critique

Run the critique skill against the page or component you want reviewed. It evaluates ten dimensions including visual hierarchy, cognitive load, and accessibility, then returns a scored report with prioritized issues and concrete fix suggestions.

It is a 0-40 score computed by rating each of Nielsen's 10 usability heuristics from 0 to 4. Scores of 28-35 indicate a good foundation, 20-27 means significant improvements are needed, and below 20 signals major UX problems.

The review selects 2-3 personas relevant to the interface type, such as a power user or first-timer, and walks the primary user action as each one. It reports specific red flags like missing keyboard shortcuts or jargon-heavy error messages rather than generic concerns.

Yes. The critique includes a dedicated AI slop detection pass that checks for common tells of AI-generated design, such as gradient text, glassmorphism, glowing dark-mode accents, and identical card grids, and reports them as a pass/fail verdict.

The critique relies on heuristic evaluation and persona simulation rather than real user testing, so it cannot measure actual task completion rates or emotional responses. It works best as a fast first-pass review before usability testing with real users.