What problem does it solve? Before a design is shown to stakeholders or delivered, subtle issues in accessibility, visual hierarchy, interaction states, and generic AI-generated styling often slip through. This Skill runs a final quality gate that catches these problems systematically instead of relying on a single ad-hoc review pass. ## Core Features & Use Cases - Parallel Multi-Agent Review: Launches four focused reviewers covering accessibility (WCAG contrast, semantics, keyboard navigation), AI slop detection (clichéd gradients, default fonts, decorative emoji), hierarchy and rhythm (spacing scales, 5-second test), and interaction states (hover, focus, disabled, loading, reduced motion). - Deduplicated Prioritization: Merges overlapping findings and sorts them into blocker, quality, and polish tiers so teams fix what matters first. - Report-First, Fix-on-Request: Outputs a quality gate report by default and only modifies files when the user explicitly asks for fixes, then re-verifies high-risk areas like contrast changes and focus rings. - Use Case: After finishing an HTML landing page, run a final check before stakeholder review to catch a missing focus ring, low-contrast secondary text, and an overused default font in one pass. ## Quick Start Ask the agent to run a polish pass on the HTML page you just finished before delivering it.
