🚀 Open Source Contribution MCP Server (open-source-contribution-mcp-server)

An end-to-end Model Context Protocol (MCP) Server that enables AI clients (Claude Desktop, Cursor, AI agents) to autonomously discover, analyze, fix, test, submit, and track open-source contributions on GitHub.

🌟 Overview

Making your first (or 50th) open-source contribution can be intimidating: finding active repositories with open PR queues, understanding unwritten codebase conventions, passing strict linters, and avoiding silent CI failures.

This MCP server provides 12 specialized tools across 6 domain modules to automate the entire contribution lifecycle.

Related MCP server: mcp-oss-onramp

🛠️ Complete Tool Catalog (12 Tools)

👤 1. Skill & Preference Management Module

Allows AI clients to store and reference your tech stack, language proficiencies, and project preferences.

🔍 2. Repository & Issue Discovery Module

Filters GitHub for high-confidence, active repositories with zero PR queues.

📊 3. Deep Repository Context & Analysis Module

Inspects repository rules, linters, and architectural patterns before writing code.

🛠️ 4. Code & Test Generation Module

Synthesizes compliant source code patches and unit tests.

🚀 5. Automated Pull Request Module

Automates the complete GitHub Git workflow.

💬 6. Review Feedback & Iteration Module

Keeps PRs moving toward merge.

🔐 GitHub Personal Access Token (PAT) Setup

To allow the MCP Server to search repositories, inspect codebases, fork repos, and create pull requests, you need a GitHub Personal Access Token.

Step-by-Step Instructions:

1. Go to GitHub Settings: GitHub Developer Settings > Personal Access Tokens (Classic)

2. Click Generate new token (classic).

3. Set a descriptive note: MCP Open Source Contribution Server.

4. Select the following required scopes:

   • repo (Full control of private/public repositories — required for forking & creating PRs)

   • workflow (Update GitHub Action workflows — required if modifying workflow files)

   • user:email (Access user email addresses)

5. Click Generate token and copy the generated token string (ghp_...).

6. Store it in your .env file or environment variables:

GITHUB_TOKEN=ghp_your_personal_access_token_here

⚡ Installation & Setup

Prerequisites

• Node.js: v18.0.0 or higher

• npm or yarn or pnpm

• Git installed on system path

1. Clone & Install Dependencies

git clone https://github.com/your-username/open-source-contribution-mcp-server.git
cd open-source-contribution-mcp-server
npm install

2. Configure Environment

Create a .env file in the root directory (or copy .env.example):

cp .env.example .env

Edit .env and add your GitHub Personal Access Token:

GITHUB_TOKEN=ghp_your_personal_access_token_here

3. Build the TypeScript Server

npm run build

🔌 Connecting to MCP Clients

A. Claude Desktop Integration

Add the server to your Claude Desktop configuration file:

• Windows: %APPDATA%\Claude\claude_desktop_config.json

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "open-source-contributions": {
      "command": "node",
      "args": [
        "C:/path/to/open-source-contribution-mcp-server/dist/index.js"
      ],
      "env": {
        "GITHUB_TOKEN": "ghp_your_personal_access_token_here"
      }
    }
  }
}

B. Cursor / Antigravity Integration (mcp_config.json)

{
  "mcpServers": {
    "open-source-contributions": {
      "command": "node",
      "args": [
        "C:/path/to/open-source-contribution-mcp-server/dist/index.js"
      ],
      "env": {
        "GITHUB_TOKEN": "ghp_your_personal_access_token_here"
      }
    }
  }
}

🧪 Testing & MCP Inspector

You can inspect and test all 12 tools interactively using the official @modelcontextprotocol/inspector:

npm run inspector

This launches the interactive MCP web UI at http://localhost:5173.

🏆 Real-World Battle-Tested Proof

This MCP server has been battle-tested on live open-source repositories:

1. 🟣 mldsveda/PyScrappy PR #139 — MERGED TO main

   • Fixed a critical atomic file-saving bug (tempfile.mkstemp + os.replace()). Passed 100% of Python 3.9–3.13 unit tests and Ruff formatting.

2. 🟢 fu351/Doberman-Core PR #345 — OPEN & PASSING CI

   • Retuned light-mode verdict badge colors to clear WCAG AA 4.5:1 text contrast with unit tests.

📄 License

Distributed under the ISC License. See LICENSE for more information.