Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Write high-converting, punchy, and professional copy for landing pages, emails, and marketing materials.
.claude/skills/aiskillstore-copywriter/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | -30% | 0% |
You are the Lead Copywriter. Your words drive action. You write copy that is clear, compelling, and converts.
Every sentence must earn its place.
Don't sacrifice clarity for a pun. If the user has to think about what you mean, you've lost them.
H1: Action Verb] Benefit] without Pain Point]. Sub: Trusted by Number] companies. The Adjective] way to Goal]. CTA: Start Action]
Example: > Build your SaaS in a weekend. > Stop wasting time on setup. The production-ready starter kit for serious developers. > Get Started]
H3: Benefit-Driven Title] Body: One sentence explaining the problem. One sentence explaining the solution. Bullet: Specific metric or capability.
Example: > Authentication handled. > Security is hard. We make it easy. Pre-configured Auth.js with social logins, ready to go.
When asked to "Write copy" or "Improve this text":
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 12,516 | 7,817 | -38% | 1 | 1 | 0% | 2,049 | 1,803 | -12% | 0 | 0 | — |
case-02 | fail→fail | 9,986 | 7,676 | -23% | 1 | 1 | 0% | 1,537 | 1,837 | +20% | 0 | 0 | — |
case-03 | fail→fail | 7,496 | 6,432 | -14% | 1 | 1 | 0% | 1,288 | 1,589 | +23% | 0 | 0 | — |
case-04 | fail→fail | 10,046 | 7,328 | -27% | 1 | 1 | 0% | 1,681 | 1,702 | +1% | 0 | 0 | — |
case-05 | pass→pass | 12,621 | 5,923 | -53% | 1 | 1 | 0% | 2,136 | 1,504 | -30% | 0 | 0 | — |
case-06 | fail→pass | 9,932 | 5,544 | -44% | 1 | 1 | 0% | 1,438 | 1,393 | -3% | 0 | 0 | — |
case-07 | pass→pass | 12,388 | 7,098 | -43% | 1 | 1 | 0% | 1,942 | 1,604 | -17% | 0 | 0 | — |
case-08 | fail→pass | 7,682 | 7,061 | -8% | 1 | 1 | 0% | 1,146 | 1,682 | +47% | 0 | 0 | — |
case-09 | fail→fail | 8,919 | 5,612 | -37% | 1 | 1 | 0% | 1,473 | 1,435 | -3% | 0 | 0 | — |
case-10 | pass→pass | 6,660 | 5,408 | -19% | 1 | 1 | 0% | 1,104 | 1,326 | +20% | 0 | 0 | — |
case-11 | pass→pass | 11,689 | 9,312 | -20% | 1 | 1 | 0% | 1,782 | 1,993 | +12% | 0 | 0 | — |
case-12 | fail→pass | 8,199 | 5,588 | -32% | 1 | 1 | 0% | 1,274 | 1,436 | +13% | 0 | 0 | — |
case-13 | pass→pass | 7,963 | 4,607 | -42% | 1 | 1 | 0% | 1,273 | 1,198 | -6% | 0 | 0 | — |
case-14 | pass→pass | 6,339 | 6,686 | +5% | 1 | 1 | 0% | 967 | 1,592 | +65% | 0 | 0 | — |
case-15 | pass→pass | 7,617 | 7,110 | -7% | 1 | 1 | 0% | 1,125 | 1,539 | +37% | 0 | 0 | — |
case-16 | fail→pass | 7,606 | 5,773 | -24% | 1 | 1 | 0% | 1,170 | 1,402 | +20% | 0 | 0 | — |
case-17 | pass→pass | 12,627 | 7,695 | -39% | 1 | 1 | 0% | 1,946 | 1,698 | -13% | 0 | 0 | — |
case-18 | pass→pass | 9,561 | 7,977 | -17% | 1 | 1 | 0% | 1,373 | 1,708 | +24% | 0 | 0 | — |
case-19 | pass→pass | 14,542 | 8,602 | -41% | 1 | 1 | 0% | 2,150 | 1,870 | -13% | 0 | 0 | — |
case-20 | pass→pass | 15,914 | 11,351 | -29% | 1 | 1 | 0% | 3,306 | 2,721 | -18% | 0 | 0 | — |
case-21 | pass→pass | 13,192 | 10,258 | -22% | 1 | 1 | 0% | 2,251 | 2,088 | -7% | 0 | 0 | — |
case-22 | pass→pass | 12,738 | 7,323 | -43% | 1 | 1 | 0% | 1,864 | 1,658 | -11% | 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 +18 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.