The full upstream README, mirrored here for reference. Install config, tool schemas, adoption signals, and an original overview live on the Mooring listing page.
Mooring lines for your code — Git and GitHub operations for AI tools.
Mooring is an MCP server that gives AI assistants structured access to local Git repositories and the GitHub API. Local operations use GitPython (no subprocess calls). GitHub operations use the PyGithub library with token masking and rate limit handling built in.
| Tool | Description | Key Parameters |
|---|---|---|
repo_status | Branch, ahead/behind, stash count, staged/unstaged/untracked files | repo_path |
repo_log | Commit log with optional filters | repo_path, max_count, author, since, path, search |
repo_diff | Unified diff — working tree, staged, or between refs | repo_path, staged, from_ref, to_ref |
repo_blame | Git blame with optional line range | repo_path, file_path, start_line, end_line |
repo_branches | All branches with tracking info, last commit, ahead/behind | repo_path |
repo_stash | Stash operations: list, push, pop, apply | repo_path, action, message |
| Tool | Description | Key Parameters |
|---|---|---|
gh_pr_list | List pull requests | repo, state, author, label |
gh_pr_detail | PR detail with reviews, comments, and check runs | repo, number |
gh_pr_create | Create a pull request | repo, title, body, head, base, labels, reviewers |
gh_issues | List, create, or update issues | repo, state, action, title, body, number |
gh_actions | List recent GitHub Actions workflow runs | repo, workflow, status |
| Tool | Description | Key Parameters |
|---|---|---|
health | Server version and status check | (none) |
Run the server directly:
Add to your Claude Desktop configuration (claude_desktop_config.json):
The GITHUB_TOKEN environment variable is required for all GitHub tools (gh_*). Local Git tools work without it.
ghp_*, gho_*, github_pat_*) before being returnedIf you're running multiple MCP servers, route them through a gateway like tbxark/mcp-proxy to manage all child processes from a single persistent service. The proxy handles process lifecycle, centralized config, and crash recovery — each server still gets its own SSE endpoint but you manage everything from one config file instead of scattered Claude Code entries.
For a full walkthrough of how this works in practice, see The Hidden Cost of a Loaded MCP Stack on charlieseay.com.