Review & audit

Reviews UI code against Vercel's Web Interface Guidelines, always pulling the latest rule list, and reports short findings with file and line.

Runs all of Jakub Krehel's better-* skills as one review across accessibility, layout, writing, typography, color and polish, ranked by impact.

Writes and reviews interface copy: labels, errors, empty states and confirmations in the product's own voice.

Reviews a branch or pull request for interface regressions it introduced and classifies each finding.

Stress-tests a UI with worst-case data, such as long names, missing fields and huge counts, behind a toggle, and reports what broke.

An evidence-based audit of one product screen against its own design system, with a self-contained change spec for each finding you pick.

Finds AI design clichés such as stacked decoration, repeated templates, motion for show and fake proof, and ranks what to remove.

Audits and fixes hierarchy, spacing, color, depth and typography with the Refactoring UI principles, with a score and a quick diagnostic.

Usability review with Krug's principles and Nielsen's ten heuristics, scored by severity.

Walks a live web app as a persona and reports usability problems with hard gates, verdicts and evidence.

Craft standards, a catalog of warning signs, and purpose and delivery gates that keep UI, text and code from reading as generated.

A deterministic design critique: a CLI runs 73 checks with exact thresholds on HTML and CSS, and the skill explains severity, exceptions and CI setup.

152 prioritized UI rules across animation principles, timing functions, exit animations, sound, the laws of UX and prefetching.

A structured design review of UI code by category, from hierarchy and spacing to states and motion, in a fixed output format.