Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Prepare a structured brief for an upcoming meeting with relevant context, open items, data points, and talking points. Used by digital twin personas to reduce meeting preparation overhead.
.claude/skills/cowork-os-twin-meeting-prep/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✓→✗ | ▼ Worse | -36% | 0% |
| case-19 | ✓→✗ | ▼ Worse | -1% | 0% |
| case-01 | ✓→✓ | = Same ✓ | -33% | 0% |
| case-02 | ✓→✓ | = Same ✓ | -29% | 0% |
| case-05 | ✓→✓ | = Same ✓ | 6% | 0% |
Prepare a structured brief for an upcoming meeting with relevant context, open items, data points, and talking points. Used by digital twin personas to reduce meeting preparation overhead.
| Name | Type | Required | Description | |---|---|---|---| | meeting_topic | string | Yes | What the meeting is about (e.g., 'Sprint planning', 'Architecture review for auth service', '1-on-1 with Sarah') |
../twin-meeting-prep.json.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 18,533 | 12,128 | -35% | 1 | 1 | 0% | 2,714 | 2,054 | -24% | 0 | 0 | — |
case-01 | pass→pass | 16,847 | 8,809 | -48% | 1 | 1 | 0% | 2,675 | 1,793 | -33% | 0 | 0 | — |
case-02 | pass→pass | 13,950 | 7,704 | -45% | 1 | 1 | 0% | 2,081 | 1,484 | -29% | 0 | 0 | — |
case-04 | fail→fail | 16,944 | 9,470 | -44% | 1 | 1 | 0% | 2,469 | 1,699 | -31% | 0 | 0 | — |
case-05 | pass→pass | 14,272 | 13,755 | -4% | 1 | 1 | 0% | 2,282 | 2,414 | +6% | 0 | 0 | — |
case-06 | fail→fail | 16,101 | 9,337 | -42% | 1 | 1 | 0% | 2,532 | 1,792 | -29% | 0 | 0 | — |
case-07 | fail→fail | 16,841 | 12,812 | -24% | 1 | 1 | 0% | 2,620 | 2,184 | -17% | 0 | 0 | — |
case-08 | fail→fail | 14,850 | 11,564 | -22% | 1 | 1 | 0% | 2,179 | 2,104 | -3% | 0 | 0 | — |
case-09 | fail→fail | 17,717 | 10,214 | -42% | 1 | 1 | 0% | 2,560 | 1,854 | -28% | 0 | 0 | — |
case-10 | fail→fail | 15,949 | 11,649 | -27% | 1 | 1 | 0% | 2,403 | 2,120 | -12% | 0 | 0 | — |
case-11 | fail→fail | 17,509 | 9,263 | -47% | 1 | 1 | 0% | 2,578 | 1,685 | -35% | 0 | 0 | — |
case-12 | fail→fail | 19,955 | 11,454 | -43% | 1 | 1 | 0% | 2,780 | 1,907 | -31% | 0 | 0 | — |
case-13 | fail→fail | 16,104 | 8,542 | -47% | 1 | 1 | 0% | 2,382 | 1,630 | -32% | 0 | 0 | — |
case-14 | fail→fail | 15,713 | 11,567 | -26% | 1 | 1 | 0% | 2,409 | 2,110 | -12% | 0 | 0 | — |
case-15 | fail→fail | 17,222 | 10,298 | -40% | 1 | 1 | 0% | 2,608 | 1,979 | -24% | 0 | 0 | — |
case-16 | pass→fail | 19,345 | 10,012 | -48% | 1 | 1 | 0% | 2,947 | 1,890 | -36% | 0 | 0 | — |
case-17 | pass→pass | 13,306 | 8,917 | -33% | 1 | 1 | 0% | 2,126 | 1,654 | -22% | 0 | 0 | — |
case-18 | fail→fail | 16,534 | 10,073 | -39% | 1 | 1 | 0% | 2,563 | 1,829 | -29% | 0 | 0 | — |
case-19 | pass→fail | 10,883 | 9,841 | -10% | 1 | 1 | 0% | 1,717 | 1,703 | -1% | 0 | 0 | — |
case-20 | fail→fail | 1,973 | 2,372 | +20% | 1 | 1 | 0% | 286 | 681 | +138% | 0 | 0 | — |
case-21 | pass→pass | 8,821 | 3,693 | -58% | 1 | 1 | 0% | 1,278 | 922 | -28% | 0 | 0 | — |
case-22 | fail→fail | 5,732 | 4,993 | -13% | 1 | 1 | 0% | 873 | 1,062 | +22% | 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 -60 percentage points is the difference between those two pass rates over the 22 comparable cases. 4 cases got worse with the skill loaded, and they are included in that figure.
Other measured skills in the registry, with their headline benchmark lift.