Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Apply a visual direction — an archetype (high-end agency, editorial minimal, brutalist, soft-SaaS, dark-tech) or one of 138 named design systems (apple, linear-app, stripe, vercel, notion, material, shadcn, spotify, tesla…) — by resolving it into the token system. Use when the user wants a specific look/vibe/brand feel, or asks to make a design feel premium/expensive/non-generic.
.claude/skills/plugin87-apply-aesthetic/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 32% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 299% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 41% | 0% |
Choose and apply a design direction without breaking accessibility.
taste/design-taste.md → Brief Inference + Variance Mandate). Generating before deciding = slop.taste/aesthetic-systems.md:python3 scripts/design_systems.py list (or search <term> / show <name>); specs live in design-systems/library/<name>/DESIGN.md.aesthetic-systems.md): re-point semantic.* tokens to the chosen system's color roles; map typography/spacing/radius/shadow/motion to tokens/*.json.scripts/contrast.py / a11y-audit). A brand value that fails must be adjusted — taste never overrides POUR.taste/motion-choreography.md; run the pre-flight aesthetic check in design-taste.md.Updated/overridden semantic tokens + notes on type/space/motion, then render via design-code. Confirm the result passes both the aesthetic check and accessibility.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→pass | 26,646 | 21,677 | -19% | 1 | 1 | 0% | 4,849 | 4,240 | -13% | 0 | 0 | — |
case-01 | fail→pass | 25,869 | 32,174 | +24% | 1 | 1 | 0% | 5,113 | 6,200 | +21% | 0 | 0 | — |
case-03 | fail→pass | 24,826 | 29,196 | +18% | 1 | 1 | 0% | 4,953 | 6,530 | +32% | 0 | 0 | — |
case-04 | fail→pass | 8,835 | 28,658 | +224% | 1 | 1 | 0% | 1,486 | 5,931 | +299% | 0 | 0 | — |
case-05 | pass→pass | 13,002 | 16,556 | +27% | 1 | 1 | 0% | 2,541 | 3,780 | +49% | 0 | 0 | — |
case-06 | pass→pass | 15,460 | 13,283 | -14% | 1 | 1 | 0% | 3,453 | 3,213 | -7% | 0 | 0 | — |
case-07 | fail→pass | 12,060 | 14,995 | +24% | 1 | 1 | 0% | 2,388 | 3,358 | +41% | 0 | 0 | — |
case-08 | fail→fail | 10,997 | 14,883 | +35% | 1 | 1 | 0% | 1,731 | 2,983 | +72% | 0 | 0 | — |
case-09 | fail→pass | 18,105 | 12,960 | -28% | 1 | 1 | 0% | 2,806 | 2,538 | -10% | 0 | 0 | — |
case-10 | fail→fail | 21,186 | 30,208 | +43% | 1 | 1 | 0% | 3,682 | 6,508 | +77% | 0 | 0 | — |
case-11 | fail→fail | 21,441 | 24,695 | +15% | 1 | 1 | 0% | 3,803 | 5,453 | +43% | 0 | 0 | — |
case-12 | pass→pass | 20,484 | 18,852 | -8% | 1 | 1 | 0% | 3,466 | 3,726 | +8% | 0 | 0 | — |
case-13 | fail→fail | 17,945 | 29,845 | +66% | 1 | 1 | 0% | 2,644 | 6,500 | +146% | 0 | 0 | — |
case-14 | fail→pass | 22,236 | 30,700 | +38% | 1 | 1 | 0% | 3,931 | 6,498 | +65% | 0 | 0 | — |
case-15 | fail→pass | 15,483 | 12,172 | -21% | 1 | 1 | 0% | 2,718 | 2,433 | -10% | 0 | 0 | — |
case-16 | pass→pass | 13,829 | 29,368 | +112% | 1 | 1 | 0% | 2,475 | 6,445 | +160% | 0 | 0 | — |
case-17 | pass→pass | 21,633 | 30,053 | +39% | 1 | 1 | 0% | 4,313 | 6,498 | +51% | 0 | 0 | — |
case-22 | fail→fail | 28,629 | 29,276 | +2% | 1 | 1 | 0% | 6,165 | 6,502 | +5% | 0 | 0 | — |
case-18 | fail→pass | 21,126 | 22,443 | +6% | 1 | 1 | 0% | 3,923 | 4,914 | +25% | 0 | 0 | — |
case-19 | pass→pass | 25,394 | 30,135 | +19% | 1 | 1 | 0% | 3,791 | 6,507 | +72% | 0 | 0 | — |
case-20 | fail→pass | 19,136 | 20,729 | +8% | 1 | 1 | 0% | 3,123 | 4,134 | +32% | 0 | 0 | — |
case-21 | fail→fail | 21,586 | 19,795 | -8% | 1 | 1 | 0% | 3,778 | 4,016 | +6% | 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. 22 cases were attempted. The headline lift of +45 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.