Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create grounded paid-media creative concepts, copy angles, refresh hypotheses, and measurable test briefs across advertising platforms. Use for ad copy, creative briefs, UGC concepts, headlines, creative fatigue, asset ideas, or a creative refresh.
.claude/skills/nowork-studio-paid-ads-creative/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -65% | 0% |
Read ../shared/operating-contract.md and ../shared/measurement-framework.md. Route platform execution to /notfair:google-ads-copy, /notfair:google-ads-assets, /notfair:meta-ads-creative, /notfair:paid-ads-x, /notfair:paid-ads-linkedin, /notfair:paid-ads-reddit, or /notfair:paid-ads-tiktok as appropriate.
Start with a verified source: winning ads, search terms, customer research, approved landing-page claims, sales calls, reviews with permission, or a documented offer. Match each promise to its landing destination. Do not fabricate testimonials, ratings, prices, outcomes, availability, ownership, or product proof.
Deliver three distinct ready_for_review concept cards. Each card contains:
needs_substantiation;Do not make cosmetic variants of the same idea. Test one strategic variable at a time: hook, proof, format, angle, audience, or destination. Diagnose fatigue with complete-window frequency, link CTR, CPM, and delivery context before recommending a refresh.
Validate current platform specifications and policies from a verified source before publishing. A brief or generated image is not proof that a claim is permitted or that the platform accepted the asset. For platforms without a verified NotFair mutation surface, output an operator-ready brief and clear handoff rather than promising an upload.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 75,573 | 24,134 | -68% | 1 | 1 | 0% | 7,827 | 4,339 | -45% | 0 | 0 | — |
case-02 | fail→pass | 33,487 | 32,532 | -3% | 1 | 1 | 0% | 5,199 | 5,839 | +12% | 0 | 0 | — |
case-03 | fail→pass | 42,523 | 31,948 | -25% | 1 | 1 | 0% | 6,924 | 5,508 | -20% | 0 | 0 | — |
case-04 | pass→pass | 20,461 | 19,518 | -5% | 1 | 1 | 0% | 3,335 | 3,738 | +12% | 0 | 0 | — |
case-05 | pass→pass | 21,607 | 19,331 | -11% | 1 | 1 | 0% | 3,329 | 3,489 | +5% | 0 | 0 | — |
case-06 | pass→pass | 29,615 | 25,575 | -14% | 1 | 1 | 0% | 2,310 | 4,355 | +89% | 0 | 0 | — |
case-07 | fail→pass | 7,970 | 2,945 | -63% | 1 | 1 | 0% | 1,167 | 812 | -30% | 0 | 0 | — |
case-08 | fail→pass | 17,434 | 4,299 | -75% | 1 | 1 | 0% | 2,420 | 838 | -65% | 0 | 0 | — |
case-09 | fail→pass | 11,449 | 3,652 | -68% | 1 | 1 | 0% | 1,626 | 846 | -48% | 0 | 0 | — |
case-10 | fail→pass | 12,648 | 4,358 | -66% | 1 | 1 | 0% | 1,695 | 997 | -41% | 0 | 0 | — |
case-11 | fail→pass | 10,460 | 4,621 | -56% | 1 | 1 | 0% | 1,373 | 965 | -30% | 0 | 0 | — |
case-12 | fail→pass | 18,421 | 26,273 | +43% | 1 | 1 | 0% | 2,709 | 3,611 | +33% | 0 | 0 | — |
case-13 | fail→pass | 13,796 | 4,301 | -69% | 1 | 1 | 0% | 2,273 | 977 | -57% | 0 | 0 | — |
case-14 | pass→pass | 11,687 | 5,406 | -54% | 1 | 1 | 0% | 1,659 | 1,178 | -29% | 0 | 0 | — |
case-15 | pass→pass | 8,770 | 23,952 | +173% | 1 | 1 | 0% | 1,254 | 4,500 | +259% | 0 | 0 | — |
case-16 | pass→fail | 22,479 | 9,873 | -56% | 1 | 1 | 0% | 1,378 | 808 | -41% | 0 | 0 | — |
case-17 | pass→pass | 14,437 | 24,642 | +71% | 1 | 1 | 0% | 2,138 | 4,153 | +94% | 0 | 0 | — |
case-18 | pass→pass | 17,867 | 10,561 | -41% | 1 | 1 | 0% | 2,490 | 1,850 | -26% | 0 | 0 | — |
case-19 | fail→fail | 5,984 | 5,508 | -8% | 1 | 1 | 0% | 765 | 1,140 | +49% | 0 | 0 | — |
case-20 | pass→pass | 11,742 | 9,989 | -15% | 1 | 1 | 0% | 1,553 | 1,513 | -3% | 0 | 0 | — |
case-21 | pass→pass | 18,812 | 11,312 | -40% | 1 | 1 | 0% | 1,560 | 1,806 | +16% | 0 | 0 | — |
case-22 | pass→pass | 15,954 | 11,980 | -25% | 1 | 1 | 0% | 2,110 | 1,977 | -6% | 0 | 0 | — |
case-23 | pass→pass | 13,396 | 3,521 | -74% | 1 | 1 | 0% | 1,685 | 771 | -54% | 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 +39 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/29/2026 | +16% |
Other measured skills in the registry, with their headline benchmark lift.