What problem does it solve?
Systematic design evaluation of web UIs using Nielsen's heuristics, Gestalt principles, and cognitive walkthrough. Dispatches sub-agents as independent evaluators, aggregates severity-rated findings, runs fresh-eyes sessions. Use when reviewing a UI's design quality, after visual changes, or when the user invokes /design-review. Not for functional QA (use qa-web) or code review.
Core Features & Use Cases
- Phase 1 — Cognitive Walkthrough: assess primary tasks by stepping through the happy path and documenting issues at each stage.
- Phase 2 — Heuristic Evaluation (Nielsen's 10): dispatch independent evaluators to check all heuristics with structured severity notes.
- Phase 3 — Visual Evaluation (Gestalt + Norman): evaluate perceptual grouping, affordances, feedback, and state coverage.
- Phase 4 — Fresh Eyes: spawn a context-free evaluator to test the UI from a naive perspective.
- Phase 5 — Aggregate & Prioritize: de-duplicate findings, map to heuristics, assign severities, and prioritize fixes.
Quick Start
Provide a UI prototype or screen set to start a full design-review evaluation pass.