What problem does it solve? Approved plans often get executed with silent deviations, unspoken judgment calls, and unverified outcomes. This Skill enforces disciplined execution: follow the plan exactly, stop on drift, surface every judgment call, and end with an explicit completion declaration. ## Core Features & Use Cases - Plan Validation Gate: Rejects plans lacking numbered steps, concrete actions, or verifiable outcomes before any work begins. - Drift and Judgment-Call Handling: Stops immediately when reality differs from the plan or when an unspecified decision is required, asking the user instead of improvising. - Verification and Terminal Declaration: Runs the plan's verification steps plus the project's gate suite (lint/test/typecheck), maps each step to its verification, and always ends with a complete, failed, or halted declaration. - Use Case: After approving an implementation plan with /plan-create, run /execute to have the agent implement each step in order, pause for your approval on any deviation, and report exactly what was verified before committing. ## Quick Start After approving a plan in your conversation, say "execute the plan" to have the agent implement it step by step with verification.
