What problem does it solve? It removes the need for a human to supervise every intermediate build and fix cycle by running governed build, challenge, debug, and verification loops until machine-checkable gates pass, then handing the exact candidate to the human for UI acceptance. ## Core Features & Use Cases - Bounded Build Envelope: Defines objective, protected state, allowed surfaces, automated gates, recovery path, and STOP conditions before any mutation. - Multi-Role Spawn Challenge: Separates Builder, Design, and Governor lanes to challenge implementation, UX, authority, and evidence quality. - Failure Memory and Anti-Loop Discipline: Tracks failure fingerprints, attempt counts, and ruled-out causes so repeated failed paths are never retried unchanged. - Use Case: A UI bug keeps reappearing after fixes. The Skill loads known-error history, builds the smallest correction, runs automated regression gates, and only then presents a USER UI TEST handoff for the human to confirm. ## Quick Start Ask the agent to keep building and fixing the current UI problem using the build-spawn-debug loop until automated gates pass and a user UI test is ready.