Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →5 ecology & environmental science skills. Trigger: biodiversity surveys, species data, environmental monitoring. Design: field data collection, spatial analysis, and conservation biology workflows.
.claude/skills/brycewang-stanford-ecology-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -75% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -64% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -79% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | biodiversity-data-guide | Biodiversity data access, species occurrence, and ecological tools | | conservation-biology-guide | Apply conservation biology methods, databases, and assessment tools | | gbif-api | Global biodiversity data API for species occurrences and datasets | | inaturalist-api | Citizen science platform API for biodiversity observations | | species-distribution-guide | Species distribution modeling with MaxEnt, SDM methods, and GBIF data |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 14,017 | 15,230 | +9% | 1 | 1 | 0% | 2,579 | 1,777 | -31% | 0 | 0 | — |
case-02 | fail→fail | 22,053 | 3,481 | -84% | 1 | 1 | 0% | 3,725 | 455 | -88% | 0 | 0 | — |
case-03 | fail→fail | 19,877 | 7,818 | -61% | 1 | 1 | 0% | 3,034 | 554 | -82% | 0 | 0 | — |
case-04 | fail→pass | 9,349 | 2,500 | -73% | 1 | 1 | 0% | 1,666 | 681 | -59% | 0 | 0 | — |
case-05 | fail→fail | 14,963 | 2,685 | -82% | 1 | 1 | 0% | 2,599 | 644 | -75% | 0 | 0 | — |
case-06 | fail→pass | 14,541 | 2,735 | -81% | 1 | 1 | 0% | 2,613 | 652 | -75% | 0 | 0 | — |
case-07 | fail→pass | 8,720 | 2,042 | -77% | 1 | 1 | 0% | 1,552 | 576 | -63% | 0 | 0 | — |
case-08 | fail→pass | 8,208 | 3,332 | -59% | 1 | 1 | 0% | 1,378 | 501 | -64% | 0 | 0 | — |
case-09 | fail→pass | 13,405 | 1,947 | -85% | 1 | 1 | 0% | 2,397 | 493 | -79% | 0 | 0 | — |
case-10 | fail→fail | 16,607 | 4,086 | -75% | 1 | 1 | 0% | 2,855 | 520 | -82% | 0 | 0 | — |
case-11 | fail→pass | 11,347 | 2,277 | -80% | 1 | 1 | 0% | 1,849 | 649 | -65% | 0 | 0 | — |
case-12 | fail→pass | 12,842 | 1,935 | -85% | 1 | 1 | 0% | 2,411 | 549 | -77% | 0 | 0 | — |
case-17 | fail→pass | 9,550 | 2,615 | -73% | 1 | 1 | 0% | 1,784 | 688 | -61% | 0 | 0 | — |
case-13 | fail→pass | 13,515 | 2,049 | -85% | 1 | 1 | 0% | 2,199 | 563 | -74% | 0 | 0 | — |
case-14 | fail→pass | 9,625 | 1,900 | -80% | 1 | 1 | 0% | 1,662 | 531 | -68% | 0 | 0 | — |
case-15 | fail→fail | 18,308 | 4,553 | -75% | 1 | 1 | 0% | 1,486 | 604 | -59% | 0 | 0 | — |
case-16 | fail→pass | 18,985 | 2,375 | -87% | 1 | 1 | 0% | 3,216 | 654 | -80% | 0 | 0 | — |
case-18 | fail→pass | 12,917 | 2,534 | -80% | 1 | 1 | 0% | 2,023 | 661 | -67% | 0 | 0 | — |
case-19 | fail→fail | 15,413 | 4,265 | -72% | 1 | 1 | 0% | 2,615 | 532 | -80% | 0 | 0 | — |
case-20 | pass→pass | 26,571 | 24,653 | -7% | 1 | 1 | 0% | 4,479 | 4,177 | -7% | 0 | 0 | — |
case-21 | pass→pass | 19,685 | 15,215 | -23% | 1 | 1 | 0% | 3,778 | 3,184 | -16% | 0 | 0 | — |
case-22 | pass→pass | 25,191 | 22,583 | -10% | 1 | 1 | 0% | 4,494 | 4,317 | -4% | 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 17 counted toward the lift figure. The other 5 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 +55 percentage points is the difference between those two pass rates over the 17 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.