脚本
Autonomous Execution
PROACTIVELY USE THIS SKILL when the user provides a PRD, spec, requirements doc, or tasklist and wants end-to-end execution with minimal check-ins. Trigger on: "autonomous mode", "own this", "own this end-to-end", "enact this PRD", "implement all of this", "here are the requirements", "no check-ins unless blocked", or similar. Treat the provided PRD or tasklist as the mandate. Execute through implementation, verification, profiling, docs, and landing when requested. Only interrupt for true external blockers such as missing secrets, required sandbox or network approval, destructive irreversible actions not already authorized, or contradictory requirements.