code
├── tests/ # Test files
├── docs/ # Documentation
├── scripts/ # Utility scripts
├── examples/ # Example code
├── config/ # Configuration files
├── devpods/ # DevPod setup scripts
│ ├── setup.sh # Main setup script (v9)
│ └── generate-claude-md.sh # CLAUDE.md generator script
├── openspec/ # OpenSpec specs (after os-init)
│ ├── specs/ # Current truth
│ └── changes/ # Proposed changes
├── .specify/ # Spec-kit specs (after sk-here)
├── .claude-flow/ # Claude Flow config (after cf-init)
├── package.json # Node.js config (ES modules)
├── tsconfig.json # TypeScript config
└── CLAUDE.md # Generated project context for Claude
🔧 Configuration
PAL MCP (Multi-Model AI)
Edit ~/.pal-mcp-server/.env:
GEMINI_API_KEY=your-key # Google Gemini
OPENAI_API_KEY=your-key # GPT-4, GPT-5, O3
OPENROUTER_API_KEY=your-key # 100+ models
XAI_API_KEY=your-key # Grok
ANTHROPIC_API_KEY=your-key # Claude (for PAL orchestration)
Claudish
export OPENROUTER_API_KEY='sk-or-v1-...'
export CLAUDISH_MODEL='x-ai/grok-code-fast-1' # Optional default
n8n-MCP
N8N_API_URL=https://your-n8n.com
N8N_API_KEY=your-api-key
MCP Servers
Auto-configured at ~/.config/claude/mcp.json:
{
"mcpServers": {
"playwright": {
"command": "npx",
"args": ["-y", "@playwright/mcp@latest"]
},
"chrome-devtools": {
"command": "npx",
"args": ["chrome-devtools-mcp@latest"]
},
"n8n-mcp": {
"command": "npx",
"args": ["-y", "n8n-mcp"]
},
"agtrace": {
"command": "agtrace",
"args": ["mcp", "serve"]
}
}
}
🎛️ DevPod Management
# Create workspace
devpod up https://github.com/marcuspat/turbo-flow-claude --ide vscode
# Stop (saves costs)
devpod stop turbo-flow-claude
# Resume
devpod up turbo-flow-claude --ide vscode
# Delete
devpod delete turbo-flow-claude --force
# List workspaces
devpod list
🌍 Cloud Providers
DigitalOcean (Recommended)
devpod provider add digitalocean
devpod provider use digitalocean
devpod provider update digitalocean --option DIGITALOCEAN_ACCESS_TOKEN=your_token
devpod provider update digitalocean --option DROPLET_SIZE=s-4vcpu-8gb
AWS
devpod provider add aws
devpod provider use aws
devpod provider update aws --option AWS_INSTANCE_TYPE=t3.medium
Other Providers
See devpod_provider_setup_guide.md for Azure, GCP, Rackspace, and local Docker setup.
🔧 Troubleshooting
Verify Installation
# Check all tools
claude --version
specify check
skills-list
echo "Agents: $(ls -1 agents/*.md 2>/dev/null | wc -l)"
# Check Claude Flow initialization
ls -la .claude-flow/ 2>/dev/null || echo "Claude Flow not initialized"
Claude Flow Not Initialized
If the setup summary shows ❌ not initialized for Claude Flow:
source ~/.bashrc
cf-fix
npx -y claude-flow@alpha init --force
Node.js Version < 20
curl -fsSL https://deb.nodesource.com/setup_20.x | sudo bash -
sudo apt-get install -y nodejs
Spec-Kit Not Found
uv tool install specify-cli --from git+https://github.com/github/spec-kit.git
Permission Issues
sudo chown -R $(whoami):staff ~/.devpod
Reload Aliases
source ~/.bashrc
npm Lock Issues
rm -rf ~/.npm/_locks
MCP Server Issues
# List registered servers
claude mcp list
# Remove and re-add
claude mcp remove n8n-mcp
claude mcp add n8n-mcp npx -y n8n-mcp
📚 Resources
- Turbo Flow Claude - This repository
- Claude Flow - AI orchestration
- 610 Claude Subagents - Agent library
- AI Agent Skills - Universal skill repository
- Spec-Kit - Spec-driven development
- OpenSpec - Spec-driven development
- PAL MCP Server - Multi-model AI
- agtrace - Agent observability
- Claudish - Multi-model proxy
- n8n-MCP - n8n workflow automation
- DevPod Documentation - Dev environments as code
📦 Installation Summary
The setup script installs:
| Category | Count |
|---|---|
| npm global packages | 12 |
| npm local dev packages | 2 |
| Python tools | 2 |
| Shell tools | 1 |
| Git-cloned repos | 2 |
| AI skills | 3 |
| Config files created | 4 |
| Directories created | 9 |
| Bash aliases | 50+ |
| Helper functions | 5 |
| MCP registrations | 5 |
Complete Package List
npm global:
- @anthropic-ai/claude-code
- claude-usage-cli
- agentic-qe
- agentic-flow
- agentic-jujutsu
- claudish
- @fission-ai/openspec
- ai-agent-skills
- n8n-mcp
- @playwright/mcp
- chrome-devtools-mcp
- @lanegrid/agtrace
Python (via uv):
- uv (package manager)
- specify-cli (spec-kit)
Shell:
- direnv
Git cloned:
- pal-mcp-server (~/.pal-mcp-server)
- 610ClaudeSubagents (agents/)
Skills installed:
- frontend-design
- mcp-builder
- code-review
⭐ Star History
Ready to start?
devpod up https://github.com/marcuspat/turbo-flow-claude --ide vscode