mcp-server-python-docs is a read-only MCP server for the official Python
standard library documentation. It is built for end users who want precise,
version-aware stdlib answers inside MCP clients such as Claude, Cursor, and
Codex without relying on an external hosted docs API at query time.
The repo's public credibility matters. Prefer changes that make the project easier to trust, easier to verify, and easier to contribute to over changes that merely add more AI or MCP setup.
If uv is not installed yet:
python -m pip install uvBootstrap the repo:
uv sync --devIf uv is not on your PATH after installation, reopen the shell or use
python -m uv ... as a fallback.
Core verification commands:
uv run ruff check src/ tests/
uv run pyright src/
uv run pytest --tb=short -qBuild and inspect a local docs index:
uv run mcp-server-python-docs build-index --versions 3.12,3.13
uv run mcp-server-python-docs doctor
uv run mcp-server-python-docs validate-corpusPackage smoke check:
uv buildUse this order when validating MCP behavior:
- Run the automated checks above.
- Use MCP Inspector for quick local iteration.
- Confirm real-client behavior with the runbook in
.github/INTEGRATION-TEST.md.
Client-facing integration and release runbooks live here:
.github/INTEGRATION-TEST.md.github/RELEASE.md
Before calling work complete:
- relevant lint, typecheck, and test commands have been run fresh
- user-facing docs reflect the current behavior
- MCP-related changes still work in the documented validation flow
- no runtime API/tool surface changes were made unless explicitly requested
- Use official documentation first for MCP, OpenAI/Codex, and Python SDK questions.
- Avoid MCP sprawl. Do not add new MCP servers unless they clearly improve this project's development or user experience.
- Do not add repo-local custom skills by default. Add one only if a repeated workflow is genuinely painful and no strong public pattern already covers it.
- Follow existing test and documentation patterns before inventing new structure.
Treat .planning/ as archival project history, not live repo truth.
Start with these files instead:
README.mdCONTRIBUTING.md.github/INTEGRATION-TEST.mdtests/
The generated planning files may still be useful for maintainers who want the old GSD workflow context, but they should not drive routine implementation decisions.