Mode: Cognitive/Prompt-Driven — No standalone utility script; use via agent context.

GitHub Skill

Overview

This skill provides access to the official GitHub MCP server with progressive disclosure for optimal context usage.

Context Savings: ~95% reduction

• MCP Mode: ~50,000 tokens always loaded (80+ tools)
• Skill Mode: ~500 tokens metadata + on-demand loading

Requirements

• Docker installed and running
• GITHUB_PERSONAL_ACCESS_TOKEN environment variable set

Toolsets

The server provides 80+ tools across 19 toolsets:

Quick Reference

Use the gh CLI for GitHub operations:

# Get repository info
gh repo view anthropics/claude-code

# List issues
gh issue list --repo anthropics/claude-code

# Search code
gh search code "language:python MCP"

# Create issue
gh issue create --repo me/myrepo --title "Bug" --body "Description"

# List pull requests
gh pr list --repo anthropics/claude-code

Common Tools (Default Toolsets: 40 tools)

Repository Operations

• search_repositories – Search for repositories
• create_repository – Create a new repository
• fork_repository – Fork a repository
• list_commits – List repository commits
• get_commit – Get commit details
• get_file_contents – Get file contents from a repository
• create_or_update_file – Create or update a file
• delete_file – Delete a file
• push_files – Push multiple files
• search_code – Search for code across GitHub
• list_branches – List repository branches
• create_branch – Create a new branch
• list_tags – List repository tags
• get_tag – Get tag details
• list_releases – List releases
• get_latest_release – Get latest release
• get_release_by_tag – Get release by tag

Issue Operations

• list_issues – List repository issues
• issue_read – Read issue details
• issue_write – Create/update issues
• add_issue_comment – Add a comment to an issue
• search_issues – Search for issues
• list_issue_types – List issue types (for organizations)
• get_label – Get label details
• sub_issue_write – Manage sub-issues
• assign_copilot_to_issue – Assign Copilot to an issue

Pull Request Operations

• list_pull_requests – List repository pull requests
• pull_request_read – Read PR details
• create_pull_request – Create a new PR
• update_pull_request – Update a PR
• update_pull_request_branch – Update PR branch
• merge_pull_request – Merge a PR
• search_pull_requests – Search for pull requests
• pull_request_review_write – Create/submit PR reviews
• add_comment_to_pending_review – Add comments to pending review
• request_copilot_review – Request Copilot review

User & Team Operations

• get_me – Get current authenticated user
• search_users – Search for users
• get_teams – Get organization teams
• get_team_members – Get team members

Configuration

The skill uses Docker to run the official GitHub MCP server:

• Image: ghcr.io/github/github-mcp-server
• Auth: GITHUB_PERSONAL_ACCESS_TOKEN environment variable

Environment Variables

Limiting Toolsets

When using MCP, configure toolsets via environment variables:

# Only repos and issues
GITHUB_TOOLSETS=repos,issues

# Only pull requests and code security
GITHUB_TOOLSETS=pull_requests,code_security

Error Handling

If operations fail:

1. Verify Docker is running: docker ps
2. Check GitHub token is set: echo $GITHUB_PERSONAL_ACCESS_TOKEN
3. Ensure token has required permissions for the operation
4. Use gh auth status to verify authentication

Related

• Official GitHub MCP Server: https://github.com/github/github-mcp-server
• GitHub API Documentation: https://docs.github.com/en/rest

Memory Protocol (MANDATORY)

Before starting: Read .claude/context/memory/learnings.md

After completing:

• New pattern -> .claude/context/memory/learnings.md
• Issue found -> .claude/context/memory/issues.md
• Decision made -> .claude/context/memory/decisions.md

ASSUME INTERRUPTION: If it’s not in memory, it didn’t happen.