ck:loop

Automating iterative optimization for code metrics requires executing code changes, verifying a numeric output, and keeping or discarding the results. This Skill runs N iterations automatically to improve measurable metrics like test coverage and bundle size.

Before running automated code optimization loops, you need a git repository with a completely clean working tree. You also need a verify command that outputs a single numeric value to evaluate the optimization results.

The rollback mechanism works by using git tracking to evaluate each iteration's results against the baseline metric. If a code change fails to improve the metric, the Skill discards the changes and reverts to the previous state.

Yes, you can use this to automatically increase test coverage in a specific directory. You point the loop at your target directory and provide a coverage command that outputs a numeric value for the iterations to track.

A key limitation is that the verify command must output a single numeric value, meaning complex, multi-variable performance metrics are difficult to evaluate. Additionally, the working git tree must be clean before starting.

Automated metric optimization surpasses manual code refactoring for measurable targets like ESLint errors or bundle size by removing human intervention. It systematically tests iterations and automatically discards regressions using git tracking.