agentleFS
youcislayoucisla/Agent-Foundry/skills/optional/design-language/SKILL.md

design-language

Skill1 starsChanged 23 days agoMITView on GitHub

Apply Apple-grade UI polish principles — restraint, coherence, intention,

8 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If a visual element has no function, remove it.

    line 24
  2. Distinctive:

    Every default (system font, #3b82f6 blue, 16px base) is a missed opportunity.

    line 32

“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Design Language

A skill for when you generate UI. Six principles. Apply on every visual output.

Procedure

The 6 principles

1. Restraint over decoration

If a visual element has no function, remove it. A border with no purpose is a wall. A shadow on every card is noise.

2. Coherence over variety

One type scale. One spacing scale. One motion language. A page with 4 fonts is chaos. A page with 1 font and 4 weights is a system.

3. Intention over default

Every default (system font, #3b82f6 blue, 16px base) is a missed opportunity. Pick a reason. "Why this color?" should have a one-sentence answer.

4. Hierarchy through scale

The primary action is bigger and higher-contrast than the secondary. Not "primary has a gradient, secondary doesn't." Size and weight carry the hierarchy.

5. Negative space is content

24px padding isn't empty — it's a visual breath that makes the adjacent content readable. Plan the empty.

6. Motion means something

Transitions answer "why did this change?" — not "look at me." A modal slides because it appeared in the user's mental model as elevated content. Linear motion is the AI tell. Use cubic-bezier.

The test

Show the design to someone for 3 seconds. Ask "what does this do?"

  • Specific answer (e.g., "manage my repos") → good
  • "It's a dashboard" / "It's a platform" → generic AI
  • "I don't know" → restart with intent

Concrete defaults

  • Type: 1 display + 1 body. Scale 1.125 or 1.25, 4-5 steps.
  • Spacing: 4px base, 4-7 steps. Never arbitrary 13px.
  • Color: 1-2 accent hues max. Backgrounds and grays fill the rest.
  • Motion: <300ms. Cubic-bezier curves. Respect prefers-reduced-motion.
  • Accessibility: WCAG AA contrast minimum (4.5:1 text, 3:1 UI). Visible focus rings. Keyboard navigation.

When to skip

  • Internal tools where polish doesn't matter
  • Backend / CLI work (not visual)
  • The user wants a 1:1 clone of an existing design (replicate, don't redesign)

Worked example

Three near-identical spacing values replaced by one scale.

Before: 13px, 17px, 22px gaps on the same panel.
After: 8px (1), 24px (3) from a 4-step spacing scale; the odd values deleted.

Anti-patterns

  • A second type scale, spacing scale, or accent hue without a reason.
  • Decoration with no interaction or state meaning.
  • A default system color or font chosen by accident.
  • Motion that delays the task or ignores reduced-motion.

Verification Checklist

  • One type scale, spacing scale, and motion language. (step 2)
  • Primary action is distinguished by size and contrast. (step 4)
  • Contrast meets WCAG AA for text and UI. (step 4)
  • Focus rings and keyboard paths are visible. (step 1)
  • prefers-reduced-motion is honored. (step 6)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.