Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Run meetings that actually decide — the pre-read-then-decide format, the options-on-the-table rule, the decider named before debate starts, and the recorded-or-it-didn't-happen close. Use when asked run this decision meeting, we discuss forever and never decide, structure the meeting where we pick the vendor/plan/design, or why do our decisions get relitigated. Produces the meeting design: pre-read, the in-room sequence, the decision rule, and the recording that makes it stick.
.claude/skills/mohitagw15856-decision-meeting-format/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 9% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 62% | 0% |
Meetings fail to decide for structural reasons, not character ones: the options were never written down (so debate invents them live), the decider was never named (so consensus is assumed and never arrives), and the decision was never recorded (so it gets relitigated by everyone who remembers it differently). The format fixes all three before the room opens: a pre-read with the options steelmanned, the decider and decision rule announced in the invite, a timeboxed sequence that separates clarifying from advocating, and a close where the decision — with its why and its dissent — lands in the log before anyone leaves.
Ask for these if not provided:
Options steelmanned + recommendation · the read-by expectation]
Clarify (Q-only, X min) → Advocate (timeboxed, dissent solicited) → Decide (in-room, or deferred-with-date)]
Decision · why (2 sentences) · dissent noted · owner + date · logged at (location) · reopening rule]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→fail | 15,772 | 13,284 | -16% | 1 | 1 | 0% | 2,481 | 3,093 | +25% | 0 | 0 | — |
case-01 | fail→pass | 16,728 | 14,447 | -14% | 1 | 1 | 0% | 3,077 | 3,363 | +9% | 0 | 0 | — |
case-02 | fail→pass | 17,753 | 17,917 | +1% | 1 | 1 | 0% | 3,066 | 4,113 | +34% | 0 | 0 | — |
case-03 | fail→fail | 17,595 | 15,577 | -11% | 1 | 1 | 0% | 2,966 | 3,783 | +28% | 0 | 0 | — |
case-11 | fail→pass | 12,870 | 12,778 | -1% | 1 | 1 | 0% | 2,352 | 3,280 | +39% | 0 | 0 | — |
case-05 | pass→fail | 10,675 | 13,710 | +28% | 1 | 1 | 0% | 1,817 | 3,320 | +83% | 0 | 0 | — |
case-06 | pass→fail | 21,602 | 22,861 | +6% | 1 | 1 | 0% | 4,147 | 5,099 | +23% | 0 | 0 | — |
case-07 | fail→fail | 17,422 | 13,922 | -20% | 1 | 1 | 0% | 2,829 | 3,310 | +17% | 0 | 0 | — |
case-08 | fail→pass | 14,654 | 14,535 | -1% | 1 | 1 | 0% | 2,636 | 3,386 | +28% | 0 | 0 | — |
case-09 | fail→fail | 19,321 | 15,332 | -21% | 1 | 1 | 0% | 3,339 | 3,792 | +14% | 0 | 0 | — |
case-10 | fail→pass | 11,131 | 12,148 | +9% | 1 | 1 | 0% | 2,030 | 3,290 | +62% | 0 | 0 | — |
case-12 | fail→pass | 15,311 | 12,299 | -20% | 1 | 1 | 0% | 2,461 | 3,256 | +32% | 0 | 0 | — |
case-13 | fail→pass | 15,180 | 16,618 | +9% | 1 | 1 | 0% | 2,712 | 3,831 | +41% | 0 | 0 | — |
case-14 | fail→pass | 15,690 | 13,102 | -16% | 1 | 1 | 0% | 2,884 | 3,396 | +18% | 0 | 0 | — |
case-15 | fail→pass | 17,034 | 15,506 | -9% | 1 | 1 | 0% | 3,032 | 3,690 | +22% | 0 | 0 | — |
case-16 | fail→pass | 13,201 | 11,889 | -10% | 1 | 1 | 0% | 2,372 | 3,171 | +34% | 0 | 0 | — |
case-17 | fail→pass | 12,529 | 13,450 | +7% | 1 | 1 | 0% | 2,082 | 3,348 | +61% | 0 | 0 | — |
case-18 | fail→fail | 12,920 | 14,492 | +12% | 1 | 1 | 0% | 2,567 | 3,556 | +39% | 0 | 0 | — |
case-19 | fail→pass | 14,381 | 12,462 | -13% | 1 | 1 | 0% | 2,538 | 3,118 | +23% | 0 | 0 | — |
case-20 | fail→pass | 14,819 | 12,585 | -15% | 1 | 1 | 0% | 2,575 | 3,324 | +29% | 0 | 0 | — |
case-21 | fail→pass | 17,473 | 14,781 | -15% | 1 | 1 | 0% | 2,949 | 3,564 | +21% | 0 | 0 | — |
case-22 | fail→pass | 12,953 | 12,522 | -3% | 1 | 1 | 0% | 2,239 | 3,150 | +41% | 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 +55 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.