npm-sentinel

NPM Sentinel MCP is an AI-assisted MCP server focused on analyzing NPM packages. It provides real-time insights into security, dependencies, size, and maintenance metrics, with capabilities to map dependency graphs, verify TypeScript support, and perform integrity checks. The server exposes its tools via MCP-compatible interfaces and supports both STDIO and HTTP streamable transports (via Smithery.ai), making it suitable for local development or scalable deployments. When using the JSON-RPC style tools, you can pass an ignoreCache flag to force a fresh lookup from the registry, ensuring you always get up-to-date results. For example, you can query for npmVersions or perform package-level analysis with caching bypass to get the most current data.

To run locally, you typically use npx to pull in the latest MCP server package and start the server in STDIO mode. The project also includes Docker and HTTP deployment options, useful for CI/CD or cloud deployments. The server supports environment customization via variables like NPM_REGISTRY_URL to point to a private or public registry and offers a structured, standardized MCP response format with robust input validation and caching strategies.

Prerequisites:

• Node.js (recommended LTS) and npm
• Optional: Docker if you want to run via Docker
• Optional: Smithery.ai account for HTTP streamable deployment

Install and run (STDIO / local development):

1. Ensure Node.js and npm are installed. Check versions: node -v npm -v

2. Install and run the MCP server using npx (pulls the latest release): npm install -g npm // if needed to ensure npm is up to date npx -y @nekzus/mcp-server@latest

3. Start the server (STDIO by default). The command above will start the server in STDIO mode. If the package provides a CLI, you can also run the dedicated start script, for example: npm run start

4. Optional: Run via Docker Build the image: docker build -t nekzus/npm-sentinel-mcp . Run the container with an attached /projects directory (as described in the README): docker run -i --rm -w /projects -v $(pwd):/projects nekzus/npm-sentinel-mcp node dist/index.js

5. Optional: HTTP stem/Smithery deployment Follow the Smithery.ai deployment steps and configure the mcpServers.http entry as documented in the README.

Tips and common considerations:

• Environment variable: NPM_REGISTRY_URL can be set to point to a private registry or a custom mirror. Default is https://registry.npmjs.org.
• Cache behavior: Tools accept ignoreCache: true to bypass the cache and fetch fresh data from the registry.
• When using Docker, ensure all mounted directories are accessible under /projects.
• The server supports both STDIO and HTTP streamable transports; choose the transport that fits your deployment scenario (local development vs. cloud/edge deployments).
• If you encounter type or schema validation errors, verify that your input conforms to the MCP schema and that the registry URL is reachable.
• To update to the latest MCP server, you can re-run the npx install command with @latest.