agentleFS
Joyboy-dyJoyboy-dy/felicio-ai-skills/skills/frontend-aesthetics/SKILL.md

frontend-aesthetics

Skill0 starsChanged 3 months agoMITView on GitHub

Use when improving the visual quality, hierarchy, typography, spacing, composition, density, responsive polish, or overall art direction of an existing or new frontend interface while respecting its design system.

5 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

7 rules. 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

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

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

    Do not use as a substitute for an accessibility audit (ux-ui-audit), color-system engineering (chromatic-design), or frontend architecture …

    line 14
  2. Distinctive:

    Content drives composition: Do not force every screen into identical cards or symmetrical grids.

    line 21
  3. Distinctive:

    Use the static smell scan for quick candidates:

    line 31
  4. Very specific:

    Read references/art-direction.md when a screen lacks clear hierarchy or needs a new composition.

    line 47
  5. Very specific:

    Prefer existing Button/Card/Dialog/Table/Form/etc.

    line 51
  6. Very specific:

    Read references/polish-checklist.md before finalizing a broad UI polish task.

    line 67
  7. Very specific:

    Run repository typecheck/lint/tests/build when implementation changes require them.

    line 73

“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

Frontend Aesthetics

Turn a functional interface into a deliberate visual system without replacing the project's identity or component architecture unnecessarily.

Use when

Activate for UI polish, visual hierarchy, layout composition, density, typography, spacing rhythm, responsive refinement, dashboard/card composition, empty states, or requests such as “make this look professional/premium/coherent.”

Do not use as a substitute for an accessibility audit (ux-ui-audit), color-system engineering (chromatic-design), or frontend architecture (senior-frontend-engineering).

Principles

  • Inherit before inventing. Inspect existing tokens, components, typography, icons, spacing, radius, shadows, and motion before introducing new choices.
  • Hierarchy over decoration. Improve reading order, grouping, emphasis, and whitespace before adding effects.
  • One visual language. Repeated surfaces and actions should feel related across the product.
  • Content drives composition. Do not force every screen into identical cards or symmetrical grids.
  • Responsive is a composition change. Preserve hierarchy and usability, not merely smaller dimensions.
  • Polish must survive states. Loading, empty, error, disabled, focus, overflow, and long content are part of the design.

Workflow

1. Discover the design language

Inspect global styles, theme/tokens, component primitives, layout shells, typography, icons, animation utilities, and several representative screens. Identify repeated spacing and component patterns.

Use the static smell scan for quick candidates:

python scripts/scan_visual_smells.py <project-root> --json

Raw values are not automatically wrong; use the scan to find inconsistency hotspots.

2. State the visual problem

Describe the current problem in terms of hierarchy, rhythm, density, contrast, alignment, affordance, or responsiveness. Avoid vague labels such as “ugly” or “modernize.”

3. Choose an art direction with constraints

Define a small set of decisions: page density, dominant surface treatment, typographic scale, spacing rhythm, emphasis strategy, border/radius/shadow behavior, and motion restraint. Preserve brand/design-system constraints already present.

Read references/art-direction.md when a screen lacks clear hierarchy or needs a new composition.

4. Reuse primitives first

Prefer existing Button/Card/Dialog/Table/Form/etc. primitives and theme tokens. Customize composition around them before duplicating components or hard-coding one-off styles. If a missing primitive is genuinely reusable, add it in the project's established component layer.

5. Polish high-impact details

Prioritize:

  1. page/frame spacing and max-width;
  2. heading/subheading/action hierarchy;
  3. grid alignment and grouping;
  4. typography and line length;
  5. control sizes and hit targets;
  6. surface separation;
  7. empty/loading/error states;
  8. responsive collapse/reflow;
  9. subtle motion only where it clarifies change or relationship.

Read references/polish-checklist.md before finalizing a broad UI polish task.

6. Validate rendered behavior

Inspect real output at representative viewport widths. Check long labels, localization-sized text, data-heavy states, empty states, focus-visible, reduced-motion behavior when motion is present, and light/dark themes when supported.

Run repository typecheck/lint/tests/build when implementation changes require them.

Avoid

  • Generic “AI dashboard” styling: excessive gradients, glow, glass, huge radii, decorative badges, or every section in a card.
  • Adding animation everywhere.
  • Replacing an established icon/component library for visual preference alone.
  • Hard-coded colors that bypass theme tokens.
  • Desktop-first layouts that merely stack at mobile width.
  • Removing useful information to achieve visual minimalism.
  • Redesigning the entire application when the request concerns one screen.

Output

Explain the detected design language, visual issues, chosen direction, exact changes, responsive/state checks, and any compromises or unvalidated rendered details.

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.