Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Searches for external evidence supporting or opposing specific claims. Returns structured evidence with source assessment and relevance scoring.
.claude/skills/yogsoth-ai-evidence-scout/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -46% | 0% |
Searches for external evidence supporting or opposing claims.
Subagent — spawned via subagent-spawning/spawn-agent.
Evidence gathering requires web search and paper lookup in dedicated context. Isolated execution prevents search results from biasing ongoing debate reasoning.
One unit = one evidence gathering pass (multiple searches within budget).
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | spawn-agent | Spawn a customized CC subagent with full MCP tool access. Used by SOPs that declare execution: subagent. | | stress-test-paper-overview | Paper landscape scan returning abstracts and metadata. Import of literature-engine/paper-overview skill. Abstracts only — no conclusions from abstracts. | | stress-test-web-search | Quick web scanning for landscape understanding. Import of web-browsing/web-search skill. Snippets only — no conclusions from snippets alone. |
<!-- 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-23 | pass→pass | 7,426 | 8,735 | +18% | 1 | 1 | 0% | 1,191 | 1,625 | +36% | 0 | 0 | — |
case-04 | pass→pass | 6,600 | 2,818 | -57% | 1 | 1 | 0% | 995 | 816 | -18% | 0 | 0 | — |
case-01 | fail→fail | 35,862 | 9,638 | -73% | 1 | 1 | 0% | 6,239 | 988 | -84% | 0 | 0 | — |
case-02 | fail→fail | 24,963 | 6,225 | -75% | 1 | 1 | 0% | 4,256 | 830 | -80% | 0 | 0 | — |
case-03 | fail→fail | 23,289 | 35,558 | +53% | 1 | 1 | 0% | 4,055 | 5,683 | +40% | 0 | 0 | — |
case-05 | pass→pass | 5,800 | 2,118 | -63% | 1 | 1 | 0% | 813 | 689 | -15% | 0 | 0 | — |
case-06 | fail→pass | 6,149 | 2,012 | -67% | 1 | 1 | 0% | 913 | 666 | -27% | 0 | 0 | — |
case-07 | fail→pass | 11,385 | 2,405 | -79% | 1 | 1 | 0% | 1,805 | 672 | -63% | 0 | 0 | — |
case-08 | fail→pass | 7,340 | 3,359 | -54% | 1 | 1 | 0% | 990 | 796 | -20% | 0 | 0 | — |
case-09 | pass→pass | 7,094 | 1,674 | -76% | 1 | 1 | 0% | 1,160 | 588 | -49% | 0 | 0 | — |
case-10 | pass→pass | 5,858 | 2,028 | -65% | 1 | 1 | 0% | 968 | 671 | -31% | 0 | 0 | — |
case-11 | pass→pass | 9,676 | 1,897 | -80% | 1 | 1 | 0% | 1,498 | 629 | -58% | 0 | 0 | — |
case-12 | fail→pass | 8,969 | 2,595 | -71% | 1 | 1 | 0% | 1,327 | 713 | -46% | 0 | 0 | — |
case-13 | fail→pass | 6,710 | 1,520 | -77% | 1 | 1 | 0% | 1,068 | 572 | -46% | 0 | 0 | — |
case-14 | pass→pass | 6,343 | 2,820 | -56% | 1 | 1 | 0% | 975 | 767 | -21% | 0 | 0 | — |
case-15 | fail→pass | 6,902 | 1,715 | -75% | 1 | 1 | 0% | 963 | 615 | -36% | 0 | 0 | — |
case-16 | fail→pass | 12,406 | 5,143 | -59% | 1 | 1 | 0% | 1,975 | 1,084 | -45% | 0 | 0 | — |
case-17 | pass→pass | 7,953 | 1,346 | -83% | 1 | 1 | 0% | 1,153 | 522 | -55% | 0 | 0 | — |
case-18 | fail→pass | 8,404 | 2,353 | -72% | 1 | 1 | 0% | 1,200 | 751 | -37% | 0 | 0 | — |
case-19 | fail→pass | 10,762 | 1,718 | -84% | 1 | 1 | 0% | 1,543 | 626 | -59% | 0 | 0 | — |
case-20 | pass→pass | 9,096 | 6,003 | -34% | 1 | 1 | 0% | 1,474 | 1,305 | -11% | 0 | 0 | — |
case-21 | fail→pass | 9,687 | 1,818 | -81% | 1 | 1 | 0% | 1,460 | 651 | -55% | 0 | 0 | — |
case-22 | pass→pass | 8,513 | 1,597 | -81% | 1 | 1 | 0% | 1,257 | 573 | -54% | 0 | 0 | — |
case-24 | pass→pass | 7,359 | 6,107 | -17% | 1 | 1 | 0% | 1,056 | 1,267 | +20% | 0 | 0 | — |
case-25 | pass→pass | 6,395 | 6,597 | +3% | 1 | 1 | 0% | 1,066 | 1,389 | +30% | 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 23 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 +40 percentage points is the difference between those two pass rates over the 23 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.