Plan Execution Guardrails

The Following Plans skill serves as a critical governance layer for AI agents, providing a structured decision tree to determine when to proceed with a task or halt for human intervention. By distinguishing between minor syntax corrections and significant approach changes, it prevents agents from 'rationalizing' simpler but incorrect solutions that bypass established design patterns. This skill is essential for maintaining architectural integrity in complex workflows, ensuring that every implementation detail aligns with the original technical specification and reporting status with mandatory 'OK' or 'STOPPED' flags.

주요 기능

사용 사례