Thanks for considering contributing. This is an experimental research project, and contributions that improve the evolution system, documentation, or reproducibility are welcome.
- Add new signals: Extend
SIGNAL_POOLSinauto_evolve.pyor add keyword checks toevaluate.py - Tune mutation weights: Adjust strategy probabilities in
mutate.py(line 130-133) - Write new strategies: Add mutation strategies to
mutate.pyfollowing the existing pattern - Improve runtime evaluator: Extend
evaluator/runtime_evaluator.pywith more execution checks - Add benchmark tasks: Add entries to
benchmarks/tasks.json
- Improve README clarity and accuracy
- Add architecture decision records
- Improve code comments for non-obvious logic
- Document experiment results and observations
- Fork the repo and create a branch from
main. - Make focused, single-purpose changes.
- Test by running
python eval.pyto verify scoring still works. - If adding runtime features, run a quick validation cycle.
- Open a pull request against
main.
Open an issue using the templates in .github/ISSUE_TEMPLATE/.
- Python: ruff-compatible. Run
ruff check .before committing. - Keep files self-contained when possible.
- No exaggeration in documentation — accurately describe capabilities.
- Comments for non-obvious behavior, not for basic operations.
- Do not claim AGI, sentience, or consciousness for this system.
- Frame contributions as evolutionary computation and prompt optimization research.
- Be honest about limitations and failure modes.
Open a discussion.