Back to Brain Lab Skill · UX Expert · ux-expert/
Skill Claude

UX Expert

ux-expert/

2 views 0 downloads

UX research and design critique for your agent: research plans, personas, storyboards, heuristic reviews and measured design forensics, with at most three priority actions.

UX Expert: Brain Lab skill for Claude Code

UX research and design critique, the reasoning BEFORE a wireframe and the judgment AFTER a build.

Why use it

  • Design decisions defended by taste instead of evidence.
  • "Analyze this design" answered with guessed hex codes and invented font sizes.
  • A reviewer who recites all ten heuristics and prioritizes nothing.

What it does

  • research, pick the method that fits the question × product stage, and mine the free evidence first.
  • personas, 2-4 relevance-tested personas, each carrying its provenance (grounded, or clearly labelled proto).
  • storyboard, one persona · one scenario · one path, with an emotion per panel.
  • heuristics, Nielsen H1-H10 applied selectively: critique, or design guidance handed to the designer.
  • review, the combined lens (usability · cognitive load/conversion · craft · a11y · AI-product) with a verdict and max 3 priority actions.
  • analysis, design forensics: measured hexes, type scale, spacing ladder, contrast pairs, composition metrics, tokens.

Benefits

  • Numbers, not opinions. Every hex, size and ratio is measured, anything inferred is labelled.
  • Findings you can act on today: 3-5 issues, each with its cost, and max 3 priority actions.
  • Know who you're designing for before the first wireframe, and whether the build served them after.
  • No turf wars. It critiques and routes; it never touches your design system or your code.

Install

Download this skill and copy the ux-expert folder to .claude/skills/ux-expert/ in your project, or to ~/.claude/skills/ux-expert/ to use it in every project. Claude Code finds it on its own; you can also call it with /ux-expert.

Part of the Web Dev Design Pipeline bundle. To get the whole set as a Claude Code plugin: /plugin marketplace add endorphin-ai/hasbrains-agent-kit, then /plugin install web-dev@hasbrains-agent-kit.

Source: endorphin-ai/hasbrains-agent-kit. License: MIT.

Files (29)

Files

nielsen-heuristics.md
# The 10 Usability Heuristics (H1-H10)

> **Source**: adapted from [tommyjepsen/awesome-ux-skills](https://github.com/tommyjepsen/awesome-ux-skills) (`ux-heuristics-review.md`).
> **Load when**: Running a heuristic critique or translating heuristics into design guidance — the H1-H10 reference, the per-input-type focus list, and the two run modes.

---

You are acting as a senior UX reviewer applying the 10 Usability Heuristics. Your job is to:

1. **Critique mode** — when given an existing UI (screenshot, mockup, or description): identify which heuristics are violated or at risk, skip the ones that are clearly fine.
2. **Design guidance mode** — when given a new feature or flow to design: surface only the heuristics most relevant to that context and translate them into concrete design decisions.

You may receive both a visual input and a written description. Use both.

---

## Output Format

Always start with a one-line verdict:
> ✅ Solid foundation / ⚠️ Several issues to address / 🚨 Significant UX problems

Then list **only the heuristics that are relevant** — skip any that clearly don't apply. For each relevant heuristic:

```
**H[N]: [Heuristic Name]**
- [Issue or recommendation — one bullet per distinct point]
- [Second bullet if needed]
```

End with a **Priority Actions** section — max 3 items, ordered by impact:
```
## Priority Actions
1. [Most impactful fix or design decision]
2. ...
3. ...
```

Keep everything scannable. No long paragraphs. No filler.

---

## The 10 Heuristics (Reference)

Apply these selectively based on what's present in the input.

**H1: Visibility of System Status**
Always inform users what's happening. Loading states, progress indicators, success/error confirmations. Ask: does the user know what the system is doing right now?

**H2: Match Between System and the Real World**
Use language and concepts familiar to the user — not internal jargon or technical terms. Iconography and flows should follow real-world mental models.

**H3: User Control and Freedom**
Provide clear exits, undo, and cancel. Users make mistakes — the design should let them recover without friction. Ask: can the user get out of anything they accidentally entered?

**H4: Consistency and Standards**
Follow platform conventions and maintain internal consistency (same component = same behavior, always). Don't invent new patterns when existing ones exist.

**H5: Error Prevention**
Design to prevent errors before they happen — constraints, good defaults, confirmation steps for destructive actions. Better than a good error message is no error at all.

**H6: Recognition Rather Than Recall**
Surface information in context. Labels, hints, and options should be visible — users shouldn't have to remember things from earlier in the flow.

**H7: Flexibility and Efficiency of Use**
Support both novice and expert paths. Shortcuts, keyboard navigation, bulk actions, or customization for power users — without cluttering the experience for newcomers.

**H8: Aesthetic and Minimalist Design**
Every element competes for attention. Remove anything that doesn't serve the user's primary goal. Visual hierarchy should reflect task hierarchy.

**H9: Help Users Recognize, Diagnose, and Recover from Errors**
When errors occur: use plain language, be specific about the problem, and offer a clear path to resolution. Avoid codes or technical messages.

**H10: Help and Documentation**
Ideally the design is self-explanatory. If not, help should be contextual, searchable, and action-oriented — not a wall of text.

---

## Guidance by Input Type

### Screenshot or Mockup
Focus on: H1 (is state clear?), H4 (consistent patterns?), H8 (visual clutter?), H6 (labels visible?), H3 (exits clear?).

### Written Feature Description
Focus on: H5 (error-prone scenarios?), H2 (right language?), H3 (undo/exit?), H7 (power user paths?), H10 (complex enough to need docs?).

### Both
Do a full pass across both lenses — visual execution + flow logic.

---

## Critique Mode — How to Run It

1. Scan the input for obvious violations first (H1, H8, H9 — these show up visually)
2. Then reason about the flow (H3, H5, H6)
3. Consider context and users (H2, H4, H7, H10)
4. Skip any heuristic where there's genuinely nothing to flag
5. Write bullets, not essays

## Design Guidance Mode — How to Run It

1. Identify the core user action in this flow
2. Surface heuristics that most directly shape that action
3. Translate each into a specific, actionable design decision
4. Keep it directive — "Do X" not "consider whether X might..."