← ClaudeAtlas

ui-ux-auditlisted

Comprehensive UI/UX quality audit and prevention system that catches the failures AI consistently ships, including broken or shifting layouts, text overflow, inconsistent components, missing states, poor contrast, undersized touch targets, accessibility violations, performance jank, and generic "AI slop" aesthetics. Use whenever the user asks to audit, review, check, critique, fix, polish, optimize, or improve any UI, page, screen, component, screenshot, or design, or says things like "the layout looks off", "something feels wrong", "this looks AI-generated", "is this ready to ship", "weird spacing", "ugly", or "broken page". ALSO consult this skill silently BEFORE writing or editing any UI code (React, JSX, HTML, CSS, Tailwind, React Native, Expo) so these failures never ship in the first place. Supports per-project profiles that override the universal defaults; an example kids-app profile doubling as a template is included.
EnchStyle/ui-ux-audit-skill · ★ 2 · Web & Frontend · score 75
Install: claude install-skill EnchStyle/ui-ux-audit-skill
# UI/UX Audit Catch the UI problems AI consistently introduces, and either report them with concrete fixes (audit) or avoid them while writing code (prevention). ## Why this skill exists Almost every AI UI failure traces back to one root cause: the model generates a plausible static snapshot of a single state (usually the happy path on a desktop viewport), it cannot see the rendered result, and it carries no persistent design system across generations. The output looks right in a screenshot and breaks the moment content grows, the viewport shrinks, an error occurs, or a second component is generated with slightly different padding. A human developer carries an internalised model of "what happens when this grows, shrinks, errors, or gets tapped." This skill forces that model into view: every check below exists because AI reliably gets it wrong. ## The three modes **1. Full audit.** Triggered by explicit requests like "run a full UI audit", "audit this app/page/component", "full design review", "is this ready to ship". This is the heavyweight, systematic pass: read ALL reference files listed below, walk every category in order, and produce the scored report defined in this file. Do not skip categories because they look fine at a glance; AI failures hide in the categories that were not checked. **2. Quick review.** Triggered by narrower requests like "review this button", "fix the spacing here", "why does this look off". Identify which categories the request touches, read