Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Abstract-level paper scanning for broad coverage. Import of literature-engine/literature-overview skill. Abstract-level only — no methodology conclusions from abstracts.
.claude/skills/yogsoth-ai-creative-ideation-paper-overview/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -57% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -19% | 0% |
Abstract-level paper scanning for broad coverage.
Import — strictly follow literature-engine/literature-overview skill protocol.
Abstract-level only — do not draw methodology conclusions from abstracts alone. Abstracts provide leads for deeper investigation (paper-search or paper-research).
Quantity target is set by the calling strategy's budget table. This SOP executes one unit = one paper scanned at abstract level.
literature-engine repo → skills/literature-overview/SKILL.md
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | literature-overview | Quick landscape scan — discover papers on a topic without full-text reading |
<!-- 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-12 | pass→pass | 16,664 | 13,185 | -21% | 1 | 1 | 0% | 2,585 | 2,467 | -5% | 0 | 0 | — |
case-01 | fail→pass | 20,661 | 20,193 | -2% | 1 | 1 | 0% | 3,756 | 3,107 | -17% | 0 | 0 | — |
case-02 | fail→pass | 15,139 | 11,802 | -22% | 1 | 1 | 0% | 2,755 | 2,218 | -19% | 0 | 0 | — |
case-03 | pass→pass | 17,020 | 11,185 | -34% | 1 | 1 | 0% | 2,809 | 1,875 | -33% | 0 | 0 | — |
case-04 | fail→pass | 27,796 | 12,497 | -55% | 1 | 1 | 0% | 5,888 | 2,539 | -57% | 0 | 0 | — |
case-05 | fail→pass | 18,112 | 9,529 | -47% | 1 | 1 | 0% | 2,856 | 1,712 | -40% | 0 | 0 | — |
case-06 | fail→pass | 16,884 | 11,427 | -32% | 1 | 1 | 0% | 2,428 | 1,970 | -19% | 0 | 0 | — |
case-13 | fail→pass | 24,998 | 13,108 | -48% | 1 | 1 | 0% | 4,042 | 2,231 | -45% | 0 | 0 | — |
case-07 | fail→pass | 16,902 | 10,443 | -38% | 1 | 1 | 0% | 2,612 | 1,814 | -31% | 0 | 0 | — |
case-08 | fail→fail | 22,196 | 16,008 | -28% | 1 | 1 | 0% | 3,492 | 2,712 | -22% | 0 | 0 | — |
case-09 | fail→fail | 13,031 | 16,322 | +25% | 1 | 1 | 0% | 2,113 | 3,092 | +46% | 0 | 0 | — |
case-10 | pass→pass | 18,491 | 8,914 | -52% | 1 | 1 | 0% | 3,122 | 1,559 | -50% | 0 | 0 | — |
case-11 | fail→fail | 26,850 | 22,625 | -16% | 1 | 1 | 0% | 4,759 | 3,963 | -17% | 0 | 0 | — |
case-14 | pass→pass | 20,538 | 16,176 | -21% | 1 | 1 | 0% | 3,387 | 2,906 | -14% | 0 | 0 | — |
case-15 | fail→fail | 13,756 | 12,338 | -10% | 1 | 1 | 0% | 2,098 | 1,993 | -5% | 0 | 0 | — |
case-16 | pass→pass | 20,746 | 13,136 | -37% | 1 | 1 | 0% | 3,255 | 2,243 | -31% | 0 | 0 | — |
case-17 | fail→fail | 32,128 | 16,492 | -49% | 1 | 1 | 0% | 5,101 | 3,049 | -40% | 0 | 0 | — |
case-22 | pass→pass | 4,898 | 3,276 | -33% | 1 | 1 | 0% | 928 | 865 | -7% | 0 | 0 | — |
case-18 | pass→pass | 17,869 | 8,194 | -54% | 1 | 1 | 0% | 2,731 | 1,505 | -45% | 0 | 0 | — |
case-19 | pass→pass | 13,744 | 18,952 | +38% | 1 | 1 | 0% | 2,128 | 2,280 | +7% | 0 | 0 | — |
case-20 | fail→fail | 4,230 | 2,887 | -32% | 1 | 1 | 0% | 584 | 633 | +8% | 0 | 0 | — |
case-21 | pass→pass | 7,221 | 5,762 | -20% | 1 | 1 | 0% | 659 | 804 | +22% | 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. The headline lift of +32 percentage points is the difference between those two pass rates over the 22 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.