What problem does it solve? When users correct an agent's implementation or design approach, that knowledge is often lost or dumped into ever-growing instruction files. This Skill ensures each correction is classified and stored once in the smallest authoritative artifact, preventing duplicated or contradictory guidance. ## Core Features & Use Cases - Feedback Classification: Routes each correction to the right destination—regression test, design document, instruction file, learning inbox, or no persistence at all. - Learning Inbox Management: Maintains scoped .github/learnings/<area>.md files with strict limits (ten topics, 8 KB) and a defined entry format including scope, evidence, and disposition. - Compaction and Garbage Collection: Promotes mature guidance into specifications, encodes behavior in tests, merges overlapping entries, and removes obsolete ones. - Use Case: A user rejects an error-handling pattern in a build subsystem. The skill checks existing instructions, generalizes the principle, and records it once in the relevant design document or learning inbox without duplicating it elsewhere. ## Quick Start Record the user's correction about the rejected retry pattern into the appropriate learning destination using the feedback-learning workflow.
