agentleFS
egoxx1egoxx1/unified-master-harness/.claude/skills/dev-ui-verify/SKILL.md

dev-ui-verify

Skill1 starsChanged 17 days agoMITView on GitHub

Build or verify user-facing UI changes in the running app. Use for any change to how the UI looks or behaves — style, spacing, size, color, layout, components, interaction — including one-line CSS fixes (quick check). Not for text-only copy edits or backend-only work.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 7 are distinctive, and 3 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.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Text-only copy edits: do not need this skill unless the new text can overflow, wrap or truncate.

    line 14
  2. Distinctive:

    If a stronger visual direction is requested, use a small number of relevant references as evidence for hierarchy, density, navigation, and …

    line 21
  3. Very specific:

    Prefer content-first hierarchy, consistent spacing/type scales, semantic colors, clear primary/secondary actions, real loading/empty/error …

    line 25
  4. Distinctive:

    Avoid generic AI defaults unless the project already uses them: purple/indigo everywhere, gratuitous gradients, maximum-radius everything, …

    line 27
  5. Distinctive:

    use native semantic elements when available,

    line 32
  6. Distinctive:

    do not rely on color alone,

    line 35
  7. Distinctive:

    Start the actual application with verified project commands.

    line 40
  8. Distinctive:

    Check the changed route/state at a representative desktop viewport and mobile viewport, using project breakpoints when known.

    line 41
  9. Very specific:

    Use screenshots when they materially help compare quality or prove the result.

    line 54
  10. Very specific:

    If a visible issue is found, fix the underlying layout/style/interaction cause and rerun the affected browser check.

    line 55

≈ “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

UI Engineering and Browser Verification

A visual change is not complete from code inspection alone: shared styles, the cascade and responsive rules make CSS results hard to predict from the diff.

Depth: match the check to the change

  • Small visual fix (one component's spacing, size, color, alignment, a single style rule): start the app, open the affected screen, and capture one screenshot of the changed area. Add a mobile-width screenshot when the component is responsive. Glance for collateral changes in places that share the style. Skip the full checklist below.
  • Meaningful UI work (new or reworked screens, layout or navigation changes, new interactions, responsive redesign): follow the whole procedure below.
  • Text-only copy edits do not need this skill unless the new text can overflow, wrap or truncate.
  • The app cannot be run here (no dev command, missing services): say the change is visually unverified and why. Never imply a visual check that did not happen.

Before editing

  1. Inspect the product's existing design system, tokens, shared components, and closest relevant screens.
  2. Preserve existing information architecture and interaction contracts unless redesign is requested.
  3. If a stronger visual direction is requested, use a small number of relevant references as evidence for hierarchy, density, navigation, and interaction patterns. Rebuild principles in the product's own language; do not copy branding or proprietary layouts.

Production-quality defaults

Prefer content-first hierarchy, consistent spacing/type scales, semantic colors, clear primary/secondary actions, real loading/empty/error states where affected, responsive behavior designed with the change, and accessible native controls.

Avoid generic AI defaults unless the project already uses them: purple/indigo everywhere, gratuitous gradients, maximum-radius everything, uniform card grids that ignore priority, oversized equal padding, excessive shadows, generic hero layouts unrelated to content.

Accessibility baseline

For affected interactions:

  • use native semantic elements when available,
  • preserve keyboard operation and visible focus,
  • label controls without visible text,
  • do not rely on color alone,
  • maintain adequate contrast.

Runtime verification

Start the actual application with verified project commands. Check the changed route/state at a representative desktop viewport and mobile viewport, using project breakpoints when known.

Inspect where relevant:

  • layout, density, spacing, alignment,
  • typography and information hierarchy,
  • overflow/clipping/wrapping,
  • navigation and primary interactions,
  • hover/focus/active states,
  • loading, empty, and error states,
  • browser console,
  • failed network requests,
  • responsive behavior.

Use screenshots when they materially help compare quality or prove the result. If a visible issue is found, fix the underlying layout/style/interaction cause and rerun the affected browser check. Do not keep layering CSS exceptions over a wrong layout approach.

See references/ui-quality.md for a deeper review checklist.

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.