visual-design-foundations — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited visual-design-foundations (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.
Visual design foundations is the discipline of shaping an interface's visual craft — palette, type, spacing rhythm, density, elevation, optical balance, hierarchy, contrast intent, motion feel, and brand fit — so that the surface's structure becomes legible rather than merely decorated. Its central premise is that visual design is not a coat of paint applied after the layout is done; it is the layer through which priority, grouping, affordance, and tone are made visible, so a user can scan and act without reading every label. The craft works through a small set of primitives chosen by role before raw value: palette roles (surface, text, accent, success/warning/danger/info, disabled) before specific colors, type roles (page title, section heading, control label, body, metadata, numeric emphasis) before specific sizes, spacing/density rules tuned to repeated real use rather than a single screenshot, and optical alignment that centers by perceived weight rather than only geometry — with elevation, border, and background reserved for clarifying grouping and affordance instead of adding noise. Crucially, it is a foundation skill by design: it gives the entry-level orientation across every craft dimension but explicitly hands off anything that needs the deep contract — formal color math (color-system-design), font-loading engineering and full type systems (typography-system), deep hierarchy modeling (visual-hierarchy), token governance (design-system-architecture / theme-system-design), sign-meaning analysis (semiotics), or accessibility compliance (a11y) — to the skill that owns it, and it cross-checks its decisions against semiotics and a11y before treating them as done.
Design and audit visual craft for interface surfaces at foundation level. Covers:
visual-hierarchy.color-system-design.typography-system.interaction-feedback.Visual design is not decoration after structure; it is how the structure becomes legible. Good visual craft makes priority, grouping, affordance, and tone visible without asking the user to parse every label.
Two working rules follow. First, choose by role before raw value — name what a color or type size is for before assigning a specific value, so the surface stays consistent and changeable rather than hand-tuned per screen. Second, spend the cheapest hierarchy lever first — reach for size, weight, and spacing before color, because color is the lever most easily exhausted and most entangled with meaning (semiotics) and contrast compliance (a11y).
Generated drafts are inputs, not authority. Visual drafts produced by AI tools, design-system templates, or UI kits can speed exploration, but they are material to evaluate, not finished designs. The final surface still needs the same craft audit — role clarity, hierarchy, density, elevation restraint, motion purpose, brand fit, and the boundary handoffs — applied to refine, correct, or reject the generated input.
Keep this skill at foundation level. This is the entry layer; when a single dimension needs its deep contract, hand off to the specialization that owns it — formal color math to color-system-design, font-loading engineering and full type systems to typography-system, deep hierarchy modeling to visual-hierarchy, theming and token governance to theme-system-design / design-system-architecture.
Visual hierarchy is built from a small set of pre-attentive levers — features the eye resolves in well under a quarter-second, before conscious reading. Spend them in roughly this order, structural first:
a11y.semiotics), and it must clear contrast and non-color-only checks (route those to a11y). Prefer to establish hierarchy with the first four levers and use color to confirm, not to carry, priority.The most common scan failure — "everything looks the same / flat and hard to scan" — is almost always under-use of levers 1–3 and over-reliance on lever 5. Fix it by differentiating type roles, weight, and spacing first, not by adding more accent colors. A fast diagnostic: if the hierarchy collapses in grayscale, at low resolution, or when accent color is removed, the surface is leaning on hue instead of structure — rebuild it on size, weight, and spacing.
This foundation gives the ranked-lever model for everyday craft. When hierarchy needs deeper modeling — formal type-scale ratio selection, contrast tuned as an explicit ordering signal, and the layered primary/secondary/tertiary information structure of a complex surface — hand off to visual-hierarchy, which owns that contract.
surface, text (and a muted/secondary text), accent, the status set (success, warning, danger, info), and disabled before assigning any raw color. A palette handed over as a list of raw hex codes "that look nice" has no roles; map each color onto a role, drop or merge the ones with no role, and ensure every status state has a dedicated role color rather than an overloaded or missing one.color-system-design's contract — hand them off rather than building a full color system here.a11y's contract — defer it rather than asserting compliance from this skill.rem) so the scale respects user text-size settings. Picking a concrete starting ratio for one surface is foundation-level direction; engineering the product-wide scale and using type-scale ratios as a formal ordering signal hand off to typography-system / visual-hierarchy.font-display, variable fonts) are typography-system's contract; this foundation only sets the role-and-scale direction.layout-composition's concern; the spacing rhythm within and around elements is this skill's.Optical alignment is the craft of making elements look centered, balanced, and intentional at real size. Geometric center is a useful starting point, but it is not always the visual center:
layout-composition), what a sign means (semiotics), or accessibility requirements (a11y).prefers-reduced-motion; never ship motion that has no reduced-motion fallback. (The deeper a11y obligations of motion sit with a11y; the feel and the existence of a reduced path are foundation-level.)interaction-feedback's contract; this skill owns the motion feel, not the feedback-state system.State the surface's intended tone and scanning demand before making craft choices, and match the visual register to the surface type. The same craft levers serve very different tones: a dense, restrained, fast-scanning admin tool and an airy, expressive marketing surface are both "correct" only relative to their purpose. Use this matrix as the starting register per archetype, then adjust:
| Surface archetype | Density | Color / noise | Personality |
|---|---|---|---|
| Operational tool / admin | Compact, high information density | Minimal accent; neutrals carry the surface | Restrained, fast-scanning, calm |
| Dashboard | Compact–comfortable | Accent reserved for status, alerts, and key metrics | Data-forward, quiet chrome |
| Docs / reference | Comfortable, generous measure | Very low chrome; near-monochrome | Quiet, legible, unobtrusive |
| Editorial / content | Generous, airy | Expressive but disciplined accent | Voiced, considered |
| Marketplace / listing | Medium; scannable cards | Accent for CTAs and trust signals | Approachable, trustworthy |
| Brand / marketing page | Airy, large scale | Bold, expressive color and motion | Distinctive, high-personality |
This skill sets foundation-level brand fit; it does not replace brand-specific guidelines or a formal brand/token system (hand those to design-system-architecture / theme-system-design).
visual-hierarchy.semiotics; contrast ratios, non-color-only, focus, and AT behavior to a11y; and the deep contract of any single dimension to its owner (color-system-design, typography-system, visual-hierarchy, theme-system-design, interaction-feedback).This skill ships comprehension and application eval artifacts in its sibling evals/ directory (evals/comprehension.json, evals/application.json). The checklist in ## Verification below is the authoring gate for visual-design decisions; the eval files are the grader surface.
visual-hierarchycolor-system-designtypography-systeminteraction-feedbacka11y; sign-meaning deferred to semiotics; deep hierarchy / color math / font engineering / token governance handed off to their owning skills| Use instead | When |
|---|---|
semiotics | The question is what a color, icon, badge, shape, or visual metaphor means. |
visual-hierarchy | The task needs deep hierarchy modeling — formal type-scale ratios, contrast as an ordering signal, layered primary/secondary/tertiary information structure, or detailed proximity/spacing-hierarchy mechanics. |
color-system-design | The task is a full color system — palette construction, semantic color tokens, OKLCH/LCH perceptual scale generation, contrast-pair matrices, wide-gamut fallbacks, or light/dark/high-contrast parity. |
typography-system | The task is a full typography system — typeface selection and pairing, vertical rhythm, line-height/measure rules, or web-font delivery (subsetting, font-display, variable fonts). |
design-system-architecture | The task is semantic tokens, component APIs, theming contracts, or governance. |
theme-system-design | The task is the theming/token-delivery layer — design tokens, semantic token layering, CSS custom-property strategy, or runtime theme switching. |
interaction-feedback | The task is feedback around actions and system state — loading, skeletons, optimistic updates, progress, success/error/empty states, retries, pending/disabled, autosave, undo, perceived latency. |
layout-composition | The task is responsive structure, grid tracks, breakpoints, or section order. |
a11y | The task is WCAG contrast, focus, labels, keyboard access, reduced-motion compliance, or assistive technology. |
microcopy | The task is the wording inside buttons, dialogs, empty states, errors, toasts, tooltips, or validation. |
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.