Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Catalog all failure modes in a domain, classify them systematically, and generate targeted solutions for each failure type.
.claude/skills/yogsoth-ai-failure-taxonomy/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 140% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 72% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 42% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 169% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 78% | 0% |
Catalog all failure modes, classify them into a taxonomy, and generate targeted solutions that address each specific failure type.
| Resource | Target | Current | % | |----------|--------|---------|---| | web-search | 25 | 0 | 0% | | web-research | 10 | 0 | 0% | | paper-overview | 25 | 0 | 0% | | paper-search | 15 | 0 | 0% | | paper-research | 5 | 0 | 0% |
Cannot exit strategy until ≥80% of each budget line is consumed OR yield targets are met with justification for remaining budget.
| Tactic | Role | |--------|------| | gap-driven-generation | Generate solutions from identified failure gaps | | evaluation-filtering | Score and filter failure-targeted ideas |
| SOP | Role | |-----|------| | failure-mode-cataloging | Systematically catalog all failure modes | | failure-driven-generation | Generate targeted solution per failure mode | | coverage-gap-detection | Detect failure modes without existing solutions | | enumeration-synthesis | Synthesize failure analysis into idea report |
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Tactic | When to use | | --- | --- | | gap-driven-generation | Generate solutions targeting specific coverage gaps — detect gaps, generate failure-driven solutions, and design factor-level experiments. |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | coverage-gap-detection | Detect uncovered regions in the solution space, producing a prioritized gap list. | | creative-ideation-failure-mode-cataloging | Systematically catalog all failure modes in a domain or method, producing a classified failure taxonomy. | | enumeration-synthesis | Synthesize all systematic enumeration outputs into a structured idea report with prioritized recommendations. | | failure-driven-generation | Generate targeted solutions for each identified failure mode, ensuring every failure has at least one proposed mitigation. |
<!-- 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 | 31,800 | 58,407 | +84% | 1 | 1 | 0% | 5,482 | 6,826 | +25% | 0 | 0 | — |
case-02 | fail→fail | 38,573 | 38,690 | +0% | 1 | 1 | 0% | 5,969 | 6,818 | +14% | 0 | 0 | — |
case-03 | fail→fail | 36,677 | 45,021 | +23% | 1 | 1 | 0% | 6,197 | 6,817 | +10% | 0 | 0 | — |
case-04 | pass→fail | 15,531 | 31,209 | +101% | 1 | 1 | 0% | 2,736 | 6,045 | +121% | 0 | 0 | — |
case-05 | pass→pass | 19,444 | 31,193 | +60% | 1 | 1 | 0% | 3,872 | 6,796 | +76% | 0 | 0 | — |
case-06 | pass→fail | 17,565 | 33,947 | +93% | 1 | 1 | 0% | 2,983 | 6,811 | +128% | 0 | 0 | — |
case-07 | fail→pass | 6,735 | 9,400 | +40% | 1 | 1 | 0% | 987 | 2,366 | +140% | 0 | 0 | — |
case-08 | pass→pass | 15,784 | 35,054 | +122% | 1 | 1 | 0% | 2,433 | 4,941 | +103% | 0 | 0 | — |
case-09 | fail→fail | 16,773 | 7,283 | -57% | 1 | 1 | 0% | 2,579 | 1,099 | -57% | 0 | 0 | — |
case-10 | fail→pass | 8,809 | 9,885 | +12% | 1 | 1 | 0% | 1,355 | 2,327 | +72% | 0 | 0 | — |
case-11 | pass→fail | 15,617 | 8,132 | -48% | 1 | 1 | 0% | 2,352 | 1,142 | -51% | 0 | 0 | — |
case-12 | fail→pass | 9,240 | 8,464 | -8% | 1 | 1 | 0% | 1,437 | 2,039 | +42% | 0 | 0 | — |
case-13 | fail→pass | 10,331 | 24,944 | +141% | 1 | 1 | 0% | 1,628 | 4,385 | +169% | 0 | 0 | — |
case-14 | fail→pass | 12,771 | 1,354 | -89% | 1 | 1 | 0% | 483 | 862 | +78% | 0 | 0 | — |
case-15 | fail→pass | 16,090 | 1,737 | -89% | 1 | 1 | 0% | 1,046 | 890 | -15% | 0 | 0 | — |
case-16 | fail→pass | 2,420 | 1,427 | -41% | 1 | 1 | 0% | 354 | 830 | +134% | 0 | 0 | — |
case-17 | fail→pass | 4,723 | 1,374 | -71% | 1 | 1 | 0% | 717 | 847 | +18% | 0 | 0 | — |
case-18 | fail→pass | 4,192 | 1,944 | -54% | 1 | 1 | 0% | 598 | 889 | +49% | 0 | 0 | — |
case-19 | pass→pass | 16,275 | 17,005 | +4% | 1 | 1 | 0% | 2,352 | 3,529 | +50% | 0 | 0 | — |
case-20 | pass→pass | 6,900 | 1,924 | -72% | 1 | 1 | 0% | 1,029 | 932 | -9% | 0 | 0 | — |
case-21 | fail→pass | 8,964 | 8,212 | -8% | 1 | 1 | 0% | 1,397 | 1,274 | -9% | 0 | 0 | — |
case-22 | pass→pass | 8,851 | 4,690 | -47% | 1 | 1 | 0% | 1,402 | 1,418 | +1% | 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 18 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 +32 percentage points is the difference between those two pass rates over the 18 comparable cases. 4 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.