agentleFS
fpindejfpindej/netrock/.claude/skills/polish-ui/SKILL.md

polish-ui

Skill234 starsChanged 3 months agoMITView on GitHub

Design-quality workflow that combines design taste, aesthetic direction, image-to-code, a Web Interface Guidelines audit, and real-browser verification. Use when building or polishing a UI surface, implementing a mockup/screenshot, or when asked to make a page "look great".

3 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

1 rule. About 0 are common, 0 are distinctive, and 1 is 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.

  • 1

    Very specific

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

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

    Standard review pass: Run frontend-reviewer + ux-designer in parallel per CLAUDE.md, then run /verify and commit.

    line 34

≈ “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

Polish or build a frontend surface to a high design standard, then prove it in a real browser.

This skill composes user-level skills installed at ~/.claude/skills/: design-taste-frontend, image-to-code, web-design-guidelines, playwright-cli, and the aesthetic packs shadcn, clean, premium. If any are missing, install them first:

npx skills add leonxlnx/taste-skill --skill design-taste-frontend --skill image-to-code -g -a claude-code -y
npx skills add vercel-labs/agent-skills --skill web-design-guidelines -g -a claude-code -y
npx skills add microsoft/playwright-cli --skill playwright-cli -g -a claude-code -y
npx skills add bergside/awesome-design-skills --skill shadcn --skill clean --skill premium -g -a claude-code -y

Arguments

  • Target: a route (/settings), a component path, or an image/mockup file. If omitted, ask what surface to work on.
  • Style (optional): shadcn (default - matches this app), clean, or premium. The style pack informs direction only; the app's existing tokens and components always win. Unified UX is a hard rule - never make one page feel like a different product.

Steps

  1. Set design direction: Invoke the design-taste-frontend skill and the chosen style pack skill. Distill their guidance into a short brief scoped to the target: hierarchy, spacing rhythm, typography, restraint. Filter out anything that conflicts with project hard rules (shadcn-svelte components, existing tokens, logical CSS, 44px touch targets, no overflow).
  2. If the target is an image or mockup: Invoke the image-to-code skill for its analysis method (section-by-section, readable reference views, no cards-inside-cards). Translate the output into Svelte 5 terms - the implementation must use runes, interface Props + $props(), shadcn-svelte components, and feature folders. Ignore the skill's instructions about generating images with Codex tooling; analyze the provided image instead.
  3. Implement via delegation: Pass the distilled brief to frontend-engineer (Delegation Rule applies - the orchestrator does not write src/ code). Include the specific taste/style constraints in the agent prompt so the direction survives the handoff.
  4. Audit: Invoke the web-design-guidelines skill against the changed files for Web Interface Guidelines compliance (accessibility, focus states, interaction details). Also run /review-design for project-specific standards. Fix findings via the engineer agent; loop until clean.
  5. Verify in a real browser: Invoke the playwright-cli skill (or use the project Playwright MCP) against the running app (Aspire full stack, or the frontend dev server). Probe for the frontend URL - do not assume a port. Screenshot the target at the canonical breakpoints from /review-design's design-tokens reference (320/375/768/1024/1440/2560), plus landscape on mobile. Check: no scrollbars in dialogs/modals, no horizontal overflow, touch targets >= 44px, both light and dark themes. Save screenshots to the scratchpad first.
  6. Standard review pass: Run frontend-reviewer + ux-designer in parallel per CLAUDE.md, then run /verify and commit.
  7. PR with visual evidence: When creating the PR, commit the meaningful screenshots under docs/sessions/assets/<date>-<topic>/ on the branch and embed them in the PR description, grouped by breakpoint and theme. Use SHA-pinned URLs (https://raw.githubusercontent.com/<owner>/<repo>/<full-sha>/docs/sessions/assets/...) so images survive branch deletion after squash-merge; raw URLs render only for public repos - in a private repo use https://github.com/<owner>/<repo>/blob/<full-sha>/<path>?raw=1.

Notes

  • Steps 1-2 produce a brief, not code. All src/frontend/ edits go through engineer agents unless the user overrides.
  • Screenshots from step 5 are evidence - if a breakpoint looks broken, that is a finding to fix, not a caveat to report.

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.