Contributing¶

Reporting issues¶

Open an issue on GitHub. Include:

• What you expected to happen
• What actually happened
• Steps to reproduce
• Relevant logs (Cloud Run logs, mcp-remote output)

Making changes¶

1. Fork the repo and create a feature branch.
2. Make your changes — see Development for the local setup.
3. Run hatch run check to ensure lint, types, and tests all pass.
4. Open a pull request against main.

Code standards¶

• Python 3.12, typed with mypy --strict
• Ruff for linting and formatting (line-length = 100)
• Google-style docstrings on public classes and methods
• Tests live in tests/ and use pytest with pytest-asyncio

The CI pipeline enforces all of these on every push.

Adding a new MCP tool¶

1. Define the function in src/the_curator/main.py decorated with @mcp.tool().
2. Add a docstring — it becomes the tool description visible to MCP clients.
3. Wire any new dependencies through PodcastGeneration or VertexClient rather than adding ad-hoc SDK calls in main.py.
4. Add a test in tests/test_main.py.
5. Document the new tool in API Reference → Server.