agentleFS
dlaumetsdlaumets/agent-design-pack/design-auto/SKILL.md

design-auto

Skill1 starsChanged 4 months agoMITView on GitHub

Designs, builds, audits, polishes, prototypes, and reviews frontend interfaces through one integrated workflow. Use for UI/UX work, visual design, accessibility, responsive behavior, design systems, animation, motion audits, gesture-driven interfaces, and final product polish. Add --deep only for exhaustive phase-gated work.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. VladTheProgrammer/claude-asimov★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 9 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before judging or changing an existing interface:

    line 44
  2. Distinctive:

    Read the target and at least one representative neighboring component.

    line 46
  3. Distinctive:

    Read the project's tokens, theme, global styles, and shared primitives.

    line 47
  4. Very specific:

    If .superdesign/init/ exists and the mode is prototype, read all of it.

    line 50
  5. Distinctive:

    Do not block ordinary work merely because PRODUCT.md, DESIGN.md, or a specific design tool is absent.

    line 52
  6. Distinctive:

    Ask only when the missing answer materially changes the design or causes destructive work.

    line 58
  7. Distinctive:

    Use real content when available.

    line 59
  8. Distinctive:

    Never commit, publish, or deploy without explicit permission.

    line 62
  9. Distinctive:

    Verify in proportion to risk: static inspection, tests, responsive checks, keyboard flow, and browser interaction where available.

    line 63
  10. Distinctive:

    Start with a phase plan containing:

    line 67

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “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

Integrated Design Workflow

Handle the request directly. Do not invoke a chain of overlapping design skills. Select the smallest mode below, read only its reference file, and continue to the mode's natural stopping point.

Modes

GoalModeRequired reference
New page, feature, component, or redesignbuildBUILD.md
UX, accessibility, visual, or technical reviewauditAUDIT.md
Nearly finished UI needs refinementpolishPOLISH.md
Explore visual directions before implementationprototypePROTOTYPE.md
Add, review, audit, name, or improve motionmotionMOTION.md

Combine references only when the outcome truly crosses modes:

  • Build with important motion: BUILD.md + MOTION.md.
  • Full release review: AUDIT.md + MOTION.md.
  • Motion-only review or codebase audit: MOTION.md alone.
  • Prototype then implement only when the user has asked for both.

Invocation variants

  • --shape: clarify users, hierarchy, flows, and acceptance criteria before code.
  • --craft: implement from a clear or confirmed brief.
  • --component: constrain work to one component and direct dependencies.
  • --visual, --tech, --a11y, --full: select audit depth.
  • --bolder, --quieter, --typeset, --layout, --clarify, --harden: select polish focus.
  • --animate: add purposeful motion using MOTION.md.
  • --motion-review: review motion in a diff and return Block/Approve.
  • --motion-audit [quick|standard|deep]: survey the codebase and produce prioritized findings; write implementation plans only after selection.
  • --gesture: design drag, swipe, sheet, spring, or momentum interactions.
  • --vocabulary: identify the precise name for a described animation effect.
  • --deep: exhaustive, phase-gated design work. Use only when explicitly requested.

Required context

Before judging or changing an existing interface:

  1. Read the target and at least one representative neighboring component.
  2. Read the project's tokens, theme, global styles, and shared primitives.
  3. Infer the product register: task-focused product UI or expressive brand/marketing.
  4. Preserve established conventions unless a concrete UX gain justifies changing them.
  5. If .superdesign/init/ exists and the mode is prototype, read all of it.

Do not block ordinary work merely because PRODUCT.md, DESIGN.md, or a specific design tool is absent. Existing code and the user's brief are valid context.

Operating rules

  • With a target but an unclear goal, perform lightweight triage and choose the safest mode.
  • With neither a target nor an actionable outcome, ask one concise blocking question.
  • Ask only when the missing answer materially changes the design or causes destructive work.
  • Use real content when available. Do not ship placeholder-only UI unless asked for a wireframe.
  • Extend existing primitives before creating new abstractions.
  • Implementation requests authorize relevant code edits; reviews and audits remain read-only unless fixes are requested.
  • Never commit, publish, or deploy without explicit permission.
  • Verify in proportion to risk: static inspection, tests, responsive checks, keyboard flow, and browser interaction where available.

--deep

Start with a phase plan containing:

  1. Context and current-system discovery
  2. Shape or prototype
  3. Build
  4. Audit
  5. Polish

State the aesthetic direction, measurable success criteria, and stop condition before implementation. Do not use --deep for ordinary broad requests.

Output

Lead with the outcome. End with a compact execution summary:

  • completed work
  • design decisions
  • verification performed
  • blockers or assumptions
  • artifacts or files produced
  • next safe action, only if one remains

For audits, prioritize evidence and concrete fixes over generic advice. For implementations, report files changed rather than repeating the code.

Sources

This integrated workflow incorporates and adapts guidance from Emil Kowalski's skills repository, Apple's fluid-interface principles, Vercel's Web Interface Guidelines, and the previously installed local design workflows. Emil-derived material is used under the MIT license; see LICENSE-EMIL-KOWALSKI.md.

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.