UX Heuristics
Usability evaluation and principles
Evaluate and improve usability using Nielsen's heuristics and Krug's "Don't Make Me Think" principles. This skill gives your AI agent a systematic framework for finding and fixing UX problems before users encounter them.
Don’t Make Me Think
npx skills add wondelai/skills/ux-heuristics --global This skill is compatible with Claude, Claude Code, Claude Cowork, Codex, Cursor, OpenClaw, Hermes Agent, and other agentskills.io-compatible agents.
What is the UX Heuristics skill?
UX Heuristics folds two usability traditions into one skill: Jakob Nielsen’s ten heuristics and Steve Krug’s Don’t Make Me Think. It has your agent walk an interface the way an evaluator does — naming each violation, rating its severity from 0 to 4, and proposing the specific fix — instead of offering a general opinion on whether a screen looks usable.
Key concepts from Don’t Make Me Think your agent applies
Visibility of System Status
Always keep users informed about what's going on through timely feedback.
Recognition over Recall
Minimize memory load by making options, actions, and information visible or easily retrievable.
Error Prevention
Design to prevent problems from occurring rather than relying on error messages after the fact.
Consistency & Standards
Follow platform conventions and maintain internal consistency throughout the interface.
The Trunk Test
Every page should clearly answer: Where am I? What are my options? How did I get here?
When to use UX Heuristics — and when not to
Reach for it when
- Support answers the same “where do I find…” question every week and you suspect the navigation.
- Your signup form is abandoned halfway and you need to know which fields cause it.
- A page looks beautiful and a first-time visitor can’t tell what the site is or where to go.
- You need an audit that produces a ranked issue list, not a design opinion.
Reach for something else when
- The interface is usable and simply looks unprofessional — Refactoring UI handles spacing, color, and hierarchy.
- You want more signups from the same traffic — CRO Methodology tests for revenue, not comprehension.
- You need real reasons from real users instead of an expert walkthrough — start with The Mom Test.
- It is an iOS app and the issue is convention rather than clarity — iOS HIG Design knows what users expect.
Example prompts for the UX Heuristics skill
Run a heuristic evaluation on my checkout flow using ux-heuristics skill
UX auditIdentify usability issues on this dashboard page using ux-heuristics skill
Usability reviewSimplify this navigation so users don't have to think using ux-heuristics skill
Information architectureAdd better error prevention to my form validation using ux-heuristics skill
Error handlingUX Heuristics vs the other Design & UX skills
All 9 Design & UX skills are free and install the same way — the only question is which framework fits the job in front of you.
| Skill | Based on | Best for |
|---|---|---|
| Refactoring UI | Refactoring UI — Adam Wathan & Steve Schoger | Practical UI design system |
| iOS HIG Design | Human Interface Guidelines — Apple | Native iOS app design guidelines |
| UX Heuristics | Don’t Make Me Think — Steve Krug & Jakob Nielsen | Usability evaluation and principles |
| Web Typography | On Web Typography — Jason Santa Maria | Web typography principles and implementation |
| Design of Everyday Things | The Design of Everyday Things — Don Norman | Foundational design principles: affordances, signifiers, feedback |
| Top Design | Elite Agency Techniques — Locomotive, Studio Freight, AREA 17 | Award-winning 10/10 web design |
| Lean UX | Lean UX — Jeff Gothelf & Josh Seiden | Collaborative, experiment-driven UX design |
| Microinteractions | Microinteractions — Dan Saffer | Designing details that delight users |
| Steve Jobs Design Review | Steve Jobs — Walter Isaacson | Jobs-style reviews: ruthless simplicity, focus, binary verdicts |
Frequently asked questions
How do I install the UX Heuristics skill?
Run npx skills add wondelai/skills/ux-heuristics --global. It takes about 30 seconds and needs no account. The UX Heuristics skill then works in Claude, Claude Code, Claude Cowork, Codex, Cursor, OpenClaw and Hermes Agent — anything that reads the open agentskills.io format — and your agent loads it on its own when a task calls for it. It is free and MIT-licensed, and the source is at https://github.com/wondelai/skills.
Which book is the UX Heuristics skill based on?
It packages Don’t Make Me Think by Steve Krug & Jakob Nielsen — pioneering usability experts and co-founders of modern UX. The skill distils the book's method into instructions your agent follows while it works, covering visibility of system status, recognition over recall and error prevention. It sits in the Design & UX part of the library.
What do Nielsen’s heuristics do that Krug’s laws don’t?
They do different halves of the job. Nielsen’s ten are an inspection checklist — system status, user control, error prevention, recognition over recall — designed so a reviewer covers the whole surface instead of noticing whatever catches their eye. Krug supplies the standard those violations are judged against: people scan rather than read, satisfice rather than optimise, and muddle through rather than learn. The skill enumerates with Nielsen and sets the bar with Krug.
Can a heuristic evaluation replace user testing?
No, and treating it as a substitute is the most common misuse. Expert inspection finds violations of known principles; it cannot tell you whether people want the feature, what they expected instead, or why they gave up. It was designed as a cheap pre-filter — clear the obvious problems before you spend a session on them. Krug himself argued for small, frequent tests with a handful of real users alongside it.
How is this different from the Design of Everyday Things skill?
Scope and output. Norman gives you a theory of why interaction fails at all, applicable to a stove as much as a checkout. This skill is narrower and more procedural: it walks your actual screens, tags each issue against a named heuristic, scores severity, and hands back a triage list. Use Norman when you are arguing about a design; use this when you need something a team can work through on Monday.
Does it cover accessibility?
Partly, and you should know where the line is. It flags the usability-adjacent things — unlabelled controls, low contrast, tap targets too small, focus and status that only exist visually, and dark patterns that pressure users. It is not a WCAG conformance audit and will not check your ARIA semantics or run assistive-technology testing. Treat what it finds as a floor, not a compliance report.
Install UX Heuristics
Free, open-source, and ready in 30 seconds.
npx skills add wondelai/skills/ux-heuristics --global MIT Licensed · Works with Claude, Claude Code, Claude Cowork, Codex, Cursor, OpenClaw, Hermes Agent & other agentskills.io agents · No account needed
Don’t guess your AI engineering level.
Measure it.
AI Developer Scorecard
How production-grade is your AI engineering?
Twenty-five questions across the practices that separate vibe-coding from production-grade engineering. Instant score, per-section breakdown, and a 30/60/90-day playbook.
Score 0–75 · 30/60/90-day playbook Score yourself For CTOs & foundersCTO Scorecard
Is your engineering team ready for AI at scale?
Twenty-five questions on how your org adopts AI — adoption, governance, automation, and ROI. See where you sit on the path from Reactive to Strategic Leader.
Score 0–75 · Reactive → Strategic Leader Score your orgWe build the skills you already use. Now we’ll build yours.
Sprints from $3K · shipped to production, or you don’t pay the final milestone.