Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Classic reductio ad absurdum: negate the core claim, derive logical consequences, seek contradiction or absurdity.
.claude/skills/yogsoth-ai-assumption-negation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-11 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -46% | 0% |
| Size | Negation chains | Max derivation depth | |---|---|---| | S | 3 | 5 steps | | M | 6 | 8 steps | | L | 10 | 12 steps |
claim-negation to produce formal negationdeductive-chain to derive consequencescontradiction-detection to evaluate resultsclaim-refinement for weakened version<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Tactic | When to use | | --- | --- | | contradiction-derivation | Negate a claim, derive logical consequences step by step, detect whether a genuine contradiction or absurdity emerges. | | counterexample-heuristics | Generate counterexamples (monsters), attempt monster-barring, incorporate surviving counterexamples as lemma refinements (Lakatos method). |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | claim-negation | Formally negate the core claim, producing the logical complement for reductio testing. | | claim-refinement | Propose a refined claim that survives counterexamples while preserving maximum explanatory power (Lakatos lemma-incorporation). | | contradiction-detection | Evaluate whether a derivation chain has reached a genuine contradiction, absurdity, or inconclusive state. | | deductive-chain | Derive logical consequences step by step from a given premise, building a traceable derivation chain. |
<!-- 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-11 | fail→pass | 25,399 | 1,680 | -93% | 1 | 1 | 0% | 1,166 | 732 | -37% | 0 | 0 | — |
case-10 | fail→pass | 24,148 | 1,704 | -93% | 1 | 1 | 0% | 1,516 | 712 | -53% | 0 | 0 | — |
case-01 | fail→pass | 6,953 | 3,979 | -43% | 1 | 1 | 0% | 1,111 | 1,258 | +13% | 0 | 0 | — |
case-02 | pass→pass | 6,274 | 6,290 | +0% | 1 | 1 | 0% | 1,049 | 1,769 | +69% | 0 | 0 | — |
case-03 | pass→pass | 8,288 | 6,520 | -21% | 1 | 1 | 0% | 1,504 | 1,650 | +10% | 0 | 0 | — |
case-04 | pass→pass | 5,668 | 7,059 | +25% | 1 | 1 | 0% | 932 | 1,665 | +79% | 0 | 0 | — |
case-05 | pass→pass | 4,927 | 5,444 | +10% | 1 | 1 | 0% | 863 | 1,374 | +59% | 0 | 0 | — |
case-06 | pass→pass | 7,418 | 10,540 | +42% | 1 | 1 | 0% | 1,323 | 2,374 | +79% | 0 | 0 | — |
case-07 | fail→pass | 18,719 | 1,805 | -90% | 1 | 1 | 0% | 841 | 755 | -10% | 0 | 0 | — |
case-08 | pass→pass | 10,611 | 1,081 | -90% | 1 | 1 | 0% | 1,622 | 624 | -62% | 0 | 0 | — |
case-09 | fail→pass | 32,607 | 1,728 | -95% | 1 | 1 | 0% | 1,416 | 767 | -46% | 0 | 0 | — |
case-12 | fail→pass | 17,114 | 1,687 | -90% | 1 | 1 | 0% | 2,664 | 760 | -71% | 0 | 0 | — |
case-13 | fail→pass | 8,507 | 2,006 | -76% | 1 | 1 | 0% | 1,278 | 807 | -37% | 0 | 0 | — |
case-14 | fail→pass | 8,880 | 2,016 | -77% | 1 | 1 | 0% | 1,433 | 840 | -41% | 0 | 0 | — |
case-15 | pass→fail | 11,040 | 2,298 | -79% | 1 | 1 | 0% | 1,779 | 880 | -51% | 0 | 0 | — |
case-16 | fail→pass | 9,799 | 2,455 | -75% | 1 | 1 | 0% | 1,543 | 892 | -42% | 0 | 0 | — |
case-17 | pass→pass | 9,352 | 2,848 | -70% | 1 | 1 | 0% | 1,635 | 990 | -39% | 0 | 0 | — |
case-18 | pass→pass | 9,977 | 2,168 | -78% | 1 | 1 | 0% | 1,701 | 811 | -52% | 0 | 0 | — |
case-19 | pass→pass | 8,335 | 1,937 | -77% | 1 | 1 | 0% | 1,325 | 815 | -38% | 0 | 0 | — |
case-20 | pass→pass | 5,799 | 2,430 | -58% | 1 | 1 | 0% | 883 | 906 | +3% | 0 | 0 | — |
case-21 | pass→pass | 12,270 | 2,405 | -80% | 1 | 1 | 0% | 1,892 | 885 | -53% | 0 | 0 | — |
case-22 | pass→pass | 12,700 | 1,841 | -86% | 1 | 1 | 0% | 1,895 | 803 | -58% | 0 | 0 | — |
case-23 | pass→fail | 12,942 | 19,561 | +51% | 1 | 1 | 0% | 2,229 | 3,986 | +79% | 0 | 0 | — |
case-24 | pass→pass | 6,249 | 4,892 | -22% | 1 | 1 | 0% | 1,155 | 1,328 | +15% | 0 | 0 | — |
case-25 | pass→pass | 5,067 | 5,805 | +15% | 1 | 1 | 0% | 989 | 1,557 | +57% | 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. 25 cases were attempted, and 21 counted toward the lift figure. The other 4 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 +28 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.