Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Systematic stress testing of assumptions — surface, classify by vulnerability, attack, assess fragility. Combines assumption-surfacing (shared), abp-vulnerability-classification, and clr-validation SOPs.
.claude/skills/yogsoth-ai-assumption-stress-test/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 128% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-21 | ✓→✗ | ▼ Worse | 25% | 0% |
Systematically stress-test assumptions to find dangerous ones.
Shared: assumption-surfacing Subagent: abp-vulnerability-classification, clr-validation Import: paper-research
Surface all assumptions (shared SOP). Classify each by load-bearing × vulnerable matrix. Validate causal logic of most critical assumptions (High-Load × High-Vulnerable quadrant).
<HARD-GATE>
- assumptions surfaced: >= 5
- vulnerability classifications: >= 5
- CLR validations on critical assumptions: >= 2
- dangerous assumptions identified: >= 1
</HARD-GATE><!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | abp-vulnerability-classification | Classify assumptions on 2 axes — load-bearing (how much conclusion depends on it) × vulnerable (how likely to be false). Focuses attention on High-Load × High-Vulnerable quadrant. | | clr-validation | Apply Goldratt's 8 Categories of Legitimate Reservation to validate causal claims. Tests clarity, existence, sufficiency, and logical integrity. | | deep-insight-assumption-surfacing | Systematically extract implicit assumptions from methods, frameworks, or arguments. Identifies what is taken for granted without explicit justification. | | deep-insight-paper-research | Full-text paper reading via three-pass Keshav method. Import of literature-engine/literature-research skill. Authoritative source for claims about paper content. |
<!-- 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-22 | pass→pass | 30,184 | 37,204 | +23% | 1 | 1 | 0% | 4,643 | 6,595 | +42% | 0 | 0 | — |
case-01 | pass→pass | 16,100 | 29,691 | +84% | 1 | 1 | 0% | 2,321 | 4,852 | +109% | 0 | 0 | — |
case-02 | pass→pass | 19,557 | 17,766 | -9% | 1 | 1 | 0% | 3,261 | 3,110 | -5% | 0 | 0 | — |
case-03 | pass→pass | 13,856 | 18,967 | +37% | 1 | 1 | 0% | 2,092 | 3,554 | +70% | 0 | 0 | — |
case-04 | pass→pass | 7,664 | 16,796 | +119% | 1 | 1 | 0% | 1,341 | 3,219 | +140% | 0 | 0 | — |
case-05 | fail→fail | 5,214 | 3,598 | -31% | 1 | 1 | 0% | 828 | 1,001 | +21% | 0 | 0 | — |
case-06 | pass→pass | 22,761 | 26,240 | +15% | 1 | 1 | 0% | 3,380 | 4,416 | +31% | 0 | 0 | — |
case-07 | pass→pass | 15,455 | 21,511 | +39% | 1 | 1 | 0% | 2,419 | 3,765 | +56% | 0 | 0 | — |
case-08 | pass→pass | 17,142 | 26,194 | +53% | 1 | 1 | 0% | 2,512 | 4,778 | +90% | 0 | 0 | — |
case-09 | pass→pass | 11,674 | 25,419 | +118% | 1 | 1 | 0% | 1,777 | 4,551 | +156% | 0 | 0 | — |
case-10 | fail→pass | 12,238 | 25,616 | +109% | 1 | 1 | 0% | 1,972 | 4,506 | +128% | 0 | 0 | — |
case-11 | pass→pass | 13,233 | 23,852 | +80% | 1 | 1 | 0% | 2,017 | 4,231 | +110% | 0 | 0 | — |
case-12 | pass→pass | 11,146 | 4,998 | -55% | 1 | 1 | 0% | 1,741 | 991 | -43% | 0 | 0 | — |
case-13 | pass→pass | 18,737 | 25,215 | +35% | 1 | 1 | 0% | 2,882 | 4,398 | +53% | 0 | 0 | — |
case-14 | pass→pass | 15,168 | 16,339 | +8% | 1 | 1 | 0% | 2,359 | 3,192 | +35% | 0 | 0 | — |
case-15 | fail→pass | 21,792 | 22,309 | +2% | 1 | 1 | 0% | 3,193 | 4,075 | +28% | 0 | 0 | — |
case-16 | fail→pass | 20,550 | 24,904 | +21% | 1 | 1 | 0% | 3,172 | 4,272 | +35% | 0 | 0 | — |
case-21 | pass→fail | 28,890 | 35,646 | +23% | 1 | 1 | 0% | 5,310 | 6,622 | +25% | 0 | 0 | — |
case-17 | pass→fail | 14,531 | 23,968 | +65% | 1 | 1 | 0% | 2,321 | 4,300 | +85% | 0 | 0 | — |
case-18 | pass→pass | 14,066 | 10,137 | -28% | 1 | 1 | 0% | 2,047 | 1,916 | -6% | 0 | 0 | — |
case-19 | fail→pass | 31,247 | 2,759 | -91% | 1 | 1 | 0% | 1,351 | 825 | -39% | 0 | 0 | — |
case-20 | pass→pass | 10,942 | 26,255 | +140% | 1 | 1 | 0% | 2,383 | 5,254 | +120% | 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, and 21 counted toward the lift figure. The other 1 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 +9 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.