Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Deep web research with full page fetching via Apify. Import of web-browsing/web-research skill. Must fetch full page — no conclusions from previews.
.claude/skills/yogsoth-ai-deep-insight-web-research/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-20 | ✓→✓ | ＝ Same ✓ | 37% | 0% |
| case-21 | ✓→✓ | ＝ Same ✓ | 81% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | 86% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | -89% | 0% |
| case-02 | ✗→✗ | ＝ Same ✗ | 0% | 0% |
Deep web research with full page fetching.
Import — strictly follow web-browsing/web-research skill protocol.
Must fetch full page via Apify rag-web-browser. No conclusions from search previews or snippets alone. Full page content is the minimum evidence standard.
Quantity target is set by the calling strategy's budget table. This SOP executes one unit = one full page fetch + analysis.
web-browsing repo → skills/web-research/SKILL.md
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | web-research | Deep web research — fetches full page content for analysis. Snippets alone are PROHIBITED for conclusions. |
<!-- 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 | 37,572 | 10,614 | -72% | 1 | 1 | 0% | 6,191 | 653 | -89% | 0 | 0 | — |
case-02 | fail→fail | 31,182 | 7,545 | -76% | 1 | 1 | 0% | 635 | 634 | -0% | 0 | 0 | — |
case-03 | fail→fail | 38,825 | 11,646 | -70% | 1 | 1 | 0% | 6,191 | 976 | -84% | 0 | 0 | — |
case-04 | fail→fail | 6,910 | 23,628 | +242% | 1 | 1 | 0% | 1,001 | 3,887 | +288% | 0 | 0 | — |
case-05 | fail→fail | 8,732 | 18,028 | +106% | 1 | 1 | 0% | 897 | 3,194 | +256% | 0 | 0 | — |
case-06 | fail→fail | 16,771 | 27,058 | +61% | 1 | 1 | 0% | 2,793 | 4,843 | +73% | 0 | 0 | — |
case-07 | fail→fail | 4,319 | 22,048 | +410% | 1 | 1 | 0% | 296 | 3,599 | +1116% | 0 | 0 | — |
case-08 | fail→fail | 14,402 | 5,437 | -62% | 1 | 1 | 0% | 2,841 | 516 | -82% | 0 | 0 | — |
case-09 | fail→fail | 11,188 | 9,908 | -11% | 1 | 1 | 0% | 1,613 | 1,059 | -34% | 0 | 0 | — |
case-10 | fail→fail | 5,038 | 7,280 | +45% | 1 | 1 | 0% | 262 | 608 | +132% | 0 | 0 | — |
case-19 | fail→fail | 3,999 | 14,706 | +268% | 1 | 1 | 0% | 650 | 2,569 | +295% | 0 | 0 | — |
case-11 | fail→fail | 20,691 | 31,540 | +52% | 1 | 1 | 0% | 3,399 | 4,319 | +27% | 0 | 0 | — |
case-12 | fail→fail | 21,406 | 6,746 | -68% | 1 | 1 | 0% | 3,925 | 592 | -85% | 0 | 0 | — |
case-13 | fail→fail | 6,839 | 27,070 | +296% | 1 | 1 | 0% | 335 | 3,474 | +937% | 0 | 0 | — |
case-14 | fail→fail | 10,270 | 21,768 | +112% | 1 | 1 | 0% | 1,901 | 3,662 | +93% | 0 | 0 | — |
case-15 | fail→fail | 14,134 | 19,025 | +35% | 1 | 1 | 0% | 2,399 | 2,871 | +20% | 0 | 0 | — |
case-16 | fail→fail | 5,142 | 8,137 | +58% | 1 | 1 | 0% | 225 | 626 | +178% | 0 | 0 | — |
case-17 | fail→fail | 18,061 | 30,855 | +71% | 1 | 1 | 0% | 2,808 | 4,936 | +76% | 0 | 0 | — |
case-18 | fail→fail | 35,989 | 29,149 | -19% | 1 | 1 | 0% | 4,730 | 4,614 | -2% | 0 | 0 | — |
case-20 | pass→pass | 2,993 | 3,134 | +5% | 1 | 1 | 0% | 483 | 661 | +37% | 0 | 0 | — |
case-21 | pass→pass | 1,624 | 1,930 | +19% | 1 | 1 | 0% | 320 | 578 | +81% | 0 | 0 | — |
case-22 | pass→pass | 1,579 | 1,866 | +18% | 1 | 1 | 0% | 267 | 496 | +86% | 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 12 counted toward the lift figure. The other 10 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 0 percentage points is the difference between those two pass rates over the 12 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Other measured skills in the registry, with their headline benchmark lift.