Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Produce final meta-analysis protocol document assembling all planning outputs into PRISMA-compliant protocol
.claude/skills/yogsoth-ai-meta-analysis-synthesis/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 3% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -16% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -39% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -33% | 0% |
Assemble all planning outputs into a complete, PRISMA-compliant meta-analysis protocol document ready for registration (PROSPERO) and execution.
all_planning_outputs: Combined outputs from all preceding SOPs (PICO, inclusion criteria, effect size plan, extraction form, RoB plan, heterogeneity plan, bias plan, sensitivity plan, and optionally network construction)Complete meta-analysis protocol document in PRISMA-P format, suitable for PROSPERO registration.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | spawn-agent | Spawn a customized CC subagent with full MCP tool access. Used by SOPs that declare execution: subagent. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 13,997 | 13,751 | -2% | 1 | 1 | 0% | 1,628 | 1,682 | +3% | 0 | 0 | — |
case-02 | pass→pass | 6,627 | 9,028 | +36% | 1 | 1 | 0% | 1,024 | 865 | -16% | 0 | 0 | — |
case-03 | pass→pass | 17,675 | 5,357 | -70% | 1 | 1 | 0% | 1,905 | 1,169 | -39% | 0 | 0 | — |
case-04 | pass→pass | 10,287 | 5,607 | -45% | 1 | 1 | 0% | 1,787 | 1,198 | -33% | 0 | 0 | — |
case-05 | pass→pass | 19,265 | 7,051 | -63% | 1 | 1 | 0% | 2,065 | 1,450 | -30% | 0 | 0 | — |
case-06 | fail→pass | 16,085 | 12,405 | -23% | 1 | 1 | 0% | 1,679 | 1,353 | -19% | 0 | 0 | — |
case-19 | pass→pass | 11,319 | 3,837 | -66% | 1 | 1 | 0% | 1,756 | 807 | -54% | 0 | 0 | — |
case-07 | pass→pass | 10,002 | 12,884 | +29% | 1 | 1 | 0% | 1,567 | 1,508 | -4% | 0 | 0 | — |
case-08 | pass→pass | 9,427 | 9,126 | -3% | 1 | 1 | 0% | 728 | 966 | +33% | 0 | 0 | — |
case-09 | pass→pass | 20,431 | 14,741 | -28% | 1 | 1 | 0% | 1,793 | 1,650 | -8% | 0 | 0 | — |
case-10 | fail→fail | 22,753 | 14,177 | -38% | 1 | 1 | 0% | 2,839 | 2,668 | -6% | 0 | 0 | — |
case-11 | fail→fail | 4,958 | 9,448 | +91% | 1 | 1 | 0% | 814 | 957 | +18% | 0 | 0 | — |
case-12 | fail→fail | 9,140 | 5,388 | -41% | 1 | 1 | 0% | 848 | 1,219 | +44% | 0 | 0 | — |
case-13 | fail→fail | 21,169 | 14,514 | -31% | 1 | 1 | 0% | 2,795 | 2,649 | -5% | 0 | 0 | — |
case-14 | fail→fail | 4,856 | 5,886 | +21% | 1 | 1 | 0% | 806 | 1,168 | +45% | 0 | 0 | — |
case-15 | fail→fail | 13,919 | 10,727 | -23% | 1 | 1 | 0% | 1,614 | 1,144 | -29% | 0 | 0 | — |
case-16 | fail→fail | 11,865 | 11,795 | -1% | 1 | 1 | 0% | 1,219 | 1,480 | +21% | 0 | 0 | — |
case-17 | fail→fail | 8,740 | 7,409 | -15% | 1 | 1 | 0% | 650 | 1,465 | +125% | 0 | 0 | — |
case-18 | fail→fail | 16,157 | 15,248 | -6% | 1 | 1 | 0% | 1,642 | 1,854 | +13% | 0 | 0 | — |
case-20 | pass→pass | 9,746 | 21,554 | +121% | 1 | 1 | 0% | 1,806 | 3,237 | +79% | 0 | 0 | — |
case-21 | pass→pass | 19,478 | 25,102 | +29% | 1 | 1 | 0% | 2,420 | 3,707 | +53% | 0 | 0 | — |
case-22 | pass→pass | 18,980 | 13,305 | -30% | 1 | 1 | 0% | 2,504 | 2,619 | +5% | 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 +5 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.