Quick orientation for AI coding agents

Key facts an agent must know (short):

The runner: runner.py dynamically imports every module under ultimatepython and runs any callable named main with zero parameters. Keep main() functions parameterless, idempotent, and side-effect safe for CI.
Single-file runs are supported: modules are runnable with python ultimatepython/<category>/<module>.py. Most modules already include if __name__ == "__main__": main().
Examples use plain Python stdlib only. Avoid adding new third-party dependencies without updating requirements.txt and pyproject.toml.
Assertions are used as the primary verification mechanism inside examples (not pytest tests). Preserve their intent and messages when editing.

main() must be parameterless and not raise exceptions when run. Examples: ultimatepython/syntax/function.py, ultimatepython/classes/basic_class.py.
Module structure: docstring, helpers, main() demonstrating usage via asserts, if __name__ == "__main__": main().
Avoid long-running or destructive operations. For filesystem interactions, keep them local to ephemeral files and clean up after running.
Use ruff and isort (configured in pyproject.toml) for consistency with existing modules.

New modules go under the appropriate folder—the runner discovers them automatically.
If a change requires a new dependency, update requirements.txt and pyproject.toml.
When updating README.md, consider updating translations or note why they were skipped.

CI expects that running runner.py and executing each main() will succeed. Avoid adding code that requires network access, long sleeps, or interactive input.
Do not introduce external services or config files; this repo intentionally uses only the Python standard library.