What problem does it solve? Design work with AI agents often lacks consistent taste, honest evaluation, and a structured process. This Skill provides one design agent that covers the full design lifecycle—research, creation, execution, review, and improvement—with a five-dimension scoring standard and four scoring disciplines that force evidence-based, non-inflated reviews. ## Core Features & Use Cases - Five-Stage Routing: Routes user intent to one of 16 stage modules spanning design research, wireframes, design variations, interactive prototypes, five-dimension review, AI-slop detection, accessibility audits, and design improvement. - Second-Opinion Advisor: A read-only cross-stage advisor ("First Principles × Product Taste") challenges major direction choices, expensive implementations, and hard-to-reverse brand or architecture decisions, returning one fused recommendation plus engineering and experience dissents. - Honest Review Standard: Scores designs on philosophical consistency, visual hierarchy, execution detail, functionality, and originality, requiring every score to cite a file, line, element, or observable fact. - Use Case: Ask the agent to deeply review a landing page; it produces a severity-ranked findings report with per-dimension scores and cited evidence, then—only when explicitly asked—fixes the confirmed issues and re-verifies. ## Quick Start Ask the agent to run a five-dimension review of your current page or prototype and list the findings by severity.
