What problem does it solve? Agents often jump straight into coding on vague requests, causing mismatched implementations, role drift, and wasted effort. This Skill enforces progressive clarification and explicit user approval before any code changes, so the final result matches what the user actually wanted. ## Core Features & Use Cases - Risk-proportional hard gate: Low-risk work (read-only investigation, verification) proceeds immediately, while high-risk work (code changes, architecture decisions) requires an approved decision surface first. - Structured questioning rounds: Three progressive rounds covering intent and expected outcome, current reality, and constraints, followed by a 2-3 approach comparison with a recommendation. - Specialized decision surfaces: Provides templates for UI/visual drift approvals, diagram intent cards, analysis and tuning tasks, and vague correction feedback via a two-stage option flow. - Use Case: A user says "the animation feels off." Instead of guessing, the agent presents dimension options (timing, spacing, motion), pins the target state through contrast choices, and gets approval on a visible-effect card before editing any code. ## Quick Start Ask the agent to clarify the expected outcome and propose approaches before implementing my requested change.
