This MCP server is a Python-based Coreshub MCP implementation. It exposes a set of plugins and tools under src/coreshub_mcp_server/plugins that can be used to interact with Coreshub resources and perform tasks such as listing and operating on containers, filesystems, and other cloud resources via MCP prompts and tools. To operate it, run the server with the uv runtime and leverage the available plugins and prompts to perform actions. Authentication for Coreshub services is configured via environment variables (QY_ACCESS_KEY_ID, QY_SECRET_ACCESS_KEY, CORESHUB_USER_ID) and can be adjusted in your environment or in the provided MCP config. You can inspect and list plugins, execute defined tools, and use prompts to drive complex workflows through the MCP interface.

Prerequisites:

• Python 3.8+ installed on your system
• Git installed
• uv runtime available (used to run the MCP server)

Step-by-step:

1. Clone the repository or download the source: git clone https://github.com/your-org/your-repo.git cd your-repo

2. (Optional) Create a virtual environment and activate it: python -m venv venv

   On Windows

   venv\Scripts\activate

   On macOS/Linux

   source venv/bin/activate

3. Install dependencies (if a requirements.txt or setup exists): pip install -r requirements.txt

   If no requirements.txt is provided, ensure Python can import the project modules

4. Ensure you have the uv runtime available. If needed, install or set up uv as described by your environment (the project uses uv run to start the server).

5. Run the MCP server (example for development): uv run src/coreshub_mcp_server

6. Optional: configure MCP in your editor/IDE (e.g., Cherry Studio) using the provided mcpServers configuration in the README, adjusting registryUrl, command, args, and env vars as needed.

Tips and notes:

• The server relies on environment variables for authentication: QY_ACCESS_KEY_ID, QY_SECRET_ACCESS_KEY, and CORESHUB_USER_ID. Set these in your environment or in the MCP config environment block.
• The available tools and prompts live under src/coreshub_mcp_server/plugins. You can add new tools or prompts by creating a Python class derived from BaseTool or BasePrompt and registering it.
• If you experience issues with plugin loading, ensure your plugin files are in the correct directory and properly importable. Check for Python path errors or syntax issues in your plugin modules.
• When integrating with Cherry Studio or other clients, ensure the MCP type is stdio and that you provide the correct command and arguments for uvx or the runtime you are using.
• The repository exposes a range of predefined tools (e.g., operations related to epfs filesystems, container info, SSH info, distributed training tasks, etc.). Review the README sections for exact tool names and required arguments to construct correct prompts or tool invocations.