loop-engineering-skills

Make AI output measurably smarter using research-backed self-correction loops. Use this skill whenever the user asks to improve, refine, double-check, verify, or "make better" any AI-generated output — code, writing, plans, analysis, answers — or asks for "your best work", "high quality", "production ready", "make sure this is right", or gives a complex/high-stakes task where a first draft is unlikely to be good enough. Also use when the user mentions loop engineering, iterative refinement, self-review, or self-correction. Do NOT loop on simple factual lookups or short casual answers.

Give the agent episodic memory so it improves every cycle instead of repeating mistakes. Use this skill whenever working in a project that has a .loop/LESSONS.md file, whenever a task is being retried after a failure, whenever the user says "remember this", "don't make that mistake again", "learn from this", or "why do you keep doing that", and at the start and end of every task run inside a loop. Also use it to compress or maintain the lessons file when it grows long.

Verify agent work against project gates before declaring it done. Use this skill whenever working in a project with a .loop/GATES.md file, whenever finishing any coding task inside a loop, whenever the user asks "is this actually done", "verify this", "check your work", or before emitting any VERDICT line. Also use it when a task involves tests, refactors, bug fixes, or dependency updates, when deciding whether work should be escalated to a human, and when the user asks to "audit my loop", "check loop readiness", or evaluate their loop setup.