agentleFS
ryanschueryanschue/automated-product-engineer/skills/design-polish/SKILL.md

design-polish

Skill0 starsChanged 27 days agoMITView on GitHub

Implement focused UI polish, interaction refinements, responsive spacing, overlay behavior, and visual-regression fixes from design evidence or QA findings. Use for product-surface refinement rather than broad design-system infrastructure.

4 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

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 4

    Distinctive

    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:

    Use the screenshot, design node, recorded defect, or visual diff as evidence.

    line 8
  2. Distinctive:

    Read references/interaction-debugging.md for overlay, animation, dense-layout, and visual-diff failure patterns.

    line 10
  3. Distinctive:

    Add or update focused behavior and geometry tests when the bug is contract-like.

    line 25
  4. Very specific:

    Use design-system when the correction changes reusable tokens, component variants, catalog metadata, or shared visual infrastructure.

    line 38
  5. Distinctive:

    Run focused component tests first, then targeted visual checks and a real-browser inspection.

    line 40

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “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 Polish

Use the screenshot, design node, recorded defect, or visual diff as evidence. Identify the source-of-truth component or token before editing.

Read references/interaction-debugging.md for overlay, animation, dense-layout, and visual-diff failure patterns.

Intake and pushback

Inspect the live surface, design system, reference frames, current screenshots, and responsive behavior first. Establish the intended hierarchy, primary task, approved visual source, relevant states, and viewport range. Ask only for visual or interaction decisions that cannot be inferred and would change the result. When a designer must provide missing direction, generate a focused request with the current screenshot or component context, the exact ambiguity, and the states or measurements needed.

Working loop

  1. Reproduce the exact state in the production component, including viewport, content density, interaction, and surrounding layout.
  2. Compare against the current design source and repository conventions. Resolve meaningful ambiguity before inventing a visual rule.
  3. Fix the narrowest owning boundary:
    • shared appearance belongs in semantic tokens or reusable components
    • domain rendering belongs in the feature component
    • state and persistence behavior belong outside presentational primitives
  4. Add or update focused behavior and geometry tests when the bug is contract-like.
  5. Inspect the running result across affected widths and interaction states.
  6. Classify visual diffs as defect, stale baseline, environment difference, or intentional change. Accept new baselines only after inspecting them, then rerun without update mode.

Interaction invariants

  • Hover and focus affordances appear next to the object they affect and do not cause unrelated reflow.
  • Menus, popovers, and dialogs anchor to their real trigger, remain inside the usable viewport, dismiss predictably, and preserve keyboard focus.
  • Animated surfaces gate pointer and accessibility behavior consistently with their visible state.
  • Icon glyph size and interaction-target size are separate decisions.
  • Compact layouts protect meaningful text and actions before decorative whitespace.
  • Loading, success, rollback, empty, error, and disabled states remain visually stable.

Use design-system when the correction changes reusable tokens, component variants, catalog metadata, or shared visual infrastructure. Use responsive-design for cross-width geometry work and figma-to-app-feature for a new or substantially redesigned experience.

Run focused component tests first, then targeted visual checks and a real-browser inspection. If the change lives in a dependent pull-request chain, fix the earliest owning branch and propagate forward rather than masking it later.

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.