Skillz MCP Server

# Mission Control Plane (MCP) Server ## Project Overview This project implements a Mission Control Plane (MCP) server using FastAPI, designed to provide a robust and extensible backend for managing and interacting with various functionalities, referred to as "skills." It leverages Docker and Docker Compose for easy deployment and includes hot-reloading for efficient development. ## Features * **FastAPI Backend:** A high-performance, easy-to-use web framework for building APIs with Python 3.11. * **Dockerized Deployment:** Packaged in a `python:3.11-slim` Docker container for consistent environments. * **Docker Compose:** Simplifies the management and orchestration of the server and its dependencies. * **Hot-Reloading:** Automatic code reloading during development for a smooth workflow. * **Skills Feature:** A dynamic system allowing LLMs to progressively discover and understand tools/functionalities defined in a structured `skillz` directory. The MCP server serves as a documentation hub for these skills, enabling LLMs to interpret and execute actions based on the provided skill definitions. When implementing skills, agents should utilize the `tmp` directory for any local file generation or copying. * **Local Skill Execution**: Skills exposed by the MCP can be retrieved and executed locally, providing flexibility for client-side operations. Refer to `GEMINI.md` for detailed instructions on how to use skills locally. ## Getting Started To set up and run the MCP server, ensure you have Docker and Docker Compose installed. 1. **Clone the repository:** ```bash git clone https://github.com/pwntato/skillz_mcp cd skillz_mcp ``` 2. **Start the server:** ```bash docker compose up -d ``` This command will build the Docker image (if not already built) and start the server in detached mode. 3. **Access the API Documentation:** The server will be available at `http://localhost:8000`. You can access the interactive API documentation (Swagger UI) by navigating to `http://localhost:8000/docs` in your web browser. ## API Endpoints * `/`: Redirects to the API documentation (`/docs`). * `/skills`: Returns an `instruction` field explaining the purpose of skills for LLMs, along with a list of available skills, including their name, description, and `skill_id` (derived from the skill's directory name). * `/skills/{skill_id}/{file_path:path}`: Retrieves an `instruction` field explaining the purpose of skills for LLMs, along with the content of a specific file within a given skill's directory. This is used by LLMs for progressive loading of skill details and associated scripts. * `/skills/{skill_id}/files`: Returns an `instruction` field explaining the purpose of skills for LLMs, along with a list of all files within a given skill's directory. ## Skills Development The "skills" feature allows for dynamic extension of the MCP server's capabilities. Each skill is defined within its own directory under the `skillz/` folder. The MCP server acts as a repository for these skill definitions, which are then interpreted and executed by an LLM. * The `skillz` directory is mounted into the container at `/skillz` and can be used for storing and accessing data. * The `tmp` directory is available for generating or copying files locally during development or agent operations. Its contents are git-ignored. ## Testing Automated tests are configured using `pytest` and can be run locally or via GitHub Actions. To run tests locally (ensure you have `pytest` and `httpx` installed in your local Python environment): ```bash PYTHONPATH=. pytest ``` ## Contributing Contributions are welcome! Please refer to `GEMINI.md` for development conventions and guidelines. ## License This project is licensed under the MIT License. See the `LICENSE` file for details.