Install any skill in seconds. Free to start, no credit card required.
Get Started Free →When the user wants to write or optimize short-form marketing copy—headlines, CTAs, ad copy, landing page copy, email copy. Also use when the user mentions "copywriting," "headline," "CTA copy," "ad copy," "landing page copy," "sales copy," "conversion copy," "PAS," "AIDA," "BAB," "copy formula," or "differentiation." For long-form article bodies (blog posts, guides), use article-content.
.claude/skills/kostja94-copywriting/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 42% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-21 | ✓→✗ | ▼ Worse | 79% | 0% |
| case-02 | ✓→✓ | = Same ✓ | 38% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 68% | 0% |
Guides short-form marketing copy—ads, landing pages, email, CTAs. Copywriting frameworks (PAS, AIDA, BAB) provide structure for conversion-focused copy. For article body content (blog posts, guides, long-form), see article-content.
When invoking: On first use, if helpful, open with 1–2 sentences on what this skill covers and why it matters, then provide the main output. On subsequent use or when the user asks to skip, go directly to the main output.
Check for project context first: If .claude/project-context.md or .cursor/project-context.md exists, read Sections 2 (Positioning), 3 (Value Proposition), 8 (Brand & Voice).
Identify:
| Framework | Structure | Best For | |-----------|-----------|----------| | PAS | Problem → Agitation → Solution | Direct-response; sales pages; landing pages | | AIDA | Attention → Interest → Desire → Action | Story-driven; ads; emails; editorials | | BAB | Before (problem) → After (outcome) → Bridge (how) | Transformation; testimonials; case studies | | FAB | Features → Advantages → Benefits | Product pages; feature lists | | 4 U's | Useful, Urgent, Unique, Ultra-specific | Headlines; subject lines |
| Formula | Example | |---------|---------| | How to outcome] | "How to Double Your Conversions in 30 Days" | | Number] Ways to benefit] | "7 Ways to Reduce Churn Without Hiring" | | Who] + benefit] | "For Marketers Who Want to Ship Faster" | | Problem]? Here's solution] | "Struggling with SEO? Here's the Fix" | | Before] → After] | "From 0 to 10K Users in 90 Days" | | The adjective] noun] for audience] | "The Ultimate Guide for SaaS Founders" |
Rules: Front-load keywords; keep under 60 chars for SERP; avoid clickbait; match ad-to-page alignment.
Information gain (differentiation): Copy that restates what competitors say adds zero value. Lead with unique angle, proprietary data, or contrarian insight. "What does this headline/promise add that others don't?" See article-content for full information gain strategy in long-form.
| Context | Focus | Skills | |---------|-------|--------| | Ad copy | Hook, benefit, CTA; platform limits | paid-ads-strategy, google-ads, meta-ads | | Landing page | Headline, value prop, CTA; above fold | landing-page-generator, hero-generator | | Email | Subject line, preview, body, CTA | email-marketing | | CTA | Value-focused; action verb; avoid "Submit" | cta-generator | | Article / blog | Headline formulas, CTA copy; body → article-content | article-content, article-page-generator |
| Element | Ad | Landing Page | |---------|-----|--------------| | Promise | Must match page headline | Same promise; expand on it | | Length | Concise; platform limits | Enough to explain value | | CTA | Click-through; "Learn More" | Conversion; "Start Free Trial" |
Avoid: Ad promise not visible on page; mismatch increases bounce.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 10,614 | 8,865 | -16% | 1 | 1 | 0% | 1,991 | 2,835 | +42% | 0 | 0 | — |
case-02 | pass→pass | 10,490 | 8,394 | -20% | 1 | 1 | 0% | 1,845 | 2,548 | +38% | 0 | 0 | — |
case-03 | pass→pass | 10,634 | 8,857 | -17% | 1 | 1 | 0% | 1,688 | 2,831 | +68% | 0 | 0 | — |
case-04 | pass→pass | 7,428 | 5,734 | -23% | 1 | 1 | 0% | 1,249 | 2,249 | +80% | 0 | 0 | — |
case-05 | pass→pass | 7,527 | 7,407 | -2% | 1 | 1 | 0% | 1,282 | 2,683 | +109% | 0 | 0 | — |
case-06 | pass→pass | 6,187 | 6,827 | +10% | 1 | 1 | 0% | 1,233 | 2,430 | +97% | 0 | 0 | — |
case-07 | pass→pass | 9,143 | 6,814 | -25% | 1 | 1 | 0% | 1,419 | 2,461 | +73% | 0 | 0 | — |
case-17 | pass→pass | 12,704 | 10,209 | -20% | 1 | 1 | 0% | 2,405 | 2,928 | +22% | 0 | 0 | — |
case-08 | pass→pass | 8,154 | 6,321 | -22% | 1 | 1 | 0% | 1,429 | 2,446 | +71% | 0 | 0 | — |
case-09 | pass→pass | 9,351 | 7,582 | -19% | 1 | 1 | 0% | 1,760 | 2,645 | +50% | 0 | 0 | — |
case-10 | pass→pass | 9,639 | 8,935 | -7% | 1 | 1 | 0% | 1,583 | 2,685 | +70% | 0 | 0 | — |
case-11 | pass→pass | 6,921 | 8,020 | +16% | 1 | 1 | 0% | 1,158 | 2,630 | +127% | 0 | 0 | — |
case-12 | pass→pass | 7,372 | 5,876 | -20% | 1 | 1 | 0% | 1,316 | 2,303 | +75% | 0 | 0 | — |
case-13 | pass→pass | 12,716 | 11,319 | -11% | 1 | 1 | 0% | 2,344 | 3,363 | +43% | 0 | 0 | — |
case-14 | pass→pass | 7,132 | 8,408 | +18% | 1 | 1 | 0% | 1,222 | 2,566 | +110% | 0 | 0 | — |
case-15 | pass→pass | 7,444 | 9,598 | +29% | 1 | 1 | 0% | 1,340 | 2,874 | +114% | 0 | 0 | — |
case-16 | pass→pass | 6,324 | 5,789 | -8% | 1 | 1 | 0% | 1,109 | 2,138 | +93% | 0 | 0 | — |
case-18 | fail→pass | 11,080 | 8,016 | -28% | 1 | 1 | 0% | 2,105 | 2,884 | +37% | 0 | 0 | — |
case-19 | pass→pass | 6,082 | 7,330 | +21% | 1 | 1 | 0% | 1,061 | 2,432 | +129% | 0 | 0 | — |
case-20 | pass→pass | 6,066 | 6,971 | +15% | 1 | 1 | 0% | 1,136 | 2,557 | +125% | 0 | 0 | — |
case-21 | pass→fail | 9,536 | 11,305 | +19% | 1 | 1 | 0% | 1,798 | 3,219 | +79% | 0 | 0 | — |
case-22 | pass→pass | 9,422 | 8,319 | -12% | 1 | 1 | 0% | 1,652 | 2,612 | +58% | 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 +5 percentage points is the difference between those two pass rates over the 22 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.