Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Draft professional emails based on context, tone, and recipient. Use for composing business emails.
.claude/skills/skrun-dev-email-drafter/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 48% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 44% | 0% |
You are a professional email writing assistant. Draft emails that are clear, appropriate for the tone, and include a compelling call to action.
Return a JSON object with:
subject: A concise, descriptive subject line (< 60 chars)body: The full email body (greeting, content, sign-off)call_to_action: The specific action you want the recipient to takeContext: "Follow up on proposal sent last week" Tone: formal Recipient: "VP of Engineering at Acme Corp" → Subject: "Follow-Up: Technical Proposal for Acme Corp" → Body: Professional follow-up with reference to key benefits → CTA: "Would you be available for a 30-minute call this Thursday?"
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 6,718 | 5,680 | -15% | 1 | 1 | 0% | 1,111 | 1,282 | +15% | 0 | 0 | — |
case-02 | fail→pass | 3,858 | 4,840 | +25% | 1 | 1 | 0% | 673 | 1,113 | +65% | 0 | 0 | — |
case-03 | fail→pass | 4,015 | 3,306 | -18% | 1 | 1 | 0% | 692 | 867 | +25% | 0 | 0 | — |
case-04 | fail→pass | 6,568 | 6,853 | +4% | 1 | 1 | 0% | 1,014 | 1,499 | +48% | 0 | 0 | — |
case-05 | pass→pass | 3,415 | 4,430 | +30% | 1 | 1 | 0% | 604 | 1,001 | +66% | 0 | 0 | — |
case-06 | pass→pass | 5,406 | 4,857 | -10% | 1 | 1 | 0% | 966 | 1,091 | +13% | 0 | 0 | — |
case-07 | pass→pass | 5,312 | 4,960 | -7% | 1 | 1 | 0% | 865 | 1,033 | +19% | 0 | 0 | — |
case-08 | pass→pass | 5,737 | 5,199 | -9% | 1 | 1 | 0% | 942 | 1,162 | +23% | 0 | 0 | — |
case-09 | pass→pass | 4,347 | 4,103 | -6% | 1 | 1 | 0% | 743 | 955 | +29% | 0 | 0 | — |
case-10 | fail→pass | 4,837 | 5,247 | +8% | 1 | 1 | 0% | 842 | 1,213 | +44% | 0 | 0 | — |
case-11 | fail→pass | 4,056 | 3,162 | -22% | 1 | 1 | 0% | 676 | 786 | +16% | 0 | 0 | — |
case-12 | fail→pass | 4,194 | 4,414 | +5% | 1 | 1 | 0% | 699 | 973 | +39% | 0 | 0 | — |
case-13 | fail→fail | 5,745 | 5,231 | -9% | 1 | 1 | 0% | 959 | 1,174 | +22% | 0 | 0 | — |
case-14 | fail→pass | 4,759 | 4,905 | +3% | 1 | 1 | 0% | 705 | 1,031 | +46% | 0 | 0 | — |
case-15 | fail→pass | 5,949 | 6,075 | +2% | 1 | 1 | 0% | 1,004 | 1,306 | +30% | 0 | 0 | — |
case-16 | pass→pass | 6,790 | 6,005 | -12% | 1 | 1 | 0% | 1,018 | 1,230 | +21% | 0 | 0 | — |
case-17 | fail→pass | 5,989 | 6,491 | +8% | 1 | 1 | 0% | 904 | 1,297 | +43% | 0 | 0 | — |
case-18 | pass→pass | 5,380 | 6,290 | +17% | 1 | 1 | 0% | 813 | 1,289 | +59% | 0 | 0 | — |
case-19 | fail→pass | 7,397 | 6,681 | -10% | 1 | 1 | 0% | 1,187 | 1,316 | +11% | 0 | 0 | — |
case-20 | fail→fail | 4,537 | 5,240 | +15% | 1 | 1 | 0% | 620 | 1,030 | +66% | 0 | 0 | — |
case-21 | pass→fail | 16,093 | 14,733 | -8% | 1 | 1 | 0% | 2,718 | 2,838 | +4% | 0 | 0 | — |
case-22 | pass→pass | 5,589 | 4,332 | -22% | 1 | 1 | 0% | 926 | 963 | +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 +45 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.