What problem does it solve? Broad quality requests like "make it feel better" or "polish this" often end with code that is merely implemented rather than genuinely improved. This Skill turns subjective quality goals into an evidence-backed loop of observing the real product surface, making targeted changes, and re-verifying the result. ## Core Features & Use Cases - Acceptance Matrix Conversion: Translates vague terms like "good" or "polished" into concrete, testable acceptance criteria before any code changes. - Observe-Change-Verify Loop: Ranks problems by user impact, implements the smallest high-impact slice, then re-observes the actual browser, game view, or device to confirm improvement. - Review Lenses Reference: Ships a curated checklist covering visual hierarchy, game feel, product flows, and writing quality to guide systematic inspection. - Use Case: A user asks to polish a game's first-time experience. The Skill captures baseline gameplay evidence, fixes the highest-impact friction point, re-plays the opening sequence, and repeats until the acceptance criteria are met. ## Quick Start Use iterate-to-evidence to inspect my web app, improve the roughest part of the onboarding flow, and keep verifying with screenshots until it genuinely feels polished.
