Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Systematically scan all known solutions, identify gaps in coverage and unexplored regions of the solution space.
.claude/skills/yogsoth-ai-benchmark-sweep/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 146% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 87% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 174% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 71% | 0% |
Systematically scan all known solutions in a domain, catalog their properties, and identify gaps where no solution exists.
| Resource | Target | Current | % | |----------|--------|---------|---| | web-search | 30 | 0 | 0% | | web-research | 10 | 0 | 0% | | paper-overview | 30 | 0 | 0% | | paper-search | 20 | 0 | 0% | | paper-research | 8 | 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 | |--------|------| | coverage-analysis | Inventory → crossing → intersection evaluation pipeline | | evaluation-filtering | Score and filter generated gap-filling ideas |
| SOP | Role | |-----|------| | benchmark-inventory | Catalog all known solutions with performance/applicability/limitations | | method-problem-crossing | Build cross-reference matrix from inventory | | intersection-evaluation | Annotate matrix cells as explored/partial/unexplored | | enumeration-synthesis | Synthesize sweep findings into structured report |
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Tactic | When to use | | --- | --- | | coverage-analysis | Systematic coverage evaluation pipeline — benchmark inventory, method-problem crossing, and intersection evaluation to map explored vs unexplored solution space. |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | creative-ideation-benchmark-inventory | Catalog all known solutions/methods in a domain with performance, applicability, and limitations. | | enumeration-synthesis | Synthesize all systematic enumeration outputs into a structured idea report with prioritized recommendations. | | intersection-evaluation | Evaluate exploration status of each cell in a method×problem matrix, annotating as explored, partial, or unexplored. | | method-problem-crossing | Build method×problem cross-reference matrix showing which methods have been applied to which problems. |
<!-- 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 | 41,484 | 37,688 | -9% | 1 | 1 | 0% | 6,206 | 6,823 | +10% | 0 | 0 | — |
case-02 | fail→fail | 41,061 | 35,759 | -13% | 1 | 1 | 0% | 6,207 | 6,160 | -1% | 0 | 0 | — |
case-03 | fail→fail | 39,986 | 38,228 | -4% | 1 | 1 | 0% | 6,210 | 6,827 | +10% | 0 | 0 | — |
case-04 | pass→pass | 35,911 | 48,083 | +34% | 1 | 1 | 0% | 6,177 | 6,794 | +10% | 0 | 0 | — |
case-05 | pass→pass | 24,580 | 39,962 | +63% | 1 | 1 | 0% | 3,790 | 6,798 | +79% | 0 | 0 | — |
case-06 | pass→pass | 16,304 | 16,227 | -0% | 1 | 1 | 0% | 3,100 | 3,783 | +22% | 0 | 0 | — |
case-07 | fail→pass | 39,352 | 37,560 | -5% | 1 | 1 | 0% | 6,184 | 6,801 | +10% | 0 | 0 | — |
case-08 | fail→pass | 15,735 | 29,900 | +90% | 1 | 1 | 0% | 2,271 | 5,576 | +146% | 0 | 0 | — |
case-09 | fail→fail | 24,299 | 39,156 | +61% | 1 | 1 | 0% | 3,786 | 6,795 | +79% | 0 | 0 | — |
case-10 | pass→fail | 27,239 | 45,082 | +66% | 1 | 1 | 0% | 4,088 | 6,788 | +66% | 0 | 0 | — |
case-11 | fail→pass | 24,793 | 40,168 | +62% | 1 | 1 | 0% | 3,563 | 6,679 | +87% | 0 | 0 | — |
case-12 | pass→pass | 26,592 | 41,664 | +57% | 1 | 1 | 0% | 4,045 | 6,786 | +68% | 0 | 0 | — |
case-13 | fail→fail | 21,118 | 21,607 | +2% | 1 | 1 | 0% | 3,703 | 1,359 | -63% | 0 | 0 | — |
case-14 | fail→pass | 17,615 | 39,232 | +123% | 1 | 1 | 0% | 2,476 | 6,794 | +174% | 0 | 0 | — |
case-15 | pass→pass | 29,080 | 37,949 | +30% | 1 | 1 | 0% | 4,189 | 6,237 | +49% | 0 | 0 | — |
case-16 | fail→pass | 25,052 | 43,044 | +72% | 1 | 1 | 0% | 3,977 | 6,784 | +71% | 0 | 0 | — |
case-17 | fail→pass | 10,992 | 8,499 | -23% | 1 | 1 | 0% | 1,600 | 1,876 | +17% | 0 | 0 | — |
case-18 | fail→pass | 24,992 | 39,659 | +59% | 1 | 1 | 0% | 3,877 | 6,782 | +75% | 0 | 0 | — |
case-19 | fail→fail | 24,796 | 38,060 | +53% | 1 | 1 | 0% | 3,832 | 6,782 | +77% | 0 | 0 | — |
case-20 | fail→fail | 23,257 | 38,151 | +64% | 1 | 1 | 0% | 3,392 | 6,791 | +100% | 0 | 0 | — |
case-21 | fail→fail | 28,462 | 8,288 | -71% | 1 | 1 | 0% | 4,736 | 1,027 | -78% | 0 | 0 | — |
case-22 | fail→fail | 39,191 | 9,695 | -75% | 1 | 1 | 0% | 6,179 | 1,528 | -75% | 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 20 counted toward the lift figure. The other 2 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 +27 percentage points is the difference between those two pass rates over the 20 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.