frontend-design — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited frontend-design (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Premise: Nothing on the page is accidental. Space is a feature. Decisions trace back to the person using the product. Working rule: Reason it out, don't recite it. When in doubt, ask.
One file is mandatory; the rest are a reference shelf, pulled in only when the work calls for them.
| File | Priority | Open it for |
|---|---|---|
| ux-psychology.md | Read every time | The behavioral foundation behind all of this |
| color-system.md | As needed | Palette and color decisions |
| typography-system.md | As needed | Choosing and pairing type |
| visual-effects.md | As needed | Shadows, blur, gradients, glow |
| animation-guide.md | As needed | Interface motion |
| motion-graphics.md | As needed | Lottie, GSAP, SVG, 3D, particles |
| decision-trees.md | As needed | Context-driven starting points |
Run these against a project; they are tools, not reading material.
| Command | What it inspects |
|---|---|
python scripts/ux_audit.py <project_path> | Broad sweep across UX, color, type, motion, and a11y heuristics |
python scripts/accessibility_checker.py <project_path> | Targeted a11y checks: labels, ARIA, focus, contrast hints |
When a brief is loose, the easy move is to fall back on the same handful of "tasteful" choices. Don't.
If the brief leaves it open, ask: a color direction (cool / warm / neutral / bolder?), a style cue (pared-back / loud / retro / futuristic / organic?), a layout hint (single column / grid / asymmetric / edge-to-edge?).
| The reflex | The problem | The better question |
|---|---|---|
| Bento grid for everything | The house style of every AI mockup | Does this content actually want a grid? |
| Left-text / right-image hero | Seen a thousand times | What about oversized type, or a vertical scroll story? |
| Aurora / mesh gradient backdrop | The new lazy "wow" | What's a color pairing nobody expects here? |
| Frosted glass panels | A stand-in for "premium" | Would flat, high-contrast blocks read stronger? |
| Safe fintech blue | The escape hatch once purple was off-limits | Why not black, a deep red, an acid green? |
| "Orchestrate," "empower," "seamless" | Copy that smells generated | How would an actual person phrase this? |
| Dark canvas + neon glow | The signature "AI look" | What does the brand genuinely call for? |
| Everything rounded | Soft by default, not by intent | Where would a hard, squared edge hit harder? |
Each "safe" pick nudges the result closer to a template. Choose the braver option more often than feels comfortable.
No visual decisions until these are answered (or asked):
| Dimension | Ask | Why it steers the work |
|---|---|---|
| Time | How much runway? | Caps the ambition |
| Content | Real or lorem ipsum? | Decides layout flexibility |
| Brand | Existing guidelines? | May lock color and type |
| Stack | Which technologies? | Bounds what's buildable |
| Audience | Who, precisely? | Drives every aesthetic call |
Audience cues: Gen Z → loud, quick, mobile-native; Millennials → tidy, values-forward; Gen X → familiar, legible; Boomers → large type, strong contrast; B2B → credible, data-forward; Luxury → quiet confidence, lots of air.
The full treatment lives in ux-psychology.md; the shortlist:
| Law | The idea | How it shows up |
|---|---|---|
| Hick's | More options slow the decision | Trim choices; reveal complexity gradually |
| Fitts's | Big and near is easy to hit | Scale targets to their importance |
| Miller's | Working memory holds only a few items | Group things into small clusters |
| Von Restorff | The odd one out is remembered | Let the primary action look different |
| Serial position | Ends of a list stick | Anchor key items first and last |
Emotion runs on three layers — visceral (the gut reaction: color, imagery, vibe), behavioral (the using: speed, feedback, smoothness), reflective (the afterthought: "this says something about me"). Earn trust with security cues where stakes are high, real proof from others, easy access to help, consistency, and honest policies.
Each has a dedicated reference file; the load-bearing rules to keep in mind:
On every design task:
1. CONSTRAINTS — time, brand, stack, audience? Unclear → ask.
2. CONTENT — what exists, and what ranks highest?
3. DIRECTION — what suits this context? Unclear → ask, don't default.
4. BUILD — apply the principles; check against the anti-patterns.
5. REVIEW — Does it serve the user? Is it different from my usual?
Would I put my name on it?| Skill | When |
|---|---|
| frontend-design (here) | Before building — reason through color, type, and UX |
| [web-design-guidelines](../web-design-guidelines/SKILL.md) | After building — audit a11y, performance, and practices |
After it's coded, run the web-design-guidelines skill to check accessibility, focus states, motion, and performance.
The throughline: design is reasoning, not duplication. Every project earns a fresh look at its own context and people. Steer wide of the default SaaS template.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.