Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →AI-powered paper summary and search. Import of literature-engine/literature-search skill. AI summary level — cite as "AI-extracted" not "paper states".
.claude/skills/yogsoth-ai-deep-insight-paper-search/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -83% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -75% | 0% |
AI-powered paper summary and search.
Import — strictly follow literature-engine/literature-search skill protocol.
AI summary level — cite findings as "AI-extracted" not "paper states". For authoritative claims about paper content, use paper-research (full text reading).
Quantity target is set by the calling strategy's budget table. This SOP executes one unit = one paper fetch + AI summary generation.
literature-engine repo → skills/literature-search/SKILL.md
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | literature-search | Medium-depth literature search — read AI-summarized reports for every paper analyzed |
<!-- 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-20 | pass→pass | 8,752 | 3,626 | -59% | 1 | 1 | 0% | 1,429 | 851 | -40% | 0 | 0 | — |
case-01 | pass→pass | 20,838 | 28,393 | +36% | 1 | 1 | 0% | 3,430 | 4,061 | +18% | 0 | 0 | — |
case-02 | pass→pass | 23,415 | 20,990 | -10% | 1 | 1 | 0% | 3,719 | 3,601 | -3% | 0 | 0 | — |
case-03 | fail→fail | 19,860 | 21,042 | +6% | 1 | 1 | 0% | 3,379 | 3,416 | +1% | 0 | 0 | — |
case-04 | fail→pass | 19,727 | 15,390 | -22% | 1 | 1 | 0% | 3,451 | 2,685 | -22% | 0 | 0 | — |
case-05 | fail→pass | 20,200 | 24,831 | +23% | 1 | 1 | 0% | 3,293 | 3,802 | +15% | 0 | 0 | — |
case-06 | fail→pass | 22,039 | 22,162 | +1% | 1 | 1 | 0% | 3,622 | 3,390 | -6% | 0 | 0 | — |
case-07 | pass→pass | 10,913 | 2,100 | -81% | 1 | 1 | 0% | 1,731 | 503 | -71% | 0 | 0 | — |
case-08 | fail→pass | 14,790 | 1,501 | -90% | 1 | 1 | 0% | 2,338 | 393 | -83% | 0 | 0 | — |
case-09 | fail→pass | 11,080 | 1,887 | -83% | 1 | 1 | 0% | 1,772 | 437 | -75% | 0 | 0 | — |
case-10 | pass→pass | 14,518 | 13,415 | -8% | 1 | 1 | 0% | 2,394 | 2,424 | +1% | 0 | 0 | — |
case-11 | pass→pass | 7,907 | 3,806 | -52% | 1 | 1 | 0% | 1,164 | 610 | -48% | 0 | 0 | — |
case-12 | pass→pass | 13,817 | 11,640 | -16% | 1 | 1 | 0% | 2,076 | 2,085 | +0% | 0 | 0 | — |
case-13 | fail→fail | 13,476 | 15,583 | +16% | 1 | 1 | 0% | 2,250 | 2,847 | +27% | 0 | 0 | — |
case-14 | pass→pass | 13,822 | 1,200 | -91% | 1 | 1 | 0% | 2,074 | 334 | -84% | 0 | 0 | — |
case-15 | pass→pass | 16,309 | 14,258 | -13% | 1 | 1 | 0% | 2,518 | 2,510 | -0% | 0 | 0 | — |
case-16 | fail→pass | 19,543 | 20,786 | +6% | 1 | 1 | 0% | 3,179 | 3,149 | -1% | 0 | 0 | — |
case-17 | fail→fail | 13,071 | 3,454 | -74% | 1 | 1 | 0% | 2,084 | 748 | -64% | 0 | 0 | — |
case-18 | pass→pass | 21,223 | 15,523 | -27% | 1 | 1 | 0% | 3,581 | 2,693 | -25% | 0 | 0 | — |
case-19 | fail→pass | 24,175 | 2,370 | -90% | 1 | 1 | 0% | 1,754 | 583 | -67% | 0 | 0 | — |
case-21 | fail→pass | 24,575 | 23,107 | -6% | 1 | 1 | 0% | 4,394 | 4,418 | +1% | 0 | 0 | — |
case-22 | pass→pass | 6,362 | 1,108 | -83% | 1 | 1 | 0% | 922 | 354 | -62% | 0 | 0 | — |
case-23 | pass→pass | 10,919 | 4,153 | -62% | 1 | 1 | 0% | 1,631 | 837 | -49% | 0 | 0 | — |
case-24 | pass→pass | 18,276 | 4,131 | -77% | 1 | 1 | 0% | 3,298 | 797 | -76% | 0 | 0 | — |
case-25 | pass→pass | 12,814 | 3,470 | -73% | 1 | 1 | 0% | 1,910 | 737 | -61% | 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. The headline lift of +32 percentage points is the difference between those two pass rates over the 25 comparable cases.
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.