Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Write a cold or warm-intro email to an investor that actually gets a reply — short, specific, traction-forward, with a clear ask. Use when asked to email an investor, write a fundraising outreach, request a warm intro, or craft a forwardable blurb. Produces a tight cold email, a forwardable intro blurb a mutual contact can paste, and the follow-up — all skimmable on a phone.
.claude/skills/mohitagw15856-investor-cold-email/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-20 | ✗→✓ | ▲ Improved | 73% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 85% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -42% | 0% |
Investors skim outreach on their phone in seconds. The emails that get replies are short, lead with the most credible proof, and make one clear ask. This skill writes them.
Given a rough company description, write the full email anyway and flag invented metrics (assumed — replace with real). Keep it ruthlessly short. Never leave placeholders an investor would see.
Ask for (if not already provided):
Acme — $30k MRR, growing 25% MoM, raising seed)A 3–4 sentence paragraph the mutual contact can paste with zero editing — written so it makes them look good for forwarding it.
A 2-line nudge to send if there's no reply in ~5 business days — adds a new data point (a milestone, a new customer), never just "bumping this."
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-20 | fail→pass | 7,706 | 8,993 | +17% | 1 | 1 | 0% | 1,270 | 2,200 | +73% | 0 | 0 | — |
case-01 | fail→fail | 12,643 | 7,364 | -42% | 1 | 1 | 0% | 2,272 | 2,091 | -8% | 0 | 0 | — |
case-02 | fail→pass | 9,060 | 7,438 | -18% | 1 | 1 | 0% | 1,585 | 2,145 | +35% | 0 | 0 | — |
case-03 | fail→fail | 10,366 | 6,772 | -35% | 1 | 1 | 0% | 1,881 | 1,878 | -0% | 0 | 0 | — |
case-04 | pass→pass | 11,370 | 11,196 | -2% | 1 | 1 | 0% | 1,853 | 2,619 | +41% | 0 | 0 | — |
case-05 | pass→pass | 14,824 | 10,024 | -32% | 1 | 1 | 0% | 2,426 | 2,172 | -10% | 0 | 0 | — |
case-06 | pass→pass | 21,250 | 10,185 | -52% | 1 | 1 | 0% | 3,264 | 2,179 | -33% | 0 | 0 | — |
case-07 | fail→pass | 11,392 | 9,039 | -21% | 1 | 1 | 0% | 1,785 | 2,116 | +19% | 0 | 0 | — |
case-08 | fail→pass | 6,038 | 8,790 | +46% | 1 | 1 | 0% | 1,140 | 2,112 | +85% | 0 | 0 | — |
case-09 | fail→fail | 8,402 | 9,437 | +12% | 1 | 1 | 0% | 1,437 | 2,280 | +59% | 0 | 0 | — |
case-10 | fail→fail | 9,775 | 7,771 | -21% | 1 | 1 | 0% | 1,702 | 2,016 | +18% | 0 | 0 | — |
case-11 | pass→pass | 6,355 | 3,805 | -40% | 1 | 1 | 0% | 1,110 | 1,206 | +9% | 0 | 0 | — |
case-12 | fail→pass | 20,838 | 8,005 | -62% | 1 | 1 | 0% | 3,571 | 2,055 | -42% | 0 | 0 | — |
case-13 | fail→pass | 7,626 | 8,503 | +12% | 1 | 1 | 0% | 1,313 | 2,111 | +61% | 0 | 0 | — |
case-14 | fail→fail | 6,988 | 11,465 | +64% | 1 | 1 | 0% | 1,215 | 2,610 | +115% | 0 | 0 | — |
case-15 | fail→fail | 9,525 | 9,403 | -1% | 1 | 1 | 0% | 1,609 | 2,260 | +40% | 0 | 0 | — |
case-16 | fail→fail | 7,653 | 8,212 | +7% | 1 | 1 | 0% | 1,410 | 2,024 | +44% | 0 | 0 | — |
case-17 | pass→pass | 8,567 | 8,482 | -1% | 1 | 1 | 0% | 1,594 | 2,058 | +29% | 0 | 0 | — |
case-18 | pass→pass | 12,065 | 8,304 | -31% | 1 | 1 | 0% | 1,933 | 2,201 | +14% | 0 | 0 | — |
case-19 | fail→pass | 9,301 | 11,014 | +18% | 1 | 1 | 0% | 1,667 | 2,523 | +51% | 0 | 0 | — |
case-21 | fail→pass | 5,314 | 4,230 | -20% | 1 | 1 | 0% | 957 | 1,331 | +39% | 0 | 0 | — |
case-22 | fail→pass | 12,131 | 9,004 | -26% | 1 | 1 | 0% | 2,039 | 2,224 | +9% | 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 +41 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.