1 entries — sorted by GitHub stars.
Let OpenClaw agents evaluate their own outputs and iteratively improve responses through self-critique loops.