All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
0.1.0 - 2026-03-22
- Initial release of autooptimise
- Karpathy-style autonomous benchmark-driven skill optimisation loop
- 10-task benchmark suite targeting the OpenClaw weather skill
- LLM judge scoring rubric (0–10 scale, four criteria)
- Autonomous loop agent instructions (
runner/run_experiment.md) - Human approval gate — no changes auto-applied in v0.1
- Experiment log (
runner/experiment_log.md, gitignored) - Maximum 3 iterations per run
- Decision rule: keep change if mean score improves by ≥0.5 points
- MIT licence