Query Package Repositories for LLMs

Pacman MCP Server

A Model Context Protocol server that provides package index querying capabilities. This server enables LLMs to search and retrieve information from package repositories like PyPI, npm, crates.io, Docker Hub, and Terraform Registry.

Available Tools

• search_package - Search for packages in package indices

  • index (string, required): Package index to search ("pypi", "npm", "crates", "terraform")
  • query (string, required): Package name or search query
  • limit (integer, optional): Maximum number of results to return (default: 5, max: 50)

• package_info - Get detailed information about a specific package

  • index (string, required): Package index to query ("pypi", "npm", "crates", "terraform")
  • name (string, required): Package name
  • version (string, optional): Specific version to get info for (default: latest)

• search_docker_image - Search for Docker images in Docker Hub

  • query (string, required): Image name or search query
  • limit (integer, optional): Maximum number of results to return (default: 5, max: 50)

• docker_image_info - Get detailed information about a specific Docker image

  • name (string, required): Image name (e.g., user/repo or library/repo)
  • tag (string, optional): Specific image tag (default: latest)

• terraform_module_latest_version - Get the latest version of a Terraform module

  • name (string, required): Module name (format: namespace/name/provider)

Prompts

• search_pypi

  • Search for Python packages on PyPI
  • Arguments:
    • query (string, required): Package name or search query

• pypi_info

  • Get information about a specific Python package
  • Arguments:
    • name (string, required): Package name
    • version (string, optional): Specific version

• search_npm

  • Search for JavaScript packages on npm
  • Arguments:
    • query (string, required): Package name or search query

• npm_info

  • Get information about a specific JavaScript package
  • Arguments:
    • name (string, required): Package name
    • version (string, optional): Specific version

• search_crates

  • Search for Rust packages on crates.io
  • Arguments:
    • query (string, required): Package name or search query

• crates_info

  • Get information about a specific Rust package
  • Arguments:
    • name (string, required): Package name
    • version (string, optional): Specific version

• search_docker

  • Search for Docker images on Docker Hub
  • Arguments:
    • query (string, required): Image name or search query

• docker_info

  • Get information about a specific Docker image
  • Arguments:
    • name (string, required): Image name (e.g., user/repo)
    • tag (string, optional): Specific tag

• search_terraform

  • Search for Terraform modules in the Terraform Registry
  • Arguments:
    • query (string, required): Module name or search query

• terraform_info

  • Get information about a specific Terraform module
  • Arguments:
    • name (string, required): Module name (format: namespace/name/provider)

• terraform_latest_version

  • Get the latest version of a specific Terraform module
  • Arguments:
    • name (string, required): Module name (format: namespace/name/provider)

Installation

Using uv (recommended)

When using uv no specific installation is needed. We will
use uvx to directly run mcp-server-pacman.

Using PIP

Alternatively you can install mcp-server-pacman via pip:

pip install mcp-server-pacman

After installation, you can run it as a script using:

python -m mcp_server_pacman

Using Docker

You can also use the Docker image:

docker pull oborchers/mcp-server-pacman:latest
docker run -i --rm oborchers/mcp-server-pacman

Configuration

Configure for Claude.app

Add to your Claude settings:

"mcpServers": {
  "pacman": {
    "command": "uvx",
    "args": ["mcp-server-pacman"]
  }
}

"mcpServers": {
  "pacman": {
    "command": "docker",
    "args": ["run", "-i", "--rm", "oborchers/mcp-server-pacman:latest"]
  }
}

"mcpServers": {
  "pacman": {
    "command": "python",
    "args": ["-m", "mcp-server-pacman"]
  }
}

Configure for VS Code

For manual installation, add the following JSON block to your User Settings (JSON) file in VS Code. You can do this by pressing Ctrl + Shift + P and typing Preferences: Open User Settings (JSON).

Optionally, you can add it to a file called .vscode/mcp.json in your workspace. This will allow you to share the configuration with others.

Note that the mcp key is needed when using the mcp.json file.

{
  "mcp": {
    "servers": {
      "pacman": {
        "command": "uvx",
        "args": ["mcp-server-pacman"]
      }
    }
  }
}

{
  "mcp": {
    "servers": {
      "pacman": {
        "command": "docker",
        "args": ["run", "-i", "--rm", "oborchers/mcp-server-pacman:latest"]
      }
    }
  }
}

Customization - User-agent

By default, the server will use the user-agent:

ModelContextProtocol/1.0 Pacman (+https://github.com/modelcontextprotocol/servers)

This can be customized by adding the argument --user-agent=YourUserAgent to the args list in the configuration.

Development

Running Tests

• Run all tests:

  uv run pytest -xvs
  

• Run specific test categories:

  # Run all provider tests
  uv run pytest -xvs tests/providers/
  
  # Run integration tests for a specific provider
  uv run pytest -xvs tests/integration/test_pypi_integration.py
  
  # Run specific test class
  uv run pytest -xvs tests/providers/test_npm.py::TestNPMFunctions
  
  # Run a specific test method
  uv run pytest -xvs tests/providers/test_pypi.py::TestPyPIFunctions::test_search_pypi_success
  

• Check code style:

  uv run ruff check .
  uv run ruff format --check .
  

• Format code:

  uv run ruff format .
  

Debugging

You can use the MCP inspector to debug the server. For uvx installations:

npx @modelcontextprotocol/inspector uvx mcp-server-pacman

Or if you've installed the package in a specific directory or are developing on it:

cd path/to/pacman
npx @modelcontextprotocol/inspector uv run mcp-server-pacman

Release Process

The project uses GitHub Actions for automated releases:

1. Update the version in pyproject.toml
2. Create a new tag with git tag vX.Y.Z (e.g., git tag v0.1.0)
3. Push the tag with git push --tags

This will automatically:

• Verify the version in pyproject.toml matches the tag
• Run tests and lint checks
• Build and publish to PyPI
• Build and publish to Docker Hub as oborchers/mcp-server-pacman:latest and oborchers/mcp-server-pacman:X.Y.Z

Project Structure

The codebase is organized into the following structure:

src/mcp_server_pacman/
├── models/             # Data models/schemas
├── providers/          # Package registry API clients
│   ├── pypi.py         # PyPI API functions
│   ├── npm.py          # npm API functions
│   ├── crates.py       # crates.io API functions
│   ├── dockerhub.py    # Docker Hub API functions
│   └── terraform.py    # Terraform Registry API functions
├── utils/              # Utilities and helpers
│   ├── cache.py        # Caching functionality
│   ├── constants.py    # Shared constants
│   └── parsers.py      # HTML parsing utilities
├── __init__.py         # Package initialization
├── __main__.py         # Entry point
└── server.py           # MCP server implementation

Tests follow a similar structure:

tests/
├── integration/        # Integration tests (real API calls)
├── models/             # Model validation tests
├── providers/          # Provider function tests
└── utils/              # Test utilities