Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Validate gap authenticity via cross-database verification, temporal sensitivity testing, and false-gap filtering. Ensures gaps are genuine absences, not search artifacts.
.claude/skills/yogsoth-ai-gap-validation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 9% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 11% | 0% |
Verify that identified gaps are genuine research absences, not artifacts of search failure.
Classified gaps need validation before prioritization — confirm they represent real absences across multiple sources and time windows.
| Base SOP | Target | ±10% Range | |----------|--------|------------| | web-search | 40 | 36–44 | | web-research | 15 | 13–17 | | paper-overview | 40 | 36–44 | | paper-search | 25 | 22–28 | | paper-research | 15 | 13–17 |
Print before every iteration:
<HARD-GATE>
| SOP | Done | Target | % |
|-----|------|--------|---|
| web-search | ? | 40 | ? |
| web-research | ? | 15 | ? |
| paper-overview | ? | 40 | ? |
| paper-search | ? | 25 | ? |
| paper-research | ? | 15 | ? |
Budget Gate: OPEN/CLOSED (>=80% required to exit)
</HARD-GATE>Import: web-search, web-research, paper-overview, paper-search, paper-research Subagent: cross-database-verification, false-gap-filtering, temporal-sensitivity-testing
For each classified gap: verify across multiple databases (Semantic Scholar, Google Scholar, arXiv, domain-specific), test temporal stability (does gap persist across 2/5/10 year windows?), filter false gaps (search failure vs genuine absence vs already solved).
Validated Gap List — each gap with: validation status (confirmed/partial/refuted), cross-database results, temporal trend, false-gap assessment.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Tactic | When to use | | --- | --- | | cross-validation | Multi-source cross-validation of gap authenticity — cross-database search, temporal sensitivity testing, false-gap filtering, stakeholder confirmation. |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | cross-database-verification | Verify gap existence across multiple databases (Semantic Scholar, Google Scholar, arXiv, domain-specific). Distinguishes database-specific gaps from universal gaps. | | false-gap-filtering | Detect false gaps — search failures, already-solved gaps, and inherently unanswerable questions masquerading as research gaps. | | temporal-sensitivity-testing | Test whether a gap persists across different time windows (2/5/10 years). Determines if gap is narrowing, widening, or stable over time. |
<!-- 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 | 29,629 | 24,875 | -16% | 1 | 1 | 0% | 4,578 | 2,152 | -53% | 0 | 0 | — |
case-02 | fail→fail | 31,672 | 7,999 | -75% | 1 | 1 | 0% | 5,252 | 1,966 | -63% | 0 | 0 | — |
case-03 | fail→pass | 35,822 | 35,836 | +0% | 1 | 1 | 0% | 5,564 | 6,889 | +24% | 0 | 0 | — |
case-04 | fail→pass | 32,492 | 2,242 | -93% | 1 | 1 | 0% | 1,845 | 1,110 | -40% | 0 | 0 | — |
case-05 | fail→pass | 13,419 | 1,689 | -87% | 1 | 1 | 0% | 894 | 962 | +8% | 0 | 0 | — |
case-06 | fail→pass | 25,055 | 2,365 | -91% | 1 | 1 | 0% | 989 | 1,079 | +9% | 0 | 0 | — |
case-07 | fail→pass | 10,174 | 5,509 | -46% | 1 | 1 | 0% | 1,576 | 1,744 | +11% | 0 | 0 | — |
case-08 | fail→pass | 12,204 | 8,885 | -27% | 1 | 1 | 0% | 2,024 | 2,286 | +13% | 0 | 0 | — |
case-09 | pass→pass | 15,622 | 2,758 | -82% | 1 | 1 | 0% | 2,341 | 1,192 | -49% | 0 | 0 | — |
case-10 | pass→pass | 11,538 | 3,591 | -69% | 1 | 1 | 0% | 1,727 | 1,335 | -23% | 0 | 0 | — |
case-11 | fail→pass | 17,008 | 2,688 | -84% | 1 | 1 | 0% | 854 | 1,104 | +29% | 0 | 0 | — |
case-12 | fail→pass | 7,649 | 2,414 | -68% | 1 | 1 | 0% | 480 | 1,066 | +122% | 0 | 0 | — |
case-13 | fail→pass | 5,276 | 1,901 | -64% | 1 | 1 | 0% | 826 | 1,026 | +24% | 0 | 0 | — |
case-14 | fail→pass | 9,975 | 2,188 | -78% | 1 | 1 | 0% | 1,663 | 987 | -41% | 0 | 0 | — |
case-15 | fail→pass | 10,522 | 5,653 | -46% | 1 | 1 | 0% | 1,725 | 1,721 | -0% | 0 | 0 | — |
case-16 | fail→pass | 8,953 | 5,180 | -42% | 1 | 1 | 0% | 1,260 | 1,689 | +34% | 0 | 0 | — |
case-17 | fail→pass | 9,375 | 2,622 | -72% | 1 | 1 | 0% | 1,476 | 1,158 | -22% | 0 | 0 | — |
case-18 | fail→pass | 9,497 | 5,718 | -40% | 1 | 1 | 0% | 1,450 | 1,743 | +20% | 0 | 0 | — |
case-19 | pass→pass | 8,118 | 5,437 | -33% | 1 | 1 | 0% | 1,203 | 1,563 | +30% | 0 | 0 | — |
case-20 | pass→fail | 18,131 | 4,286 | -76% | 1 | 1 | 0% | 2,767 | 1,346 | -51% | 0 | 0 | — |
case-21 | fail→pass | 6,895 | 34,423 | +399% | 1 | 1 | 0% | 1,104 | 6,865 | +522% | 0 | 0 | — |
case-22 | pass→fail | 12,641 | 18,592 | +47% | 1 | 1 | 0% | 1,819 | 3,812 | +110% | 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 17 counted toward the lift figure. The other 5 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 +59 percentage points is the difference between those two pass rates over the 17 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.