Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create evidence-based Meta ad creative briefs, copy angles, and testable asset concepts for Facebook and Instagram. Use when asked for Meta ad creative, Facebook ads copy, Instagram ad assets, UGC concepts, creative briefs, new ad concepts, or creative refreshes.
.claude/skills/nowork-studio-meta-ads-creative/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -17% | 0% |
Read ../../meta-ads/creative/SKILL.md completely, then follow it as the active workflow. Normalize that path from the directory containing this wrapper: the canonical file is <plugin-root>/meta-ads/creative/SKILL.md, not <plugin-root>/skills/meta-ads-creative/creative/SKILL.md. Resolve every relative reference from the canonical file against <plugin-root>/meta-ads/creative/. If the canonical file cannot be read, stop and report the packaging error; never substitute a similarly named skill from another plugin.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 30,809 | 36,059 | +17% | 1 | 1 | 0% | 4,604 | 5,189 | +13% | 0 | 0 | — |
case-02 | fail→fail | 19,690 | 22,009 | +12% | 1 | 1 | 0% | 2,924 | 3,176 | +9% | 0 | 0 | — |
case-03 | fail→fail | 32,584 | 22,631 | -31% | 1 | 1 | 0% | 3,193 | 3,270 | +2% | 0 | 0 | — |
case-04 | pass→pass | 15,794 | 9,676 | -39% | 1 | 1 | 0% | 1,383 | 1,747 | +26% | 0 | 0 | — |
case-05 | pass→pass | 19,626 | 16,467 | -16% | 1 | 1 | 0% | 2,920 | 3,210 | +10% | 0 | 0 | — |
case-06 | pass→pass | 15,187 | 11,798 | -22% | 1 | 1 | 0% | 2,548 | 1,804 | -29% | 0 | 0 | — |
case-07 | fail→fail | 15,238 | 3,542 | -77% | 1 | 1 | 0% | 2,256 | 509 | -77% | 0 | 0 | — |
case-08 | fail→pass | 5,442 | 3,378 | -38% | 1 | 1 | 0% | 697 | 590 | -15% | 0 | 0 | — |
case-09 | fail→pass | 9,293 | 3,131 | -66% | 1 | 1 | 0% | 1,351 | 509 | -62% | 0 | 0 | — |
case-10 | fail→pass | 10,566 | 5,414 | -49% | 1 | 1 | 0% | 1,500 | 636 | -58% | 0 | 0 | — |
case-11 | fail→pass | 5,725 | 38,594 | +574% | 1 | 1 | 0% | 866 | 759 | -12% | 0 | 0 | — |
case-12 | fail→fail | 15,827 | 19,798 | +25% | 1 | 1 | 0% | 2,103 | 2,642 | +26% | 0 | 0 | — |
case-13 | fail→fail | 12,543 | 11,362 | -9% | 1 | 1 | 0% | 1,641 | 2,058 | +25% | 0 | 0 | — |
case-14 | fail→fail | 17,797 | 16,638 | -7% | 1 | 1 | 0% | 2,381 | 2,631 | +10% | 0 | 0 | — |
case-15 | fail→fail | 18,602 | 17,342 | -7% | 1 | 1 | 0% | 2,336 | 2,428 | +4% | 0 | 0 | — |
case-16 | fail→fail | 14,845 | 18,010 | +21% | 1 | 1 | 0% | 1,952 | 2,413 | +24% | 0 | 0 | — |
case-17 | pass→pass | 12,036 | 3,476 | -71% | 1 | 1 | 0% | 1,491 | 539 | -64% | 0 | 0 | — |
case-18 | fail→pass | 5,249 | 4,623 | -12% | 1 | 1 | 0% | 804 | 667 | -17% | 0 | 0 | — |
case-19 | fail→pass | 13,092 | 3,842 | -71% | 1 | 1 | 0% | 1,977 | 535 | -73% | 0 | 0 | — |
case-20 | fail→pass | 15,387 | 4,593 | -70% | 1 | 1 | 0% | 2,333 | 841 | -64% | 0 | 0 | — |
case-21 | fail→fail | 14,331 | 12,131 | -15% | 1 | 1 | 0% | 1,957 | 1,763 | -10% | 0 | 0 | — |
case-22 | fail→fail | 16,292 | 20,828 | +28% | 1 | 1 | 0% | 2,434 | 2,865 | +18% | 0 | 0 | — |
case-23 | fail→fail | 14,543 | 18,560 | +28% | 1 | 1 | 0% | 1,937 | 2,633 | +36% | 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 +30 percentage points is the difference between those two pass rates over the 23 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 9/9/2026 | +43% |
| gemini-3.6-flash | verified | 8/29/2026 | +14% |
Other measured skills in the registry, with their headline benchmark lift.