Comet MCP server provides AI-assisted development capabilities for MetroStar’s Comet ecosystem. The MCP server exposes model-context-protocol endpoints that help with intelligent code tasks, such as context-aware suggestions, automated code actions, and workflow automation integrated into Comet projects. Once you start the server, you can connect your development tooling to receive contextual hints, automate repetitive coding tasks, and orchestrate model-driven actions within your editor or CI/CD pipelines. The server is designed to work alongside the Comet component libraries and data visualization packs to streamline development across React/TypeScript projects that leverage USWDS components.

To use it, install the MCP package and run the server via your preferred package manager. The MCP server can be queried by clients that implement the MCP endpoints for context, prompts, and actions. Typical workflows include integrating model-assisted code suggestions into your editor, triggering automated refactors, and generating documentation snippets or test stubs based on your codebase and project conventions. If you’re using the Comet starter or a Correlated workflow, connect the MCP server to your development environment and configure it to target your repository and build/test pipelines for a seamless AI-assisted experience.

Prerequisites:

• Node.js v14+ (for npx usage) or npm installed
• Internet access to fetch the MCP package

Installation steps:

1. Ensure you have Node.js and npm installed. Verify: node -v npm -v

2. Install and run the Comet MCP server using npx (no local install required):

   Start the MCP server (runs on the default port configured by the package)

   npx -y @metrostar/comet-mcp

3. Optional: If you prefer a local install, you can install the package first and then run it as a script provided by the package: npm install --save-dev @metrostar/comet-mcp npx @metrostar/comet-mcp

4. Configure environment variables as needed (see Additional Notes) and start your development workflow to begin leveraging MCP capabilities.

Tips and common considerations:

• If the MCP server requires a specific port or host, check the package documentation or output logs for configuration options and set environment variables accordingly.
• Ensure your editor or IDE can reach the MCP server endpoints (consider CORS and network access if running locally in a container).
• Typical environment variables may include API keys or internal feature flags; placeholder examples: VAR_MCP_API_KEY, MCP_PORT, MCP_HOST. Only set what your deployment requires.
• When upgrading the MCP package, verify compatibility with your Comet version and any dependent packages to avoid breaking changes.
• If you run into issues, try restarting the MCP server and clearing any cached connections in your editor integrations.