Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Simulate the due-diligence call where an acquirer's or investor's analyst takes your metrics apart — the questions behind the spreadsheet, the moment a number wobbles, and a debrief on which answers create risk. Use when asked simulate due diligence on my startup, stress-test my metrics before the raise, what will the acquirer's analyst ask, or prep me for the DD call. Produces the call transcript with the analyst's private notes, the internal memo they write afterward, and a debrief separating
.claude/skills/mohitagw15856-the-due-diligence-call/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 42% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 153% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 97% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 92% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 90% | 0% |
Founders present metrics; analysts reconcile them. The due-diligence call is where "150% net revenue retention" meets "walk me through the cohort math," and where numbers that were directionally fine become credibility problems because they don't tie to each other. This skill runs that call early — an analyst who has already read the data room, asking the questions the spreadsheet raised — then writes the internal memo and debriefs on the difference between numbers that need better presentation and numbers that need a better business. Honesty is structural: the exercise finds the wobbles so they can be fixed or disclosed, never dressed.
Ask for these if not provided:
> Simulation — a plausible adversarial reading, not a prediction.
The call, grounded ONLY in supplied numbers; where a metric can't be reconciled from what was given, the analyst notices — that's the finding. Analyst's note: after each exchange.]
To: partner/corp-dev. Recommendation with the real reasons — including the difference between what was said on the call and what the memo flags.]
| Wobble found | Category | The move | |---|---|---| | … | Presentation-fixable | Recompute and present both definitions side by side | | … | Disclose-and-explain | The disclosure sentence, drafted | | … | Fix-the-business | No sentence fixes this; here's what would |
Plus: the reconciliation pack to build before the real call — the 5 numbers that must tie]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 56,485 | 65,308 | +16% | 1 | 1 | 0% | 8,411 | 9,436 | +12% | 0 | 0 | — |
case-02 | fail→fail | 56,325 | 44,127 | -22% | 1 | 1 | 0% | 6,816 | 7,909 | +16% | 0 | 0 | — |
case-03 | fail→fail | 46,044 | 37,419 | -19% | 1 | 1 | 0% | 6,814 | 7,464 | +10% | 0 | 0 | — |
case-04 | pass→fail | 29,379 | 58,184 | +98% | 1 | 1 | 0% | 3,889 | 9,302 | +139% | 0 | 0 | — |
case-05 | pass→fail | 29,717 | 61,044 | +105% | 1 | 1 | 0% | 3,632 | 9,300 | +156% | 0 | 0 | — |
case-06 | pass→fail | 22,282 | 60,636 | +172% | 1 | 1 | 0% | 3,290 | 9,313 | +183% | 0 | 0 | — |
case-07 | fail→pass | 21,219 | 16,507 | -22% | 1 | 1 | 0% | 2,423 | 3,441 | +42% | 0 | 0 | — |
case-08 | fail→pass | 28,046 | 61,083 | +118% | 1 | 1 | 0% | 3,216 | 8,145 | +153% | 0 | 0 | — |
case-09 | fail→pass | 29,278 | 49,347 | +69% | 1 | 1 | 0% | 3,620 | 7,137 | +97% | 0 | 0 | — |
case-10 | fail→pass | 45,418 | 43,994 | -3% | 1 | 1 | 0% | 3,684 | 7,060 | +92% | 0 | 0 | — |
case-11 | pass→pass | 24,105 | 55,768 | +131% | 1 | 1 | 0% | 2,484 | 7,757 | +212% | 0 | 0 | — |
case-12 | fail→fail | 28,994 | 41,394 | +43% | 1 | 1 | 0% | 3,621 | 6,420 | +77% | 0 | 0 | — |
case-13 | fail→pass | 30,720 | 39,754 | +29% | 1 | 1 | 0% | 3,214 | 6,096 | +90% | 0 | 0 | — |
case-14 | fail→pass | 29,960 | 47,138 | +57% | 1 | 1 | 0% | 3,078 | 7,630 | +148% | 0 | 0 | — |
case-15 | fail→pass | 21,073 | 42,540 | +102% | 1 | 1 | 0% | 2,151 | 6,700 | +211% | 0 | 0 | — |
case-16 | fail→pass | 41,242 | 38,220 | -7% | 1 | 1 | 0% | 3,283 | 5,891 | +79% | 0 | 0 | — |
case-17 | fail→pass | 31,339 | 42,032 | +34% | 1 | 1 | 0% | 3,863 | 7,867 | +104% | 0 | 0 | — |
case-18 | fail→fail | 20,279 | 52,501 | +159% | 1 | 1 | 0% | 2,694 | 8,952 | +232% | 0 | 0 | — |
case-19 | fail→pass | 16,939 | 18,186 | +7% | 1 | 1 | 0% | 1,905 | 3,242 | +70% | 0 | 0 | — |
case-20 | fail→fail | 23,776 | 52,240 | +120% | 1 | 1 | 0% | 2,840 | 7,997 | +182% | 0 | 0 | — |
case-21 | fail→fail | 21,384 | 25,061 | +17% | 1 | 1 | 0% | 2,267 | 5,116 | +126% | 0 | 0 | — |
case-22 | fail→pass | 32,685 | 25,991 | -20% | 1 | 1 | 0% | 2,983 | 4,270 | +43% | 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 +36 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 cases got worse with the skill loaded, and they are 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.