Miggo MCP Server

Ask your AI assistant about your live application security — vulnerabilities, services, endpoints, findings, dependencies, and third-party integrations — all through natural language.

Miggo's MCP server connects Claude, Cursor, VS Code Copilot, and other AI assistants directly to your Miggo environment. Instead of clicking through dashboards, just ask questions and get answers grounded in real-time data from your running applications.

Quick start

1. Get your API token

Create an API token in the Miggo Integrations portal.

2. Install

1. Download the latest .mcpb from the releases page
2. Open the file — Claude Desktop installs it automatically
3. Enter your API token when prompted

Claude Desktop ships its own uv and provisions the Python runtime from the bundle's pyproject.toml. No system Python install required.

Click the badge to auto-install:

Or manually add to your Cursor MCP settings:

{
  "mcpServers": {
    "miggo": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/miggo-io/miggo-mcp", "miggo-mcp"],
      "env": {
        "MIGGO_PUBLIC_TOKEN": "<your-token>"
      }
    }
  }
}

Requires uv on your PATH.

Add to your VS Code settings.json (Cmd+Shift+P → "Preferences: Open User Settings (JSON)"):

{
  "mcp": {
    "servers": {
      "miggo": {
        "command": "uvx",
        "args": ["--from", "git+https://github.com/miggo-io/miggo-mcp", "miggo-mcp"],
        "env": {
          "MIGGO_PUBLIC_TOKEN": "<your-token>"
        }
      }
    }
  }
}

Requires uv on your PATH.

1. Open Settings → Tools → AI Assistant → MCP Servers
2. Add a new server with the following configuration:

{
  "miggo": {
    "command": "uvx",
    "args": ["--from", "git+https://github.com/miggo-io/miggo-mcp", "miggo-mcp"],
    "env": {
      "MIGGO_PUBLIC_TOKEN": "<your-token>"
    }
  }
}

Requires uv on your PATH.

Point your client at uvx using the standard MCP stdio configuration:

{
  "mcpServers": {
    "miggo": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/miggo-io/miggo-mcp", "miggo-mcp"],
      "env": {
        "MIGGO_PUBLIC_TOKEN": "<your-token>"
      }
    }
  }
}

Requires uv on your PATH.

Example prompts

Once installed, try these to see what Miggo + your AI assistant can do:

Configuration

Development

Running locally

Note: Ensure MIGGO_PUBLIC_TOKEN is exported in your environment.

# Start the stdio server
uv run ./run.py

# Or launch the MCP Inspector for interactive testing
uv run mcp dev ./run.py

Building artifacts

uv run python scripts/build.py

Produces the MCPB bundle at dist/miggo-mcp.mcpb (a small zip containing manifest.json, pyproject.toml, uv.lock, run.py, and the src/ tree). Claude Desktop's bundled uv provisions the runtime on the user's machine when the .mcpb is installed; no platform-specific binaries are produced. Pass --help for options.

Testing, formatting, linting

uv run pytest
uv run ruff format .
uv run ruff check .

Install pre-commit hooks:

uv run pre-commit install

Release process

We use release-please to automate releases and version bumps. In normal operation, just review and merge the release PR that release-please opens. The workflow can also be triggered manually if needed.

Contributing

See CONTRIBUTING.md for guidelines on submitting issues and pull requests.

License

MIT