ck:autoresearch

To automate iterative optimization, you can run an autonomous loop that applies atomic changes, evaluates metric impacts, and uses git revert to discard non-productive iterations. Results are logged to loop-results.tsv for traceable, repeatable software experiments.

Git-based memory records each iteration's outcome in loop-results.tsv and uses git revert to automatically back out non-productive changes. This enables safe exploration by learning from history while ensuring only beneficial modifications are kept in the project.

Yes, autonomous iteration loops are suitable for improving metrics like test coverage, bundle size, lint errors, or performance benchmarks. The loop repeatedly applies changes and evaluates their impact across frontend, backend, or tooling projects to push measurable improvements.

The best way to ensure safe changes is to use Verify and optional Guard commands within the optimization loop. These enforce strict rules, automatically triggering rework or discarding changes via git revert if the iteration does not meet the defined metric or guard criteria.

No, autonomous iterative optimization supports configurable iterations, allowing you to define the exact number of cycles to run. Each iteration applies a single atomic change and evaluates its impact, following a strict 8-phase protocol to ensure structured and safe execution.