Installable skillMIT

Design Critique Polish

Critique and polish frontend interfaces before shipping, with focused review of hierarchy, typography, spacing, states, color, and UX clarity. Use when a page, component, or app surface feels unclear, inconsistent, flat, awkward, under-tested, or not yet production-grade and needs design verifica...

BY markoblogoUPDATED Oct 28, 2026
INTENDED FOR✳Claude Code>_Codex↗Cursor◎ChatGPTInstall in a listed agent.

instructions.md

---
name: design-critique-polish-33e1f251299a
description: Critique and polish frontend interfaces before shipping, with focused review of hierarchy, typography, spacing, states, color, and UX clarity. Use when a page, component, or app surface feels unclear, inconsistent, flat, awkward, under-tested, or not yet production-grade and needs design verification instead of a broad rewrite.
license: MIT
metadata:
  abvx_status: experimental
  abvx_origin: adapted
---

# Design Critique Polish

Run a deliberate critique pass before shipping. This is not a vague “make it nicer” prompt.

## Use For

- pre-ship review of a frontend surface;
- critique after implementation but before release;
- “this feels off, but I do not know why” situations;
- tightening hierarchy, spacing, typography, copy, motion, and edge states.

## Inputs

Take:

- the page, component, or route to review;
- screenshots, live URL, or repo files;
- current `PRODUCT.md` / `DESIGN.md` if available;
- known constraints: ship date, design system, browser targets, accessibility requirements.

## Critique Order

Evaluate in this order:

1. **Surface fit**
   - does this look like the right class of product or brand?
2. **Hierarchy**
   - can a first-time user tell what matters first?
3. **Layout and rhythm**
   - spacing, grouping, alignment, section cadence, empty space.
4. **Typography**
   - font choice, scale, line length, display/body contrast, overflow.
5. **Color and contrast**
   - palette discipline, readable text, accent overuse, weak affordances.
6. **States and resilience**
   - hover, focus, loading, empty, error, disabled, overflow, long labels.
7. **Motion and interaction**
   - helpful versus noisy; reduced-motion handling. If the diff materially changes transitions, keyframes, gestures, Lottie, popovers, toasts, or hover/press motion, run `motion-review-gate` as the dedicated review pass.
8. **Technical quality**
   - responsive behavior, accessibility signals, obvious performance issues.

## Output Mode

Produce:

- `what works`
- `what fails`
- `highest-leverage fixes first`
- `ship blockers`

Keep findings concrete. Name the surface, the defect, and the consequence.

## Fix Strategy

- Prefer small, high-leverage corrections over full rewrites.
- If the core design language is wrong, say so explicitly instead of endlessly polishing local details.
- If the page is good enough to ship, say that too and focus on blockers only.

## Common Frontend Tells To Catch

- same card pattern repeated everywhere;
- weak first viewport with no real anchor;
- gray text that reads low-contrast on tinted backgrounds;
- spacing that is mechanically even instead of rhythmic;
- headings that overflow or collapse on tablet/mobile;
- decorative motion without task value;
- UI copy that sounds like implementation notes rather than product language.

## Final Report

Include:

- top issues ordered by severity;
- what should change now versus later;
- whether the surface is ship-ready;
- what to verify in-browser after changes.
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.