Claude Code Toolkit

Professional Claude Code setup with voice feedback, security protection, and MCP integrations. Get a complete development environment in 30 seconds.

🚀 Quick Start

One-line installation:

curl https://raw.githubusercontent.com/Mandalorian007/claude-code-toolkit/main/install.sh | bash

What this does:

• ✅ Installs hooks, agents, and MCP configurations
• ✅ Sets up .env with TODO comments for your API keys
• ✅ Updates .gitignore with essential entries
• ✅ Checks system dependencies (advisory only)
• ✅ Creates backup of any existing files

After installation:

# 1. Configure your API keys in .env (see TODO comments)
# 2. Launch Claude Code with environment loaded
export $(cat .env | xargs) && claude

📋 Prerequisites

The installer will check for these and guide you if missing:

• Claude Code - Anthropic's CLI (required)
• FFmpeg - Required for TTS audio playback
• Node.js - Required for MCP servers

⚡ Features

• 🎙️ Voice Notifications - ElevenLabs TTS for agent feedback
• 🛡️ Security Protection - Blocks dangerous commands (rm -rf, etc.)
• 🤖 Sub-Agents - Meta-agent creates specialized task agents
• 🔗 MCP Integration - Perplexity, Firecrawl, YouTube, Reddit, Playwright tools

📁 What Gets Installed

your-project/
├── .claude/
│   ├── settings.json              # Hook configurations & MCP permissions
│   ├── agents/                    # Sub-agent configurations
│   │   ├── meta-agent.md          # Agent that creates other agents
│   │   └── research-agent.md      # Multi-source research specialist
│   ├── commands/
│   │   ├── prime.md               # Project context loader
│   │   └── all_tools.md           # Tool discovery
│   └── hooks/
│       ├── user_prompt_submit.py  # Security hooks for prompts
│       ├── pre_tool_use.py        # Security gate for dangerous commands
│       ├── post_tool_use.py       # Post-tool processing hook
│       ├── notification.py        # Voice notifications with ElevenLabs TTS
│       ├── stop.py                # Completion messages with ElevenLabs TTS
│       └── utils/tts/elevenlabs_tts.py  # ElevenLabs voice synthesis
├── .mcp.json                      # MCP server configurations
├── README.md                      # This file
└── .env                          # Your environment variables (not in git)

🎯 Usage

• Load project context: /prime
• List available tools: /all_tools
• Create new agents: Ask the meta-agent to create specialized sub-agents
• Research tasks: Automatically delegated to research-agent

All hooks provide security protection and voice feedback automatically.

🔗 MCP Servers

• Perplexity - Real-time web search and AI Q&A
• Firecrawl - Advanced web scraping and content extraction
• YouTube - Video metadata, transcripts, and analysis
• Reddit - Community discussions and content access
• Playwright - Browser automation and testing

🔧 Manual Setup (Alternative)

If you prefer manual installation or want to understand the components:

1. Clone or download the toolkit files to your project
2. Configure API keys in .env:

   ENGINEER_NAME=your-name
   ELEVENLABS_API_KEY=your-elevenlabs-key
   ELEVENLABS_VOICE_ID=aUNOP2y8xEvi4nZebjIw
   PERPLEXITY_API_KEY=your-perplexity-key
   FIRECRAWL_API_KEY=your-firecrawl-key
   YOUTUBE_API_KEY=your-youtube-key
   YOUTUBE_TRANSCRIPT_LANG=en

3. Install dependencies:

   brew install ffmpeg node  # macOS
   # or your system's package manager

4. Launch: export $(cat .env | xargs) && claude

🛡️ Safety Features

• Automatic backups - Existing files saved to timestamped directory
• Non-destructive - Only adds missing entries to .gitignore
• No sudo required - Safe installation without elevated permissions
• Advisory dependency checks - Guides you to install missing tools

📚 Credits

Base agent configurations adapted from Contains Studio.