agentleFS
strombraatenstrombraaten/claude-skills/skills/design-explorer/SKILL.md

design-explorer

Skill0 starsChanged 5 months agoMITView on GitHub

Multi-direction UI design exploration tool. Generates 6–8 completely different design mockups as static HTML previews and displays them in a browsable carousel gallery with voting, notes, and feedback synthesis — so you can compare directions instead of reacting to one thing at a time. Use this skill whenever the user wants to explore UI design options, generate multiple design directions, mockup a screen, iterate on visual design, or says things like "show me some design options", "generate some UI mockups", "I want to explore how this could look", or "let's figure out the design for X". Also use it when the user is stuck in a design feedback loop or wants to break out of a single-direction approach. This is the right skill whenever design exploration, visual iteration, or UI direction-finding is the goal — even if the user just says "I want to see what this could look like."

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. CODE-SAURABH/OpenSkills★ 2same 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.

  • In this repository3 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 7 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Start with a focused conversation to understand the design target.

    line 24
  2. Distinctive:

    Use this real project context in the mockups.

    line 37
  3. Distinctive:

    Use the project's real field names and realistic data values (not Lorem ipsum)

    line 67
  4. Very specific:

    Read the gallery template from assets/gallery_template.html (in this skill's directory).

    line 78
  5. Very specific:

    Use preview_start (from the Claude Preview MCP) to launch a local server pointing at the _design-exploration/ directory.

    line 95
  6. Distinctive:

    When the user returns with feedback:

    line 106
  7. Distinctive:

    Extract the signal: Look for comparative statements ("the layout of #3 but the color palette of #5"), strong reactions (explicit loves or …

    line 108
  8. Distinctive:

    Don't be afraid of bold choices.

    line 132
  9. Distinctive:

    If generating a second round, explicitly call out what you took from each previous direction in a comment at the top of each new file.

    line 134

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Design Explorer

Help the user rapidly explore multiple distinct UI design directions for a screen or component. The core insight: comparing directions ("I want the warmth of #2 but the layout of #5") is far more productive than reacting to one mockup at a time.

Phase 1 — Gather Context

Start with a focused conversation to understand the design target. Ask:

  • What specific screen or component are we designing? (e.g., "reading journal entry view", "dashboard overview", "onboarding flow")
  • What's the app/project? Can you point me to the codebase or describe it briefly?
  • Are there any hard constraints? (mobile-only, must match existing brand, etc.)
  • How many directions would you like? (default: 6)

Then read the project codebase to find:

  • Existing color palette and CSS variables (look for :root blocks, Tailwind config, CSS files)
  • Real data schemas and field names (look at types, state objects, API shapes)
  • Typography choices (font families in use)
  • Any existing component patterns or UI library in use

Use this real project context in the mockups. If you can't find a codebase, ask the user to describe the data model and color preferences briefly.

Phase 2 — Generate the Mockups

Generate 6 (or the requested number) completely distinct HTML mockups. Each should represent a fundamentally different design philosophy — not just color variations.

What "genuinely different" means:

  • Different layout paradigms (single-column centered vs. sidebar+content vs. full-bleed cards)
  • Different visual registers (warm+editorial vs. data-dense+utilitarian vs. playful+illustration)
  • Different information hierarchies (what's most prominent varies across designs)
  • Different interaction models (inline editing vs. modal vs. separate view vs. bottom sheet)

Suggested direction archetypes to draw from:

  1. Editorial / magazine-style (generous whitespace, large typography)
  2. Dashboard / data-dense (compact, lots of info visible at once)
  3. Card-based / spatial (floating cards, elevation, shadows)
  4. Minimalist / monochrome (one accent color, high contrast)
  5. Warm / human (rounded corners, warm tones, handwritten-feeling elements)
  6. Dark / premium (dark background, glowing accents)
  7. Mobile-native / gesture-forward (bottom navigation, swipe affordances)
  8. Brutalist / bold (strong borders, grid-heavy, unapologetically structured)

Pick 6 that feel genuinely different from each other given the context.

Technical rules for each mockup file:

  • File naming: design-1.html, design-2.html, etc.
  • Each file is fully self-contained: all CSS inline or in <style> tags
  • External dependencies OK: Google Fonts (link tag), no JS libraries
  • Use the project's real field names and realistic data values (not Lorem ipsum)
  • Each mockup should have a clear <!-- Design Name: [descriptive name] --> comment at the top and an <h1> or visually prominent title showing the design name (e.g., "Direction 2: Warm Editorial")
  • Target a single screen/state — don't try to show the whole app flow
  • Make it look complete and polished — this isn't a wireframe

Write all mockups to a directory called _design-exploration/ in the project root (or the closest sensible location). If no project root is clear, ask the user where to save them.

Phase 3 — Create the Gallery Viewer

Read the gallery template from assets/gallery_template.html (in this skill's directory). Replace the placeholder __DESIGNS_JSON__ with a JSON array of objects, one per design:

[
  { "file": "design-1.html", "name": "Direction 1: Warm Editorial" },
  { "file": "design-2.html", "name": "Direction 2: Dark Premium" }
]

Save the result as _design-exploration/_gallery.html.

The gallery handles all browsing, voting, and feedback collection — you don't need to write any gallery logic yourself.

Phase 4 — Start the Preview

Use preview_start (from the Claude Preview MCP) to launch a local server pointing at the _design-exploration/ directory. Open _gallery.html as the entry point.

Tell the user:

"I've generated [N] design directions and opened the gallery. Browse with ← → arrow keys, vote each design up or down, and add notes. When you've seen them all, click 'Summarize Feedback' and paste the summary back here — or just describe what you liked and disliked in your own words."

Phase 5 — Synthesize Feedback and Iterate

When the user returns with feedback:

  1. Extract the signal. Look for comparative statements ("the layout of #3 but the color palette of #5"), strong reactions (explicit loves or hates), and recurring themes.
  2. Synthesize into a design brief. Summarize: what layout direction won, what visual tone resonated, what to avoid, and what specific elements to carry forward.
  3. Offer the user a choice:
    • Refine: Generate a new round of 3–4 designs that explore the winning direction more deeply, applying the feedback. Repeat from Phase 2 with narrower brief.
    • Converge: Generate 1–2 "synthesis" designs that combine the best elements from multiple directions.
    • Handoff: Help convert the winning mockup into production-ready code using the actual framework/component library from the project.

For the handoff path, read the winning design-N.html carefully and translate it to the project's actual stack (React components, Tailwind classes, ShadCN components, etc.).

Tips for Writing Great Mockups

  • Spend the most effort making each design feel complete — real labels, real numbers, real copy. A mockup that looks like a real app is 10x more useful than a careful wireframe.
  • Vary the mockups along axes the user actually cares about: information density, warmth, modernity, mobile-friendliness. If they mentioned "I want something clean", skew the archetypes accordingly but still give them real variety.
  • Don't be afraid of bold choices. The user can vote them down. A surprising direction often unlocks a conversation that a safe one doesn't.
  • If generating a second round, explicitly call out what you took from each previous direction in a comment at the top of each new file.

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.