Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Synthesizes findings from a completed campaign into typed verdict reports. Produces DebateVerdict, RedTeamReport, FailureAnticipationReport, CounterfactualMap, or AdversarialStressReport depending on campaign. Also supports cross-campaign StressTestSummary.
.claude/skills/yogsoth-ai-verdict-synthesis/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -73% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -65% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -53% | 0% |
Synthesizes campaign findings into typed verdict reports.
Subagent — spawned via subagent-spawning/spawn-agent.
Synthesis requires processing all strategy/tactic outputs from a campaign in dedicated context to produce coherent final report.
Typed report matching campaign:
DebateVerdict (multiagent-debate)RedTeamReport (red-teaming)FailureAnticipationReport (failure-anticipation)CounterfactualMap (counterfactual-probing)AdversarialStressReport (adversarial-stress-testing)StressTestSummary (cross-campaign aggregation)One unit = one synthesis pass producing a complete typed report.
<!-- 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 | fail→fail | 18,028 | 23,068 | +28% | 1 | 1 | 0% | 2,210 | 3,504 | +59% | 0 | 0 | — |
case-02 | fail→fail | 10,271 | 24,694 | +140% | 1 | 1 | 0% | 784 | 4,101 | +423% | 0 | 0 | — |
case-03 | fail→fail | 13,744 | 27,920 | +103% | 1 | 1 | 0% | 1,372 | 3,810 | +178% | 0 | 0 | — |
case-04 | fail→fail | 12,020 | 23,054 | +92% | 1 | 1 | 0% | 2,015 | 3,467 | +72% | 0 | 0 | — |
case-05 | fail→fail | 10,507 | 11,011 | +5% | 1 | 1 | 0% | 802 | 1,285 | +60% | 0 | 0 | — |
case-06 | fail→fail | 18,942 | 22,394 | +18% | 1 | 1 | 0% | 2,321 | 3,297 | +42% | 0 | 0 | — |
case-11 | pass→pass | 8,099 | 13,099 | +62% | 1 | 1 | 0% | 1,335 | 1,113 | -17% | 0 | 0 | — |
case-07 | fail→fail | 6,683 | 21,500 | +222% | 1 | 1 | 0% | 246 | 2,948 | +1098% | 0 | 0 | — |
case-08 | fail→fail | 8,463 | 12,758 | +51% | 1 | 1 | 0% | 1,262 | 2,158 | +71% | 0 | 0 | — |
case-09 | fail→fail | 16,288 | 19,713 | +21% | 1 | 1 | 0% | 1,869 | 3,563 | +91% | 0 | 0 | — |
case-10 | fail→fail | 14,639 | 15,175 | +4% | 1 | 1 | 0% | 1,458 | 2,737 | +88% | 0 | 0 | — |
case-12 | fail→pass | 16,403 | 8,893 | -46% | 1 | 1 | 0% | 2,615 | 984 | -62% | 0 | 0 | — |
case-13 | pass→pass | 27,253 | 7,578 | -72% | 1 | 1 | 0% | 3,231 | 745 | -77% | 0 | 0 | — |
case-14 | fail→pass | 17,070 | 1,583 | -91% | 1 | 1 | 0% | 2,063 | 558 | -73% | 0 | 0 | — |
case-15 | fail→pass | 14,311 | 6,936 | -52% | 1 | 1 | 0% | 1,715 | 605 | -65% | 0 | 0 | — |
case-16 | fail→pass | 15,526 | 7,104 | -54% | 1 | 1 | 0% | 1,799 | 676 | -62% | 0 | 0 | — |
case-17 | fail→pass | 28,183 | 7,280 | -74% | 1 | 1 | 0% | 1,374 | 646 | -53% | 0 | 0 | — |
case-18 | fail→pass | 14,968 | 1,998 | -87% | 1 | 1 | 0% | 1,546 | 581 | -62% | 0 | 0 | — |
case-19 | pass→pass | 14,420 | 8,217 | -43% | 1 | 1 | 0% | 1,425 | 816 | -43% | 0 | 0 | — |
case-20 | pass→fail | 15,641 | 13,011 | -17% | 1 | 1 | 0% | 1,563 | 1,451 | -7% | 0 | 0 | — |
case-21 | pass→pass | 13,873 | 19,328 | +39% | 1 | 1 | 0% | 1,479 | 2,966 | +101% | 0 | 0 | — |
case-22 | pass→pass | 13,793 | 10,346 | -25% | 1 | 1 | 0% | 2,134 | 2,015 | -6% | 0 | 0 | — |
case-23 | fail→fail | 14,509 | 17,656 | +22% | 1 | 1 | 0% | 1,783 | 2,733 | +53% | 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. The headline lift of +22 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is 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.