Feature Request: GitHub MCP Server

[afrah123456]

Overview

I propose adding a GitHub MCP server to enable AI assistants to interact with GitHub repositories, issues, pull requests, and more through natural language.

Motivation

GitHub is the most widely used development platform, but there's currently no GitHub integration in the IBM MCP collection. This server would enable developers to automate common GitHub workflows through AI assistants.

Phase 1: MVP (Initial Contribution)

I'll start with 15 core features covering essential GitHub operations:

Repository Management (3 tools):

• list_repositories - List user's repositories
• get_repository - Get detailed repository information
• create_repository - Create a new repository

Issue Management (4 tools):

• list_issues - List issues in a repository
• get_issue - Get detailed issue information
• create_issue - Create a new issue
• close_issue - Close an issue

Pull Request Management (4 tools):

• list_pull_requests - List pull requests
• get_pull_request - Get detailed PR information
• create_pull_request - Create a new pull request
• merge_pull_request - Merge a pull request

Branch Management (2 tools):

• list_branches - List repository branches
• create_branch - Create a new branch

File Operations (2 tools):

• get_file_contents - Read file contents from repository
• create_or_update_file - Create or update a file

Plus: OAuth/Token authentication, error handling, rate limiting, comprehensive tests, and documentation.

Phase 2: Enhanced Features (Future PRs)

After MVP is merged, I plan to add:

• Issue comments, assignments, and labels
• PR reviews and comments
• Branch deletion
• File deletion
• Code search
• Collaborator management
• Milestones and labels
• Commit history

This phased approach allows for early feedback and iterative improvement.

Example Use Cases

Automated Issue Creation:

User: "Create an issue: Fix login bug on mobile"
AI → GitHub MCP → Issue created

PR Management:

User: "Show me all open PRs and merge any that are approved"
AI → GitHub MCP → Lists PRs and merges approved ones

Quick Repository Info:

User: "What's the status of my resume-matcher repo?"
AI → GitHub MCP → Returns stars, forks, last update, open issues

Implementation Details

• Language: Python 3.11+
• GitHub API: PyGithub library
• Authentication: OAuth 2.0 / Personal Access Token
• MCP Framework: fastmcp
• Testing: pytest with comprehensive test coverage
• Documentation: Complete README with examples

Timeline

Phase 1 MVP:

• Days 1-2: Setup + Repository tools
• Days 3-4: Issues + PR tools
• Day 5: Branches + Files
• Day 6: Testing + Bug fixes
• Day 7: Documentation + PR submission

Phase 2+: Iterative improvements based on maintainer feedback

Questions

• Would maintainers be interested in this phased contribution approach?
• Any specific coding standards or patterns I should follow?
• Should I prioritize OAuth or Token auth (or both)?
• Any rate limiting strategies you recommend?

About Me

MS Computer Science student at Northeastern University with experience in Python, REST APIs, and backend development. Excited to contribute to IBM MCP!

Looking forward to your feedback!