Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build a quarterly business review from a client folder -- reports, usage exports, support tickets, meeting notes, emails. Extracts delivered value with receipts, surfaces risks before the client does, and drafts the QBR narrative plus expansion asks. For agencies, consultancies, and CS teams.
.claude/skills/onewave-ai-cowork-qbr-builder/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 38% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 373% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 241% | 0% |
Prepare a QBR the way a top customer-success lead does: prove the value delivered with specifics, name the problems before the client names them, and earn the expansion conversation. Input: a folder of client artifacts for the quarter -- status reports, deliverables, usage/analytics exports, support tickets, meeting notes, email threads -- plus the engagement's stated goals if documented.
qbr-[client]-[quarter].md structured for a deck: executive summary, goals scorecard, wins with receipts, misses with fixes, next-quarter plan, and -- only where the evidence supports it -- 1-2 expansion recommendations framed around the client's goals, not your revenue. Hand to pptx or presentation-design-enhancer for the deck build.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 18,364 | 41,770 | +127% | 1 | 1 | 0% | 2,936 | 6,512 | +122% | 0 | 0 | — |
case-02 | fail→fail | 35,290 | 4,841 | -86% | 1 | 1 | 0% | 5,013 | 898 | -82% | 0 | 0 | — |
case-03 | fail→pass | 12,431 | 12,401 | -0% | 1 | 1 | 0% | 1,812 | 2,495 | +38% | 0 | 0 | — |
case-04 | fail→pass | 14,748 | 14,914 | +1% | 1 | 1 | 0% | 2,236 | 3,014 | +35% | 0 | 0 | — |
case-05 | fail→pass | 8,855 | 5,791 | -35% | 1 | 1 | 0% | 1,407 | 1,638 | +16% | 0 | 0 | — |
case-06 | fail→fail | 10,245 | 7,461 | -27% | 1 | 1 | 0% | 1,561 | 1,880 | +20% | 0 | 0 | — |
case-07 | pass→pass | 9,331 | 10,396 | +11% | 1 | 1 | 0% | 1,393 | 2,151 | +54% | 0 | 0 | — |
case-08 | pass→pass | 9,189 | 7,033 | -23% | 1 | 1 | 0% | 1,354 | 1,743 | +29% | 0 | 0 | — |
case-09 | pass→pass | 10,072 | 3,922 | -61% | 1 | 1 | 0% | 1,178 | 1,245 | +6% | 0 | 0 | — |
case-10 | fail→pass | 3,395 | 11,486 | +238% | 1 | 1 | 0% | 513 | 2,424 | +373% | 0 | 0 | — |
case-11 | pass→pass | 13,282 | 15,902 | +20% | 1 | 1 | 0% | 1,959 | 2,745 | +40% | 0 | 0 | — |
case-12 | fail→pass | 6,356 | 15,477 | +144% | 1 | 1 | 0% | 845 | 2,879 | +241% | 0 | 0 | — |
case-13 | fail→pass | 12,600 | 13,259 | +5% | 1 | 1 | 0% | 1,840 | 2,417 | +31% | 0 | 0 | — |
case-14 | fail→pass | 3,396 | 22,004 | +548% | 1 | 1 | 0% | 143 | 4,179 | +2822% | 0 | 0 | — |
case-15 | fail→pass | 7,238 | 9,479 | +31% | 1 | 1 | 0% | 1,069 | 2,071 | +94% | 0 | 0 | — |
case-16 | pass→pass | 10,621 | 9,409 | -11% | 1 | 1 | 0% | 1,539 | 1,912 | +24% | 0 | 0 | — |
case-17 | pass→pass | 6,832 | 6,073 | -11% | 1 | 1 | 0% | 991 | 1,573 | +59% | 0 | 0 | — |
case-18 | pass→pass | 13,539 | 11,246 | -17% | 1 | 1 | 0% | 1,850 | 2,233 | +21% | 0 | 0 | — |
case-19 | pass→pass | 11,693 | 10,530 | -10% | 1 | 1 | 0% | 1,744 | 2,191 | +26% | 0 | 0 | — |
case-20 | fail→pass | 3,426 | 9,370 | +173% | 1 | 1 | 0% | 540 | 1,927 | +257% | 0 | 0 | — |
case-21 | fail→fail | 20,139 | 20,518 | +2% | 1 | 1 | 0% | 2,991 | 3,595 | +20% | 0 | 0 | — |
case-22 | pass→fail | 22,792 | 13,059 | -43% | 1 | 1 | 0% | 4,211 | 3,060 | -27% | 0 | 0 | — |
case-23 | fail→fail | 15,807 | 17,435 | +10% | 1 | 1 | 0% | 2,523 | 3,369 | +34% | 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. 23 cases were attempted, and 21 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +35 percentage points is the difference between those two pass rates over the 21 comparable cases. 2 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.