mcp-code-checker
MCP server providing code quality checks (pylint and pytest) with smart LLM-friendly prompts for analysis and fixes. Enables Claude and other AI assistants to analyze your code and suggest improvements.
claude mcp add --transport stdio marcusjellinghaus-mcp-code-checker python -m mcp_code_checker --project-dir /path/to/project
How to use
The MCP Code Checker server exposes a set of code quality capabilities that can be accessed by any MCP client (such as Claude Desktop, VSCode with Copilot, or Windsurf/Cursor) to inspect and improve a project's code base. It supports running pylint for static code analysis, pytest for test results, and mypy for type checking, returning structured results and optional smart prompts to help explain issues and suggest fixes. You can combine multiple checks with run_all_checks to get a comprehensive quality report, or run individual checks as needed. The server runs within a specified project directory, ensuring all tools analyze only the intended codebase and keep artifacts contained. To use it, configure the MCP client to point at the server's endpoint and invoke the available operations (e.g., run_pylint_check, run_pytest_check, run_mypy_check, or run_all_checks) with optional parameter customizations.
Typical workflow:
- Start the server for a given project directory (e.g., --project-dir /path/to/project).
- Request run_pylint_check to identify code smells and generate LLM-friendly prompts describing issues and suggested fixes.
- Optionally run run_pytest_check to see failing tests and obtain accompanying prompts to guide remediation.
- Run run_mypy_check to enforce type safety and surface type-related issues, with outputs suitable for developer feedback and automated suggestions.
- Use run_all_checks to obtain a consolidated report and combined prompts for easier review by AI assistants.
How to install
Prerequisites:\n- Python 3.8+ (recommended)\n- pip (comes with Python)\n- Git (for cloning or installing from GitHub)\n\nInstallation steps (official release):\n1) Install directly from GitHub using pip:\nbash\npip install git+https://github.com/MarcusJellinghaus/mcp-code-checker.git\n\n2) Verify installation and help:\nbash\nmcp-code-checker --help\n\n3) Run the server for a project (example):\nbash\nmcp-code-checker --project-dir /path/to/project [options]\n\n\nDevelopment installation (from source):\nbash\n# Clone the repository\ngit clone https://github.com/MarcusJellinghaus/mcp-code-checker.git\ncd mcp-code-checker\npython -m venv .venv\nsource .venv/bin/activate # On Windows: .venv\Scripts\activate\npip install -e ".[dev]"\nmcp-code-checker --help\n
Additional notes
Notes and tips:\n- This server expects a project directory and provides operations that run tools inside that directory. Ensure your project has a valid Python environment and the necessary tooling installed (pylint, pytest, mypy) if you plan to run those checks via the server.\n- You can customize tools using the CLI options described in the README: for pylint, pytest, and mypy, including target directories, extra arguments, markers, verbosity, and environment variables.\n- The server can be integrated with mcp-config for managing MCP server configurations, enabling interactive setup with Claude Desktop or VSCode.\n- When running tests or checks in parallel, be mindful of resource usage in constrained environments.\n- For local development, consider mounting your project directory or using a virtual environment to isolate dependencies.
Related MCP Servers
web-eval-agent
An MCP server that autonomously evaluates web applications.
mcp-neo4j
Neo4j Labs Model Context Protocol servers
Gitingest
mcp server for gitingest
zotero
Model Context Protocol (MCP) server for the Zotero API, in Python
fhir
FHIR MCP Server – helping you expose any FHIR Server or API as a MCP Server.
unitree-go2
The Unitree Go2 MCP Server is a server built on the MCP that enables users to control the Unitree Go2 robot using natural language commands interpreted by a LLM.