MCP Server: GitHub Intelligence

An MCP (Model Context Protocol) server that gives Claude direct access to GitHub's REST API for repository analytics, issue tracking, pull-request status, contributor insights, and code search.

Built with the official Anthropic MCP Python SDK (FastMCP).

---

Features

Tool	Description
get_repository_info	Comprehensive repo details — stars, forks, language, license, topics, and more
list_issues	Browse and filter issues by state and labels
list_pull_requests	List PRs with branch info, draft/merge status
get_contributor_stats	Top contributors ranked by commit count
search_repositories	Search all of GitHub with language and sort filters

---

Quick Start

1. Install dependencies

pip install -r requirements.txt

2. (Optional) Configure a GitHub token

Copy the example env file and add your token for higher rate limits:

cp .env.example .env
# Edit .env and paste your token

3. Run the server

python server.py

4. Connect to Claude Desktop

Add the following to your Claude Desktop configuration file:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "github-intel": {
      "command": "python",
      "args": ["/absolute/path/to/server.py"]
    }
  }
}

Restart Claude Desktop. The five GitHub tools will appear automatically.

---

Authentication

The server works in two modes:

Mode	Rate Limit	Setup
Unauthenticated	60 requests / hour	Nothing required
Authenticated	5,000 requests / hour	Set GITHUB_TOKEN in .env

To create a personal access token, visit github.com/settings/tokens. A classic token with no scopes is sufficient for public repository data.

---

Example Queries

Once the server is connected to Claude, try asking:

• "Get me an overview of the facebook/react repository."
• "Show me the open issues labeled 'bug' in vercel/next.js."
• "List the most recent pull requests for rust-lang/rust."
• "Who are the top contributors to torvalds/linux?"
• "Search GitHub for the most-starred Python machine learning frameworks."

---

Rate Limits

GitHub enforces rate limits per IP (unauthenticated) or per token (authenticated).

• Without token: 60 requests per hour.
• With token: 5,000 requests per hour.

The server detects rate-limit errors (HTTP 403) and returns a clear message with the reset timestamp so you know when to retry.

---

Tech Stack

• Python 3.10+
• mcp — Official Anthropic MCP SDK (FastMCP)
• httpx — Async HTTP client
• pydantic — Data validation (used by MCP internally)
• python-dotenv — .env file loading

---

License

MIT