Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Import SOP: Medium-depth literature search, AI summary report (from literature-engine)
.claude/skills/yogsoth-ai-hypothesis-formation-paper-search/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-17 | ✓→✗ | ▼ Worse | -90% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 16% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 16% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | -14% | 0% |
| case-10 | ✓→✓ | ＝ Same ✓ | -50% | 0% |
Medium-depth literature search capability imported from the literature-engine skill library.
Directly call the paper-search skill in literature-engine. This skill provides:
<!-- 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-01 | fail→fail | 19,913 | 15,989 | -20% | 1 | 1 | 0% | 2,859 | 528 | -82% | 0 | 0 | — |
case-02 | fail→fail | 33,828 | 36,781 | +9% | 1 | 1 | 0% | 5,077 | 5,653 | +11% | 0 | 0 | — |
case-03 | fail→fail | 26,481 | 7,398 | -72% | 1 | 1 | 0% | 4,328 | 514 | -88% | 0 | 0 | — |
case-04 | pass→pass | 7,554 | 8,372 | +11% | 1 | 1 | 0% | 1,529 | 1,779 | +16% | 0 | 0 | — |
case-05 | pass→pass | 14,475 | 15,177 | +5% | 1 | 1 | 0% | 3,133 | 3,641 | +16% | 0 | 0 | — |
case-06 | pass→pass | 15,297 | 4,907 | -68% | 1 | 1 | 0% | 1,022 | 876 | -14% | 0 | 0 | — |
case-07 | fail→fail | 17,898 | 2,959 | -83% | 1 | 1 | 0% | 2,873 | 608 | -79% | 0 | 0 | — |
case-08 | fail→fail | 27,963 | 38,646 | +38% | 1 | 1 | 0% | 3,874 | 5,924 | +53% | 0 | 0 | — |
case-09 | fail→fail | 23,346 | 9,001 | -61% | 1 | 1 | 0% | 3,467 | 818 | -76% | 0 | 0 | — |
case-10 | pass→pass | 22,295 | 8,534 | -62% | 1 | 1 | 0% | 3,245 | 1,627 | -50% | 0 | 0 | — |
case-11 | fail→fail | 22,608 | 6,542 | -71% | 1 | 1 | 0% | 3,453 | 614 | -82% | 0 | 0 | — |
case-12 | fail→fail | 27,817 | 5,919 | -79% | 1 | 1 | 0% | 4,480 | 584 | -87% | 0 | 0 | — |
case-13 | fail→fail | 24,250 | 6,089 | -75% | 1 | 1 | 0% | 3,702 | 442 | -88% | 0 | 0 | — |
case-14 | fail→fail | 21,383 | 6,177 | -71% | 1 | 1 | 0% | 3,376 | 666 | -80% | 0 | 0 | — |
case-15 | pass→pass | 21,830 | 33,841 | +55% | 1 | 1 | 0% | 3,375 | 5,599 | +66% | 0 | 0 | — |
case-16 | fail→fail | 22,884 | 6,823 | -70% | 1 | 1 | 0% | 3,656 | 694 | -81% | 0 | 0 | — |
case-17 | pass→fail | 31,334 | 6,222 | -80% | 1 | 1 | 0% | 4,736 | 451 | -90% | 0 | 0 | — |
case-18 | pass→pass | 26,923 | 37,244 | +38% | 1 | 1 | 0% | 3,971 | 5,241 | +32% | 0 | 0 | — |
case-19 | fail→fail | 29,886 | 40,440 | +35% | 1 | 1 | 0% | 4,281 | 6,385 | +49% | 0 | 0 | — |
case-20 | fail→fail | 26,197 | 5,704 | -78% | 1 | 1 | 0% | 4,016 | 514 | -87% | 0 | 0 | — |
case-21 | fail→fail | 23,271 | 5,393 | -77% | 1 | 1 | 0% | 3,480 | 437 | -87% | 0 | 0 | — |
case-22 | fail→fail | 24,114 | 9,612 | -60% | 1 | 1 | 0% | 3,754 | 1,198 | -68% | 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 10 counted toward the lift figure. The other 12 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 -5 percentage points is the difference between those two pass rates over the 10 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.