What is Mcp Package Manager

The Package Manager MCP Server is an implementation designed to facilitate package management for Node.js (using npm) and Python (using UV). It provides functionalities for installing and uninstalling packages, managing virtual environments, and initializing projects with structured metadata files like package.json and pyproject.toml.

Use cases

This server is ideal for developers who require seamless integration of Node.js and Python package management in their projects. It assists in managing dependencies, setting up development environments, and maintaining security practices within project setups. It supports project initialization and ensures that only approved packages are installed.

How to use

To use the Package Manager MCP Server, set up a virtual environment, install necessary dependencies, and configure environment variables for logging, allowed packages, and paths. Start the server by adding its configuration to your Claude Desktop setup. Use the available tools to install or uninstall packages, initialize projects, or create virtual environments as needed.

Key features

Key features include node and Python package management, requirements.txt support with whitelist validation, robust virtual environment management, cross-platform support, comprehensive logging, and various security measures like package whitelisting and resource limits. It also provides project initialization functionality for setting up new projects swiftly.

Where to use

The server can be utilized in various development environments that leverage both Node.js and Python, such as web applications, data science projects, and software development workflows. It is suitable for individual developers, teams, and organizations needing consistent package management across different technological stacks.

Package Manager MCP Server

A Model Context Protocol (MCP) server implementation for package management in Claude Desktop. This server handles npm (for Node.js) and uv (for Python) package installations, uninstallations, project initializations, and virtual environment management.

Features

• Node.js package management (using npm)
• Python package management (using uv)
  • Package installation and uninstallation
  • Requirements.txt support with whitelist validation
  • Virtual environment creation and management
  • Platform-specific path detection
• Project initialization (package.json/pyproject.toml)
• Security features:
  • Package whitelist validation
  • Path verification
  • Resource limits
  • Timeout handling
• Cross-platform support:
  • Windows path detection for npm and uv
  • Unix/Linux compatibility
• Comprehensive logging and error handling

Requirements

• Python 3.12+
• Node.js (for npm operations)
• UV for Python package management

Installation

# Create and activate a virtual environment (recommended)
uv venv .venv
source .venv/bin/activate  # Unix/Linux
# or
.venv\Scripts\activate     # Windows

# Install dependencies
uv pip install -e .

Configuration

Environment variables:

# Common Settings
LOG_LEVEL=DEBUG
NODE_ENV=development
PYTHON_ENV=development

# Package Management
ALLOWED_PACKAGES=typescript,react,express,requests,pandas...
MAX_INSTALL_SIZE=50000000
PROJECT_DIR=H:/projects

# Timeouts (in milliseconds)
INSTALL_TIMEOUT=300000
UNINSTALL_TIMEOUT=60000
INIT_TIMEOUT=30000
VENV_TIMEOUT=30000

# Package Managers Configuration
NPM_PATH=npm                  # Optional: Auto-detected
UV_PATH=uv                    # Optional: Auto-detected
USE_UV=true
VENV_NAME=.venv

Usage

Claude Desktop Configuration

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "package-manager": {
      "command": "python",
      "args": [
        "-m",
        "package_manager_mcp.server"
      ],
      "env": {
        "ALLOWED_PACKAGES": "typescript,react,express,requests,pandas...",
        "MAX_INSTALL_SIZE": "50000000",
        "PROJECT_DIR": "H:/projects",
        "LOG_LEVEL": "DEBUG",
        "NODE_ENV": "development",
        "USE_UV": "true",
        "UV_PATH": "uv",
        "VENV_NAME": ".venv",
        "PYTHON_VERSION": "3.12",
        "VENV_TIMEOUT": "30000"
      }
    }
  }
}

Available Tools

1. install

   • Install npm (Node.js) or uv (Python) packages
   • Supports requirements.txt with whitelist validation
   • Handles versioning and dependencies
   • Respects security whitelist and timeouts

2. uninstall

   • Remove installed packages
   • Supports both npm and uv
   • Clean uninstallation with dependency handling

3. init

   • Initialize new projects
   • Creates package.json (npm) or pyproject.toml (uv)
   • Configures initial project structure

4. create_venv

   • Create Python virtual environments
   • Uses UV for reliable environment creation
   • Path verification and security checks
   • Configurable venv name and location

Development

Setup

# Clone repository
git clone https://github.com/your-org/package-manager-mcp.git

# Create virtual environment
uv venv .venv
source .venv/bin/activate  # Unix/Linux
# or
.venv\Scripts\activate     # Windows

# Install development dependencies
uv pip install -r requirements-dev.txt

Testing

# Run tests
pytest

# Run with coverage
pytest --cov=package_manager_mcp

# Run specific test modules
pytest tests/test_uv_manager.py

Logging

The server provides comprehensive logging with:

• Console output
• File logging (package_manager.log)
• Configurable log levels
• Detailed error messages and stack traces

Security

• Package whitelist enforcement
• Path traversal protection
• Virtual environment path verification
• Resource usage limits
• Operation timeouts
• Comprehensive input validation
• Secure subprocess management

License

MIT License