Uxcel-Lab Visual Aesthetics Audit Skill
How this skill behaves (read first)
This is an evaluative skill — it reviews existing work, it doesn't generate. It auto-runs and is built to be called as a validation step by generative skills. It's the visual-quality backbone of the repo: generators that produce UI (pricing, login-signup, dashboard, onboarding, component skills) hand off here to confirm they didn't tip into clutter or break hierarchy.
It owns visual composition. Two deliberate deferrals so it doesn't duplicate:
- Color-contrast ratios, and any accessibility concern (keyboard, screen readers, alt text) → the accessibility audit. This skill judges visual contrast/emphasis and harmony, not WCAG ratios.
- Usability, flow, and interaction logic → the heuristics audit. This skill judges how it looks, not how it works.
Scope discipline. When invoked directly (the user named this audit), review only this concern — don't pull in sibling audits. It runs alongside other lenses only when an orchestrator (ux-design-review) or a generative skill calls it under docs/orchestration-policy.md, where it sits in Tier B — offered (and suppressed when a design system already defines visual harmony). Explicit scope always wins.
What this skill changes vs. default behavior
By default, visual feedback is vague and subjective — "make it cleaner", "add more spacing", "looks a bit busy" — with no principle, no severity, and no fix. This skill forces:
- Every finding names the violated principle (hierarchy, emphasis, proximity, etc.).
- Severity rated by whether it breaks scanning/comprehension or is just polish.
- A concrete fix, not a vibe.
- An explicit overdesign pass — the failure mode generators most need caught, and the one "more best practices" tends to cause.
It is not taste-policing: flag principle violations that cost scanning, comprehension, or perceived quality — not personal style preferences.
The evaluation framework — what to check and what a violation looks like
Work through each. Flag with principle, severity, fix.
1. Visual hierarchy & eye-flow
Users should have one clear entry point and a guided path through the content (scale, weight, color, spacing, position). Hierarchy is achieved through movement, contrast, and proportion.
Flag when: no obvious place the eye lands first; the primary action isn't the most prominent element; everything is the same visual weight; reading order fights the layout.
- ❌ Three cards, three "primary" buttons, a big hero image, and a banner all competing — eye doesn't know where to go.
- ✅ One dominant focal point (the primary action / key message); everything else recedes.
2. Emphasis & contrast
Emphasis makes one element the focal point (Von Restorff). Contrast is the difference between elements — aim for more than one point of contrast (size + color, shape + position), and pair a bold accent with a neutral anchor so the eye has somewhere to rest.
Flag when: everything is emphasized (so nothing is); a single weak contrast dimension; high-saturation colors everywhere with no neutral to absorb tension ("Willy Wonka" palette).
- ❌ Every section uses a bright accent color → visual noise, no focal point.
- ✅ Mostly neutral surface with one accent reserved for the key action.
3. Spacing & negative space
White space is breathing room, not waste. Passive space separates within groups; active space deliberately spotlights an element. You don't notice passive space "until there's too little."
Flag when: cramped, content jammed edge-to-edge; uneven/inconsistent gaps between similar elements; or the opposite — so much empty space the content feels disconnected and aimless.
- ❌ Form fields and labels packed with no breathing room → feels stressful, hard to scan.
- ✅ Consistent, generous spacing that groups related items and gives the focal point room.
4. Alignment & grid
Elements should sit on a common axis / shared grid. Alignment creates a structured, professional, calm impression.
Flag when: ragged edges, elements off the grid, inconsistent margins/gutters, headers that don't align to the column structure.
- ❌ Labels left-aligned, inputs centered, button floating off-axis.
- ✅ One consistent alignment system; intentional breaks only.
5. Grouping & proximity
Related items sit close; unrelated items are separated (Gestalt proximity). Similarity (shared shape/size/color) unifies items of equal importance.
Flag when: related items are far apart or divided by a border while unrelated items crowd together; equal-importance items (e.g. a row of cards) are styled inconsistently.
- ❌ A field's helper text sits closer to the next field than its own.
- ✅ Spacing encodes the grouping; a row of cards shares one consistent treatment.
6. Consistency & repetition
Repeating elements (color, shape, type, spacing) creates unity and a learnable rhythm. The color system should be deliberate — one blue, not three near-identical ones.
Flag when: the same component styled differently across the screen; near-duplicate colors/radii/shadows that should be one token; buttons of the same importance looking different.
- ❌ Three slightly different grays and two button radii with no reason.
- ✅ A small, consistent set of styles applied repeatably.
7. Typographic hierarchy
Build hierarchy with size, weight, style, color, case — but use no more than ~3 of those instruments, keep to 2–3 heading levels, and body text ≥16px. Color in type used sparingly.
Flag when: a heading stacks big + bold + colored + uppercase + italic all at once; more than ~3 heading levels; tiny body text; color applied to too much type.
- ❌ Headline that is huge, bold, red, AND all-caps → cluttered, shouting.
- ✅ Size + weight alone separates levels cleanly; color reserved for one accent.
8. Depth & elevation
Shadows signal elevation and draw the eye to lifted surfaces (cards, modals, FABs). A single, consistent light direction and a defined elevation scale keep depth readable.
Flag when: shadows fall in different directions on the same screen; every element has a shadow (so elevation means nothing); shadows stacked on a card and its inner buttons/icons.
- ❌ Card has a shadow, its button has a shadow, its icon has a shadow → muddy, no real hierarchy.
- ✅ One elevation level per surface; shadows share a light direction.
The overdesign pass (cross-cutting). After the eight checks, step back: are there multiple competing focal points? More than a couple of accent colors? Stacked shadows? >3 type treatments? Decorative elements that don't serve the task? If the screen feels busy, name which principles are over-applied — this is the specific signal generators rely on.
Severity rating
| Severity | Meaning |
|---|---|
| Critical | Hierarchy collapses — no discernible focal point / the primary action doesn't stand out, or clutter is severe enough that the user can't tell what matters. The visual design fails its job. |
| Major | Competing emphasis, inconsistent component styling, cramped or uneven spacing, misalignment, >3 type instruments, clashing/stacked shadows — slows scanning and erodes perceived quality. |
| Minor | Polish — slightly off proportions, one small inconsistency, a subtle-contrast element that reads as a near-mistake. |
Rate by impact on scanning, comprehension, and perceived quality — not by how subjective the fix feels. Note the aesthetic-usability effect: a polished look earns tolerance for minor usability issues, but never let "it looks nice" excuse a broken hierarchy.
Audit output format
## Visual Aesthetics Audit — [Screen / Component]
Context: [platform · primary task / focal action, if known]
### Critical
- [Issue] — what's wrong → fix. (Principle: …)
### Major
- …
### Minor
- …
### Working well ✓
- …
**Overdesign verdict:** [clean / borderline / overdesigned] — the 1–3 over-applied principles to pull back.
**Top 3 priorities:** the highest-impact fixes.
End with the overdesign verdict and the prioritized top 3 — that's what a calling generator acts on.
How to run a good audit
- Establish context — platform and the screen's primary task / focal action. Hierarchy can only be judged against what's supposed to be most important.
- Serve the task, not decoration. Every element should earn its visual weight; flag decoration that competes with content.
- Don't taste-police. Flag principle violations that cost scanning/comprehension/quality, not style you'd personally choose differently.
- Run the overdesign pass explicitly — it's the highest-value output and the reason generators call this skill.
- Defer contrast-ratio/accessibility to the accessibility audit and usability/flow to the heuristics audit; don't re-derive them here.
Common do/don't patterns
| ❌ Don't | ✅ Do |
|---|---|
| "Make it cleaner / less busy" | Name the over-applied principle + the specific fix |
| Treat every nit as equal | Rate by impact on hierarchy/scanning; surface top 3 |
| Emphasize everything | One focal point; everything else recedes |
| One weak contrast dimension | More than one point of contrast + a neutral anchor |
| Accent color everywhere | Mostly neutral; accent reserved for the key action |
| Stack big+bold+color+caps on a heading | ≤3 type instruments; 2–3 heading levels; body ≥16px |
| Shadow on the card and its inner button and icon | One elevation level per surface; consistent light direction |
| Re-derive WCAG contrast or keyboard issues | Defer to the accessibility audit |
| Flag subjective style preferences | Flag only principle violations that cost the user |