What problem does it solve? When users correct an AI agent's implementation or design choices, that knowledge is often lost or dumped into ever-growing instruction files. This Skill ensures each correction is classified and stored once in the right place—tests, design docs, instructions, or a scoped learning inbox—without duplicating rules across artifacts. ## Core Features & Use Cases - Feedback classification: Routes corrections to the right destination using a decision table covering regression tests, design documents, instruction files, learning inboxes, and skills. - Learning inbox management: Maintains scoped .github/learnings/<area>.md files with a fixed entry format, a ten-topic and 8 KB limit, and mandatory compaction before every write. - Garbage collection: Promotes mature guidance into authoritative documents, merges overlapping entries, and removes obsolete or contradicted learnings during periodic reviews. - Use Case: A user rejects an agent's error-handling pattern in a subsystem. The Skill checks existing rules, generalizes the principle, and records it once in the relevant design document or learning inbox instead of appending it to every instruction file. ## Quick Start When I correct your approach on a task, classify my feedback and record it in the appropriate repository artifact following the feedback-learning workflow.
