agentleFS
AlexanderTarAlexanderTar/swarmrun/skills/swarm-designer/SKILL.md

swarm-designer

Skill2 starsChanged 19 days agoApache-2.0View on GitHub

Extra rules for Swarm designer agents (role designer) producing a design artifact before UI is built. Use together with the `swarm` skill when your kickoff names this skill or your assignment has a `## Workflow` step naming you "designer".

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Follow the swarm skill first.

    line 8
  2. Very specific:

    Write to $SWARM_HOME/designs/<ROOT-KEY>/<ITEM-KEY>-<slug>.md.

    line 22
  3. Very specific:

    Write completed with the artifact's path in artifacts (gate artifact:design reads it from there.

    line 38

≈ “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

Designing with Swarm

Follow the swarm skill first; this adds to it.

What a designer does

You produce a design artifact for a screen, flow or component before it is built — you never write product code. Your output blocks the ui-tdd-reviewed package that builds what you designed, so be concrete: a coder should be able to implement your artifact without guessing, and a ui_reviewer should be able to check the build against it line by line.

Explore first

Use superpowers:brainstorming to explore options before committing to a design — but questions go to your parent, never to the end user directly; you have no user-facing channel. Send open questions with swarm_send (kind: "question") and keep working on parts that don't depend on the answer. Your design artifact replaces brainstorming's usual spec/plan steps: brainstorming on its own ends in a user-approved spec and superpowers:writing-plans, but here the artifact below is the spec a coder builds from — don't also write a separate spec or plan.

Tools

  • ui-ux-pro-max: --design-system when designing for a new surface with no existing system, --stack set to the target stack, e.g. react, react-native, swiftui, jetpack-compose or html-tailwind (see scripts/core.py's STACK_CONFIG for the full list), so its guidance matches what will actually get built.
  • building-components: component boundaries, prop surface, composition patterns — use it while sketching the Components section below.
  • The mobile skills (mobile-ios-design, mobile-android-design, expo-native-ui, vercel-react-native-skills) for anything with a native or React Native surface.

The artifact

Write to $SWARM_HOME/designs/<ROOT-KEY>/<ITEM-KEY>-<slug>.md. Sections, in order:

  • Goals — what this screen/flow/component is for, in 1–3 sentences.
  • Screens — each screen with its states: empty, loading, error, success. Use a Mermaid flow diagram for multi-screen flows.
  • Components — API sketch for each new or reused component (props, variants, composition).
  • Tokens — color, typography, spacing, both light and dark.
  • Interaction & motion — transitions, gestures, feedback.
  • Accessibility — contrast, focus order, screen-reader labels, reduced motion.
  • Mobile specifics — safe areas, gestures, platform conventions, touch targets (when the target includes a mobile surface).
  • Open questions — anything left unresolved; never silently pick an answer here and hide the uncertainty.

Headless Chrome: keep it in your dirs, close it before you finish

Chrome copies its app bundle into com.google.Chrome.code_sign_clone at every launch and deletes the copy only on a clean quit, so a Chrome left running leaks hundreds of thousands of inodes. If you start Chrome for screenshots or inspection:

  • Launch it with --user-data-dir pointing inside your own scratchpad or work dir, never a shared or system path.
  • Close it before your completed or handoff checkpoint, and on any early exit: send Browser.close over CDP, or SIGTERM it and wait for it to exit (never SIGKILL first).

Completing

Write completed with the artifact's path in artifacts (gate artifact:design reads it from there; the daemon registers it, not you). Don't mark the package done until the file is saved and readable — an unwritten design blocks the build package forever.

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.