Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A brand product-launch email — masthead with wordmark, hero image block, headline lockup with skewed-italic accent, body copy, primary CTA, and a specifications grid. Pure HTML email layout (centered single column, table fallback). Use when the brief asks for an "email", "newsletter blast", "MJML", "product launch email", or "email template".
.claude/skills/nexu-io-email-marketing/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 7% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 30% | 0% |
| case-06 | ✓→✗ | ▼ Worse | 207% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 47% | 0% |
Produce a single HTML email — centered, single column, no chrome around the email body. Treat it like a marketing artifact: one big idea, one CTA.
font more than any other surface — pick the loudest type token in the DS for the headline lockup.
real product name, and one real benefit sentence — no placeholders.
page background (so the email body looks like an email, not the page):
MEMBERS) on the right. Thin underline.
gradient or a stylized SVG silhouette of the product (shoe, bottle, headphones, whatever the brief implies). Add a tiny brand stamp on the top-left and a colorway tag on the bottom-left.
· characters(e.g. "NEW · MAX-CUSHION TRAINER · EMBER FLARE").
extra-tight tracking. Apply a slight skew (transform: skew(-6deg)) on one accent word to give it a sporty parallelogram feel.
the display font for the numbers.
<!doctype html> through </html>, CSS inline.margin: 0 auto. Set body { background: <tint> }so the email-on-page metaphor reads.
product photo.
data-od-id on the masthead, hero, headline, CTA, specs.Emit between <artifact> tags:
<artifact identifier="email-slug" type="text/html" title="Email — Subject Line">
<!doctype html>
<html>...</html>
</artifact>One sentence before the artifact, nothing after.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 25,141 | 22,494 | -11% | 1 | 1 | 0% | 6,001 | 6,397 | +7% | 0 | 0 | — |
case-02 | fail→fail | 34,894 | 26,229 | -25% | 1 | 1 | 0% | 6,214 | 6,926 | +11% | 0 | 0 | — |
case-03 | fail→pass | 25,931 | 23,266 | -10% | 1 | 1 | 0% | 6,201 | 6,133 | -1% | 0 | 0 | — |
case-04 | pass→pass | 22,735 | 22,215 | -2% | 1 | 1 | 0% | 3,926 | 5,764 | +47% | 0 | 0 | — |
case-05 | pass→pass | 12,261 | 25,759 | +110% | 1 | 1 | 0% | 2,328 | 6,876 | +195% | 0 | 0 | — |
case-06 | pass→fail | 11,092 | 27,779 | +150% | 1 | 1 | 0% | 2,243 | 6,876 | +207% | 0 | 0 | — |
case-07 | pass→pass | 24,028 | 29,234 | +22% | 1 | 1 | 0% | 5,442 | 6,893 | +27% | 0 | 0 | — |
case-08 | pass→pass | 26,066 | 27,047 | +4% | 1 | 1 | 0% | 6,181 | 6,893 | +12% | 0 | 0 | — |
case-09 | fail→pass | 25,481 | 27,229 | +7% | 1 | 1 | 0% | 6,183 | 6,639 | +7% | 0 | 0 | — |
case-10 | fail→fail | 18,333 | 22,569 | +23% | 1 | 1 | 0% | 3,991 | 6,099 | +53% | 0 | 0 | — |
case-11 | fail→fail | 22,101 | 26,181 | +18% | 1 | 1 | 0% | 5,093 | 6,802 | +34% | 0 | 0 | — |
case-12 | fail→fail | 25,745 | 26,589 | +3% | 1 | 1 | 0% | 5,807 | 6,198 | +7% | 0 | 0 | — |
case-13 | fail→fail | 27,629 | 32,598 | +18% | 1 | 1 | 0% | 4,827 | 6,893 | +43% | 0 | 0 | — |
case-14 | fail→fail | 27,152 | 23,921 | -12% | 1 | 1 | 0% | 6,176 | 6,148 | -0% | 0 | 0 | — |
case-15 | fail→fail | 32,935 | 27,392 | -17% | 1 | 1 | 0% | 5,985 | 6,878 | +15% | 0 | 0 | — |
case-16 | fail→fail | 26,579 | 28,800 | +8% | 1 | 1 | 0% | 6,169 | 6,881 | +12% | 0 | 0 | — |
case-17 | fail→fail | 27,181 | 26,389 | -3% | 1 | 1 | 0% | 6,175 | 6,887 | +12% | 0 | 0 | — |
case-18 | fail→fail | 28,317 | 24,800 | -12% | 1 | 1 | 0% | 6,173 | 6,610 | +7% | 0 | 0 | — |
case-19 | fail→fail | 28,125 | 26,640 | -5% | 1 | 1 | 0% | 6,179 | 6,891 | +12% | 0 | 0 | — |
case-20 | fail→pass | 19,727 | 23,438 | +19% | 1 | 1 | 0% | 4,604 | 5,998 | +30% | 0 | 0 | — |
case-21 | fail→fail | 17,148 | 20,038 | +17% | 1 | 1 | 0% | 3,578 | 5,372 | +50% | 0 | 0 | — |
case-22 | pass→pass | 26,369 | 25,257 | -4% | 1 | 1 | 0% | 6,176 | 6,445 | +4% | 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 +9 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.