skills.rest
IkeaNorweegway
IkeaNorweegwayCOMMUNITY
Explore all skills by this author →
GitHub

critique

Evaluate interface designs with heuristic scoring, persona testing, and automated anti-pattern detection.

Updated Apr 16, 2026
Search Tags:#frontend#heuristic evaluation#cognitive load#ux review#design critique#persona testing#anti-pattern detection
One-click install
npx skills add https://github.com/IkeaNorweegway/Classroom-Tools --skill critique-ikeanorweegway
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: critique
Source: https://github.com/IkeaNorweegway/Classroom-Tools/tree/main/.claude/skills/critique
Command: npx skills add https://github.com/IkeaNorweegway/Classroom-Tools --skill critique-ikeanorweegway

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires impeccable, and includes references (resource) components.

What problem does it solve? Design reviews are often subjective and inconsistent, missing AI-generated design tells, cognitive overload, and usability failures. This Skill provides a structured, dual-assessment critique that combines an LLM design review with deterministic automated detection to produce honest, actionable feedback. ## Core Features & Use Cases - Dual Independent Assessments: Runs an LLM design review (visual hierarchy, AI slop detection, cognitive load, emotional journey) alongside a deterministic CLI/browser detector that flags 25 anti-patterns, then synthesizes both into one report. - Quantitative Scoring: Scores all 10 Nielsen heuristics (0-4 each, 40 total) with rating bands, plus P0-P3 severity tagging for every issue found. - Persona-Based Red Flags: Tests the interface through 2-3 relevant personas (power user, first-timer, accessibility-dependent, stress tester, mobile user) and reports specific failures per persona. - Use Case: After building a new dashboard page, run the critique to get a Design Health Score, a list of prioritized issues with concrete fixes, and a recommended sequence of follow-up commands to address them. ## Quick Start Ask the assistant to critique the design of your checkout page component and provide a scored report with prioritized fixes.

Frequently Asked Questions about critique

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get a UX design review of my web page?▼

Run the critique skill against your page or component files. It launches two independent assessments—an LLM design review and an automated detector—then merges them into a scored report with prioritized issues and concrete fixes.

How to detect AI-generated design patterns in frontend code?▼

The skill runs npx impeccable against your markup files to deterministically flag 25 AI slop tells such as gradient text, glassmorphism, and generic card grids. The LLM review separately judges overall aesthetic sameness, and both verdicts are reported together.

What file types does the design critique support?▼

The CLI scan accepts HTML, JSX, TSX, Vue, and Svelte files or directories containing markup. CSS-only files should not be passed. For URLs, browser automation with the live detection server is used instead of the CLI.

Does the critique work on large codebases?▼

Yes, with limits. For directories with 200 or more scannable files, use the --fast flag for regex-only scanning. For 500 or more files, you should narrow the scope or confirm the target with the user first.

How is the design quality score calculated?▼

Each of Nielsen's 10 usability heuristics is scored 0-4 based on defined criteria, for a total out of 40. Scores map to rating bands: 36-40 excellent, 28-35 good, 20-27 acceptable, 12-19 poor, and 0-11 critical.

Why does the critique use two separate assessments?▼

Independent assessments prevent bias: the LLM review catches holistic and emotional issues while the deterministic detector catches specific patterns the LLM might miss. The final report notes where they agree, diverge, and flags detector false positives.

Related Skills

Explore More Design & Creative Skills→
mikaelcharbonneaumikaelcharbonneau

brand-context

Captures and stores core brand identity, audience, and positioning in a structured Markdown file.

Community
Basic
AcidicSoilAcidicSoil
1

apply-qa-fixes

Apply prioritized QA fixes and update the task record.

Community
Advanced
fusenginefusengine
22

validating-accessibility

Validate digital interfaces against WCAG 2.2 Level AA accessibility standards.

Community
Intermediate
oakossoakoss
14

usability-tester

Plan and conduct usability tests with heuristic evaluations and accessibility audits.

Official
Advanced