tkm:auto-research

Automate code optimization by iteratively applying changes and measuring metrics like test coverage or performance. This Skill uses machine learning to predict impacts and git history to prevent regressions via automatic rollback.

Yes, machine learning predicts the impact of code changes. The Skill leverages scikit-learn and numpy to analyze iterative testing results, helping target measurable metrics like bundle size and performance improvements.

Integrate git history for iterative testing by tracking changes with gitpython. The Skill uses this history to learn from previous runs and automatically rolls back changes if regressions are detected during code optimization.

You need Python, a git repository with a clean working tree, and dependencies like pandas, scikit-learn, and gitpython. You must also define a goal, scope, and verification command to measure the target metric.

Avoid using machine learning for code optimization if your repository does not have a clean working tree or lacks measurable verification commands. The Skill requires git history and quantifiable metrics like bundle size to function correctly.