What problem does it solve? Visual inconsistencies, spacing problems, weak hierarchy, and generic AI-looking interfaces slip into production because manual design QA is slow and subjective. This Skill performs a structured designer-grade audit of a live site, then fixes each issue in source code with before/after screenshot verification. ## Core Features & Use Cases - Design Audit with Grades: Visits 5-8 pages, extracts the actual rendered design system (fonts, colors, heading scale, touch targets), and produces a report with letter grades and a first-impression critique. - Iterative Fix Loop: Fixes each finding in source code, commits atomically, and re-verifies with before/after screenshots. - Multiple Modes: Full, quick, deep, diff-aware (feature-branch scoped), and regression modes that compare against a previous design baseline. - Use Case: Before launching a marketing site, run a deep audit to catch undersized tap targets, inconsistent heading scales, and console errors, then have each issue fixed and committed automatically. ## Quick Start Ask the AI to run a design review on your local dev server URL and fix the visual issues it finds.
