Install any skill in seconds. Free to start, no credit card required.
Get Started Free →SOP for gathering source material relevant to the ontology domain — fetch papers and web pages into source pages.
.claude/skills/yogsoth-ai-source-gathering/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✓→✗ | ▼ Worse | -60% | 0% |
| case-04 | ✓→✗ | ▼ Worse | -63% | 0% |
| case-13 | ✓→✗ | ▼ Worse | -79% | 0% |
| case-18 | ✓→✗ | ▼ Worse | -57% | 0% |
| case-05 | ✓→✓ | = Same ✓ | -6% | 0% |
Gather source material (papers, web pages) relevant to the ontology domain and ingest as source pages.
wiki-ingest-source (SOP) + web/paper search tools
For public X sources, read references/xquik-apify-x-research.md before choosing an Actor.
<HARD-GATE> Must ingest ≥2 source pages per invocation. </HARD-GATE>
Returns: { sources_ingested: number, paths: string[] }
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | wiki-ingest-source | SOP for source page creation — write immutable source page capturing raw material, then update search index. |
<!-- 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-06 | pass→fail | 29,121 | 21,447 | -26% | 1 | 1 | 0% | 3,327 | 1,337 | -60% | 0 | 0 | — |
case-01 | fail→fail | 17,585 | 18,352 | +4% | 1 | 1 | 0% | 2,282 | 1,055 | -54% | 0 | 0 | — |
case-02 | fail→fail | 12,489 | 46,998 | +276% | 1 | 1 | 0% | 568 | 560 | -1% | 0 | 0 | — |
case-03 | fail→fail | 32,022 | 15,290 | -52% | 1 | 1 | 0% | 4,925 | 492 | -90% | 0 | 0 | — |
case-04 | pass→fail | 20,828 | 22,072 | +6% | 1 | 1 | 0% | 3,336 | 1,219 | -63% | 0 | 0 | — |
case-05 | pass→pass | 35,415 | 30,892 | -13% | 1 | 1 | 0% | 5,114 | 4,793 | -6% | 0 | 0 | — |
case-07 | fail→fail | 10,393 | 19,987 | +92% | 1 | 1 | 0% | 808 | 1,004 | +24% | 0 | 0 | — |
case-08 | fail→fail | 15,582 | 27,149 | +74% | 1 | 1 | 0% | 1,987 | 724 | -64% | 0 | 0 | — |
case-09 | fail→fail | 30,644 | 16,165 | -47% | 1 | 1 | 0% | 5,071 | 640 | -87% | 0 | 0 | — |
case-10 | fail→fail | 22,911 | 15,585 | -32% | 1 | 1 | 0% | 3,361 | 560 | -83% | 0 | 0 | — |
case-11 | fail→fail | 12,027 | 15,752 | +31% | 1 | 1 | 0% | 1,091 | 545 | -50% | 0 | 0 | — |
case-12 | fail→fail | 28,180 | 20,082 | -29% | 1 | 1 | 0% | 4,404 | 893 | -80% | 0 | 0 | — |
case-13 | pass→fail | 19,681 | 16,688 | -15% | 1 | 1 | 0% | 2,623 | 540 | -79% | 0 | 0 | — |
case-14 | fail→fail | 18,327 | 19,053 | +4% | 1 | 1 | 0% | 2,200 | 1,163 | -47% | 0 | 0 | — |
case-15 | fail→fail | 17,709 | 16,995 | -4% | 1 | 1 | 0% | 2,002 | 695 | -65% | 0 | 0 | — |
case-16 | fail→fail | 20,186 | 15,297 | -24% | 1 | 1 | 0% | 2,613 | 507 | -81% | 0 | 0 | — |
case-17 | fail→fail | 21,096 | 16,337 | -23% | 1 | 1 | 0% | 2,851 | 548 | -81% | 0 | 0 | — |
case-18 | pass→fail | 18,566 | 20,446 | +10% | 1 | 1 | 0% | 2,164 | 930 | -57% | 0 | 0 | — |
case-19 | fail→fail | 42,904 | 16,537 | -61% | 1 | 1 | 0% | 5,108 | 564 | -89% | 0 | 0 | — |
case-20 | fail→fail | 10,921 | 18,647 | +71% | 1 | 1 | 0% | 1,047 | 1,112 | +6% | 0 | 0 | — |
case-21 | fail→fail | 16,056 | 15,211 | -5% | 1 | 1 | 0% | 258 | 537 | +108% | 0 | 0 | — |
case-22 | fail→fail | 14,550 | 16,771 | +15% | 1 | 1 | 0% | 297 | 757 | +155% | 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 1 counted toward the lift figure. The other 21 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. A headline lift is not published for this run.
Other measured skills in the registry, with their headline benchmark lift.