Mcp Server Github Actions

What is Mcp Server Github Actions?

Mcp Server Github Actions is an MCP server that lets AI assistants manage GitHub Actions workflows directly from the editor. It provides tools to list workflows and runs, read logs, re-run failed jobs, cancel builds, and trigger deployments, bridging the gap left by GitHub’s official MCP server which does not cover Actions. It is designed for developers using Claude Desktop, Cursor, VS Code Copilot, or any MCP client.

How to use Mcp Server Github Actions?

Run npx mcp-server-github-actions with a GITHUB_TOKEN environment variable set. Configure the server in your MCP client’s settings (e.g., ~/.cursor/mcp.json for Cursor, claude_desktop_config.json for Claude Desktop, or environment for VS Code). The token must have Actions read and write permissions.

Key features of Mcp Server Github Actions

• List all workflow files in a repository
• List workflow runs with optional filtering
• Get detailed information for a specific run
• Retrieve the logs URL for a run
• Re-run an entire workflow or only failed jobs
• Cancel an in-progress or queued run
• List artifacts produced by a workflow run
• Trigger a workflow via workflow_dispatch

Use cases of Mcp Server Github Actions

• Diagnose why the last CI run failed without leaving the editor
• Re-run failed jobs after fixing a flaky test
• Cancel a stuck or unnecessary workflow run
• Trigger a deployment workflow on a specific branch
• Inspect artifacts from a completed build

FAQ from Mcp Server Github Actions

How does this differ from GitHub’s official MCP server?

GitHub’s official MCP server covers repos, issues, and PRs but does not include Actions. This server fills that gap, allowing you to manage CI/CD workflows directly from your AI assistant.

What are the runtime requirements?

You need Node.js (to run npx) and a GitHub personal access token with Actions read and write permissions. The server communicates over the MCP stdio transport.

Where does the data live?

All data is fetched from the GitHub REST API. No data is stored locally; the server acts as a thin proxy between your AI assistant and GitHub.

What authentication is needed?

A GitHub Personal Access Token (fine-grained recommended) with the “Actions: Read and Write” permission (or classic repo scope). Set it as the GITHUB_TOKEN environment variable.

Are there any known limitations?

The server relies on the GitHub API; rate limits apply. It does not support Actions OIDC or self-hosted runner management out of the box.