| name | ui-polish |
| description | Iterative UI/UX polishing for Stripe-level visual quality. Use when app already works and looks decent, wanting to elevate to world-class through multiple passes. Not for complete overhauls. |
UI/UX Polish — Iterative Enhancement Workflow
When to Use: App already works and looks decent, you want to elevate it.
When NOT to Use: App is broken, styling is fundamentally wrong, or starting from scratch.
Key Insight: Asking for agreement ("don't you agree?") motivates better polish. Instructing to consider desktop vs mobile separately leads to much better outcomes.
THE EXACT PROMPT — UI/UX Polish
This prompt is used so frequently it's worth putting on a Stream Deck button:
I still think there are strong opportunities to enhance the UI/UX look and feel and to make everything work better and be more intuitive, user-friendly, visually appealing, polished, slick, and world-class in terms of following UI/UX best practices like those used by Stripe, don't you agree? And I want you to carefully consider desktop UI/UX and mobile UI/UX separately while doing this and hyper-optimize for both separately to play to the specifics of each modality. I'm looking for true world-class visual appeal, polish, slickness, etc. that makes people gasp at how stunning and perfect it is in every way. Think carefully and step-by-step; this is harder than it looks.
Best Models
| Model | Configuration |
|---|
| Claude Code + Opus 4.7 | Append "Think carefully and step-by-step; this is harder than it looks." (triggers adaptive thinking). Fixed-budget keywords like ultrathink are no-ops on 4.7 |
| Claude Code + Opus 4.5/4.6 | Append "Use ultrathink" (fixed extended-thinking budget) |
| Codex + GPT 5.2 | High/Extra-High reasoning |
| Gemini CLI | Good |
Why This Prompt Works
| Element | Effect |
|---|
| "don't you agree?" | Engages reasoning about whether improvements are possible |
| "desktop and mobile separately" | Prevents compromise solutions that are mediocre on both |
| "world-class", "Stripe" | Sets high anchor point |
| "makes people gasp" | Pushes toward exceptional quality |
| "Think carefully and step-by-step" | Triggers Opus 4.7 adaptive thinking for thorough analysis (pre-4.7, "ultrathink" was the equivalent fixed-budget keyword) |
Iteration Protocol
The Workflow
1. App already works and looks decent
↓
2. Run the polish prompt
↓
3. Agent makes incremental improvements
↓
4. Repeat many times (10+ iterations)
↓
5. Small improvements compound dramatically
Why Multiple Passes Work
- Each pass makes some incremental improvement (even if minor)
- These really add up after 10 iterations
- Multiple agents can work simultaneously on different areas
- Cumulative effect is dramatic
What Typically Gets Improved
Visual Polish
- Spacing and padding consistency
- Typography hierarchy
- Color contrast and accessibility
- Shadow and depth effects
- Border radius consistency
- Hover/focus states
Interaction Design
- Button feedback
- Loading states
- Transitions and animations
- Error/empty state design
Desktop vs Mobile
| Desktop Optimizations | Mobile Optimizations |
|---|
| Keyboard navigation | Touch target sizes |
| Hover states | Mobile-specific navigation |
| Multi-column layouts | Gesture support |
| Sidebar navigation | Performance |
| Power user shortcuts | Responsive breakpoints |
When to Use vs When NOT
USE When:
- App works correctly
- Basic styling is in place
- Want to elevate from "decent" to "world-class"
- Ready for iterative refinement
DON'T Use When:
- App is broken or buggy (fix bugs first)
- Styling is fundamentally wrong (need complete overhaul)
- No basic design system in place
- Starting from scratch
For complete overhauls, establish a design system and component library first.
Tips
- Don't skip iterations — Even when changes seem small, keep going
- Review changes — Make sure improvements don't break things
- Test on real devices — Desktop browser != mobile experience
- Consider accessibility — WCAG compliance matters
- Keep performance in mind — Pretty but slow is bad UX
References
