oms0401/repopack-MCP-git-flatner

🚀 Repopack - GitHub Repository Flattener for Claude

Pack entire GitHub repositories into Claude-ready context in seconds.

RepoStack is a powerful Model Context Protocol (MCP) server that transforms any GitHub repository into a single, beautifully formatted file that Claude can analyze instantly. No more manual copying, no more missing files—just pure codebase intelligence.

🎯 What It Does

RepoStack empowers Claude to:

• 📦 Instant Access - Fetch any GitHub repository with a single command
• 🔍 Smart Extraction - Automatically identifies and extracts relevant code files
• 📄 Perfect Formatting - Concatenates everything into a clean, navigable document
• 🤖 Full Context - Gives Claude the complete picture of any codebase

Perfect for code reviews, security audits, documentation generation, learning new frameworks, architecture analysis, and technical interviews.

✨ Features

• ✅ Public & Private Repos - Works with any GitHub repository
• ✅ Intelligent Filtering - Excludes dependencies, build files, and noise
• ✅ 40+ Languages - Python, JavaScript, TypeScript, Go, Rust, and more
• ✅ Configurable Limits - Control how many files to include
• ✅ Auto Branch Detection - Finds main/master automatically
• ✅ Beautiful Output - Clean separators and file headers
• ✅ Lightning Fast - Async operations for maximum speed

🎬 Demo

You: "RepoStack, analyze https://github.com/facebook/react"

Claude: *fetches entire React codebase*
        "React is a declarative JavaScript library for building user interfaces..."
        *provides deep insights from the actual source code*

📋 Prerequisites

• Python 3.8+ installed on your system
• Claude Desktop application
• GitHub Account (for private repos)

🚀 Quick Start

Step 1: Installation

# Clone RepoStack
git clone <your-repo-url>
cd repostack

# Install dependencies
pip install fastmcp httpx

💡 Pro Tip: Use a virtual environment:

python -m venv venv
venv\Scripts\activate  # Windows
source venv/bin/activate  # macOS/Linux
pip install fastmcp httpx

Step 2: Configure Claude Desktop

Windows 🪟

1. Press Win + R and type: %APPDATA%\Claude
2. Create or edit claude_desktop_config.json
3. Add this configuration:

{
  "mcpServers": {
    "repostack": {
      "command": "python",
      "args": ["D:\\path\\to\\repostack\\github_flattener.py"]
    }
  }
}

macOS 🍎

code ~/Library/Application\ Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "repostack": {
      "command": "python3",
      "args": ["/path/to/repostack/github_flattener.py"]
    }
  }
}

Linux 🐧

nano ~/.config/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "repostack": {
      "command": "python3",
      "args": ["/path/to/repostack/github_flattener.py"]
    }
  }
}

Step 3: Restart Claude

Important: Completely quit Claude Desktop (not just minimize) and restart it.

💬 Usage

Simply chat with Claude naturally:

🎯 Basic Commands

"Flatten https://github.com/openai/whisper"

"Use RepoStack on github.com/vercel/next.js"

"Analyze the codebase at https://github.com/microsoft/typescript"

🎨 Creative Prompts

Code Review:

"RepoStack https://github.com/user/project and review the code quality"

Security Audit:

"Find security vulnerabilities in https://github.com/user/webapp"

Architecture Analysis:

"Explain the architecture of https://github.com/nestjs/nest"

Learning:

"How does https://github.com/django/django handle authentication?"

Documentation:

"Create comprehensive docs for https://github.com/user/api"

Migration Planning:

"How would I migrate https://github.com/user/js-app to TypeScript?"

⚙️ Configuration

Parameters

Examples

Limit files:

"Flatten https://github.com/large/repo with max 50 files"

Private repository:

"Flatten https://github.com/user/private with token ghp_your_token_here"

🔐 Private Repositories

For private repositories, you'll need a GitHub Personal Access Token:

1. Visit https://github.com/settings/tokens
2. Click "Generate new token (classic)"
3. Name it: RepoStack MCP
4. Select scope: repo (Full control of private repositories)
5. Click "Generate token"
6. Copy and save your token securely

Usage:

"Flatten https://github.com/mycompany/private-api with token ghp_abc123xyz"

📁 Supported File Types

RepoStack intelligently includes:

Languages

py js ts jsx tsx java cpp c h cs go rs rb php swift kt scala r m sh

Web

html css scss sass less vue svelte

Config

json yaml yml toml xml md txt

Special

Makefile Dockerfile .gitignore .env.example

🚫 Auto-Excluded

RepoStack automatically skips:

• 📁 Dependencies: node_modules/, vendor/, dist/, build/
• 🔒 Cache: __pycache__/, .next/, .nuxt/, coverage/
• 🔐 Env: venv/, env/, .venv/
• 📦 Lock files: package-lock.json, yarn.lock, pnpm-lock.yaml
• 🗜️ Minified: .min.js, .min.css, bundle.js

🐛 Troubleshooting

"ModuleNotFoundError: No module named 'fastmcp'"

pip install fastmcp httpx

"Server disconnected immediately"

Check your Python path:

where python      # Windows
which python3     # macOS/Linux

Update your config with the full path:

{
  "mcpServers": {
    "repostack": {
      "command": "C:\\Python313\\python.exe",
      "args": ["D:\\path\\to\\github_flattener.py"]
    }
  }
}

"GitHub API rate limit exceeded"

You've hit GitHub's rate limit (60 requests/hour without auth).

Solution: Add a GitHub token to increase limit to 5,000/hour.

"Can't see RepoStack in Claude"

1. ✅ Verify config file has no JSON syntax errors
2. ✅ Completely quit and restart Claude Desktop
3. ✅ Check logs: Claude → Help → View Logs
4. ✅ Test the server independently (see Testing section)

"Repository not found (404)"

• Verify the repository URL is correct
• For private repos, ensure you're using a valid token
• Check that the repository hasn't been deleted or renamed

🧪 Testing

Method 1: MCP Inspector (Recommended)

npx @modelcontextprotocol/inspector

Connect with:

• Command: python (or full path)
• Args: D:\path\to\github_flattener.py

Method 2: Direct Test Script

Create test.py:

import asyncio
import httpx
import os

# Copy the helper functions from your main file
CODE_EXTENSIONS = {
    '.py', '.js', '.ts', '.jsx', '.tsx', '.java', '.cpp', '.c', '.h', '.hpp',
    '.cs', '.go', '.rs', '.rb', '.php', '.swift', '.kt', '.scala', '.r',
    '.m', '.mm', '.sh', '.bash', '.zsh', '.sql', '.html', '.css', '.scss',
    '.sass', '.less', '.vue', '.svelte', '.json', '.yaml', '.yml', '.toml',
    '.xml', '.md', '.txt', '.gitignore', '.env.example', 'Makefile', 'Dockerfile'
}

EXCLUDE_PATTERNS = {
    'package-lock.json', 'yarn.lock', 'pnpm-lock.yaml',
    '.min.js', '.min.css', 'bundle.js'
}

async def fetch_repo_tree(owner: str, repo: str, token: str = None):
    url = f"https://api.github.com/repos/{owner}/{repo}/git/trees/main?recursive=1"
    headers = {"Accept": "application/vnd.github.v3+json"}
    if token:
        headers["Authorization"] = f"token {token}"
    
    async with httpx.AsyncClient() as client:
        response = await client.get(url, headers=headers, timeout=30.0)
        if response.status_code == 404:
            url = f"https://api.github.com/repos/{owner}/{repo}/git/trees/master?recursive=1"
            response = await client.get(url, headers=headers, timeout=30.0)
        
        response.raise_for_status()
        data = response.json()
        return data.get("tree", [])

async def fetch_file_content(owner: str, repo: str, path: str, token: str = None):
    url = f"https://raw.githubusercontent.com/{owner}/{repo}/main/{path}"
    headers = {}
    if token:
        headers["Authorization"] = f"token {token}"
    
    async with httpx.AsyncClient() as client:
        response = await client.get(url, headers=headers, timeout=30.0)
        if response.status_code == 404:
            url = f"https://raw.githubusercontent.com/{owner}/{repo}/master/{path}"
            response = await client.get(url, headers=headers, timeout=30.0)
        
        if response.status_code == 200:
            try:
                return response.text
            except:
                return f"[Binary file: {path}]"
        return ""

def should_include_file(path: str):
    for pattern in EXCLUDE_PATTERNS:
        if pattern in path:
            return False
    
    ignore_dirs = ['node_modules/', '.git/', 'dist/', 'build/', '__pycache__/', 
                   'venv/', 'env/', '.next/', '.nuxt/', 'coverage/']
    if any(ignore_dir in path for ignore_dir in ignore_dirs):
        return False
    
    _, ext = os.path.splitext(path)
    filename = os.path.basename(path)
    return ext.lower() in CODE_EXTENSIONS or filename in CODE_EXTENSIONS

async def test_flatten(repo_url: str, max_files: int = 20):
    """Test the flattening logic directly"""
    repo_url = repo_url.strip().rstrip('/')
    if "github.com/" in repo_url:
        parts = repo_url.split("github.com/")[1].split("/")
        owner, repo = parts[0], parts[1]
    else:
        parts = repo_url.split("/")
        owner, repo = parts[0], parts[1]
    
    print(f"Fetching repo: {owner}/{repo}")
    
    tree = await fetch_repo_tree(owner, repo)
    print(f"Total items in tree: {len(tree)}")
    
    files_to_fetch = [
        item for item in tree 
        if item["type"] == "blob" and should_include_file(item["path"])
    ][:max_files]
    
    print(f"Files to fetch: {len(files_to_fetch)}")
    
    output_lines = [
        f"# Flattened Repository: {owner}/{repo}",
        f"# Total files included: {len(files_to_fetch)}",
        "=" * 80,
        ""
    ]
    
    for item in files_to_fetch:
        path = item["path"]
        print(f"Fetching: {path}")
        content = await fetch_file_content(owner, repo, path)
        
        if content:
            output_lines.extend([
                "",
                f"{'=' * 80}",
                f"# FILE: {path}",
                f"{'=' * 80}",
                content,
                ""
            ])
    
    result = "\n".join(output_lines)
    return result

async def main():
    result = await test_flatten(
        repo_url="https://github.com/fastmcp/fastmcp",
        max_files=10
    )
    print("\n" + "="*80)
    print("RESULT (first 2000 chars):")
    print("="*80)
    print(result[:2000])
    print(f"\n... (Total length: {len(result)} characters)")

if __name__ == "__main__":
    asyncio.run(main())

Run:

python test.py

📊 Performance

RepoStack is optimized for speed:

• Small repos (<50 files): ~2-5 seconds
• Medium repos (50-200 files): ~5-15 seconds
• Large repos (200+ files): ~15-30 seconds

Times vary based on network speed and file sizes

🔒 Security & Privacy

• ✅ All data stays between you, GitHub, and Claude
• ✅ No data is stored or logged by RepoStack
• ✅ Tokens are used only for API authentication
• ✅ Open source - audit the code yourself

🛠️ Development

Project Structure

repostack/
├── github_flattener.py    # Main MCP server
├── requirements.txt       # Dependencies
├── README.md             # This file
├── LICENSE               # MIT License
└── tests/                # Test files
    └── test_mcp.py

Contributing

We love contributions!

1. Fork the repository
2. Create a feature branch: git checkout -b feature/amazing-feature
3. Commit your changes: git commit -m 'Add amazing feature'
4. Push to the branch: git push origin feature/amazing-feature
5. Open a Pull Request

📝 Requirements

Create requirements.txt:

fastmcp>=0.1.0
httpx>=0.27.0

Install:

pip install -r requirements.txt

🎓 Use Cases

For Developers

• 🔍 Quick codebase exploration
• 🐛 Bug hunting across large projects
• 📚 Learning from popular open-source projects
• 🔄 Refactoring assistance

For Teams

• 👀 Code review automation
• 📖 Documentation generation
• 🏗️ Architecture analysis
• 🔐 Security audits

For Learners

• 📖 Study real-world codebases
• 🎯 Understand design patterns
• 🚀 Learn best practices
• 💡 Get explanations of complex code

🌟 Pro Tips

1. Start Small: Test with smaller repos first (< 50 files)
2. Use Tokens: Always use a GitHub token to avoid rate limits
3. Limit Files: Use max_files for very large repositories
4. Be Specific: Ask Claude specific questions about the code
5. Combine Tools: Use with Claude's other tools for maximum power

📜 License

MIT License - use it, modify it, ship it! See LICENSE file for details.

🙏 Credits

Built with love using:

• FastMCP - Simplified MCP development
• httpx - Modern async HTTP
• Model Context Protocol - By Anthropic

📞 Support

Need help? Here's how to get it:

1. 📖 Check the Troubleshooting section
2. 🔍 Review MCP Documentation
3. 💬 Open an
4. 📧 Contact: your-email@example.com

🚀 What's Next?

Upcoming features:

• Support for GitLab and Bitbucket
• Custom file filtering rules
• Diff analysis between branches
• Direct Git clone support
• Repository comparison tool
• Code statistics and insights

⭐ If RepoStack helped you, give it a star on GitHub!

Made with ❤️ by Om Manoj Sharma for the developer community