Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Stop coding agents from shipping generic UI with UIZZE's 800,000+ real web and iOS screens, product-specific contracts, required states, and a hard finish gate.
.claude/skills/sickn33-anti-ui-slop/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-11 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 9% | 0% |
| case-21 | ✗→✓ | ▲ Improved | 4% | 0% |
| case-23 | ✗→✓ | ▲ Improved | -17% | 0% |
Build product-specific UI with 800,000+ real web and iOS screens via UIZZE.
Use the product brief, existing interface, components, and local design system before reaching for outside references. UIZZE evidence is optional: it should answer a concrete visual question, not turn every interface task into a research project.
Use this skill when designing, implementing, redesigning, critiquing, or doing a pre-ship review of a web or iOS interface.
and important loading, empty, error, success, disabled, and permission states.
interaction conventions before adding a new abstraction or visual language.
hierarchy, workflow shape, allowed components, required states, responsive behavior, and observable acceptance criteria. Keep smaller changes smaller.
testimonials, users, or placeholder workflows to make a layout look complete.
The free skill and public catalogue work without an account, token, dependency, script, or executable. If a concrete unresolved visual question would benefit from evidence, use the smallest relevant set of screens or materials.
The optional authenticated UIZZE MCP exposes exactly find_ui_references and find_ui_materials. Use it only when those tools are actually available. If a search returns nothing, continue silently from repository evidence. Never claim an MCP-backed result that was not returned by the host.
Treat references as evidence, not templates. Transfer useful decisions about hierarchy, density, navigation, controls, responsive behavior, and state handling; never copy another product's branding, proprietary text, imagery, or exact layout.
Install the current free skill directly from its canonical source:
bashnpx skills add https://uizze.com --skill anti-ui-slop
When the environment supports it, render and inspect the result once. Fix observable breakage such as clipping, overlap, distorted media, inaccessible or inert controls, missing required states, and unintentional responsive behavior. Run the project's normal checks and keep the handoff concise.
security review, or project-specific tests.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-13 | pass→pass | 19,264 | 16,382 | -15% | 1 | 1 | 0% | 2,076 | 2,275 | +10% | 0 | 0 | — |
case-14 | pass→pass | 18,819 | 14,704 | -22% | 1 | 1 | 0% | 2,103 | 2,077 | -1% | 0 | 0 | — |
case-15 | pass→pass | 20,674 | 20,713 | +0% | 1 | 1 | 0% | 2,566 | 3,213 | +25% | 0 | 0 | — |
case-16 | fail→fail | 22,394 | 17,371 | -22% | 1 | 1 | 0% | 2,635 | 2,330 | -12% | 0 | 0 | — |
case-01 | pass→pass | 42,137 | 44,044 | +5% | 1 | 1 | 0% | 8,248 | 8,853 | +7% | 0 | 0 | — |
case-02 | fail→fail | 22,268 | 22,577 | +1% | 1 | 1 | 0% | 3,138 | 3,696 | +18% | 0 | 0 | — |
case-03 | fail→fail | 42,885 | 43,544 | +2% | 1 | 1 | 0% | 8,232 | 8,384 | +2% | 0 | 0 | — |
case-04 | pass→pass | 31,691 | 39,348 | +24% | 1 | 1 | 0% | 5,218 | 7,580 | +45% | 0 | 0 | — |
case-05 | pass→pass | 23,936 | 28,126 | +18% | 1 | 1 | 0% | 2,537 | 3,405 | +34% | 0 | 0 | — |
case-06 | pass→pass | 24,389 | 21,654 | -11% | 1 | 1 | 0% | 4,124 | 4,236 | +3% | 0 | 0 | — |
case-07 | pass→pass | 27,286 | 25,429 | -7% | 1 | 1 | 0% | 3,398 | 3,567 | +5% | 0 | 0 | — |
case-08 | fail→fail | 40,663 | 42,217 | +4% | 1 | 1 | 0% | 8,226 | 8,831 | +7% | 0 | 0 | — |
case-09 | pass→pass | 13,776 | 8,810 | -36% | 1 | 1 | 0% | 1,268 | 1,216 | -4% | 0 | 0 | — |
case-10 | fail→fail | 37,485 | 29,421 | -22% | 1 | 1 | 0% | 7,045 | 5,636 | -20% | 0 | 0 | — |
case-11 | fail→pass | 40,392 | 13,934 | -66% | 1 | 1 | 0% | 2,198 | 2,042 | -7% | 0 | 0 | — |
case-12 | pass→pass | 15,946 | 10,680 | -33% | 1 | 1 | 0% | 1,571 | 1,473 | -6% | 0 | 0 | — |
case-17 | fail→pass | 20,184 | 8,862 | -56% | 1 | 1 | 0% | 2,761 | 1,212 | -56% | 0 | 0 | — |
case-18 | pass→pass | 9,103 | 12,478 | +37% | 1 | 1 | 0% | 1,879 | 1,773 | -6% | 0 | 0 | — |
case-19 | fail→pass | 25,611 | 25,867 | +1% | 1 | 1 | 0% | 4,249 | 4,644 | +9% | 0 | 0 | — |
case-20 | pass→pass | 16,738 | 11,384 | -32% | 1 | 1 | 0% | 1,659 | 1,691 | +2% | 0 | 0 | — |
case-21 | fail→pass | 22,238 | 20,007 | -10% | 1 | 1 | 0% | 2,721 | 2,833 | +4% | 0 | 0 | — |
case-22 | pass→pass | 18,771 | 17,359 | -8% | 1 | 1 | 0% | 2,142 | 2,527 | +18% | 0 | 0 | — |
case-23 | fail→pass | 16,665 | 4,619 | -72% | 1 | 1 | 0% | 1,709 | 1,420 | -17% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +22 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.