Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →9 grants & funding skills. Trigger: grant applications, funding search, budget planning, data repositories. Design: funder-specific guides with eligibility, submission requirements, and timelines.
.claude/skills/brycewang-stanford-funding-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-22 | ✗→✓ | ▲ Improved | -57% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | eu-horizon-guide | Navigate EU Horizon Europe funding programs and proposal writing | | figshare-api | Research data sharing and repository | | grant-budget-guide | Prepare and justify research grant budgets across funding agencies | | grant-writing-guide | Write competitive research proposals with clear objectives and budgets | | nih-reporter-api-guide | Search NIH-funded grants and research projects via RePORTER API | | nsf-award-api-guide | Search NSF awards and grants with free public API, no auth required | | nsf-grant-guide | Navigate NSF grant applications, program selection, and strategies | | open-science-guide | Pre-registration, open data, and FAIR principles for research | | zenodo-api | Open research repository for all disciplines |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-23 | pass→fail | 6,672 | 7,145 | +7% | 1 | 1 | 0% | 858 | 1,568 | +83% | 0 | 0 | — |
case-01 | pass→fail | 7,359 | 65,934 | +796% | 1 | 1 | 0% | 1,453 | 606 | -58% | 0 | 0 | — |
case-02 | pass→fail | 9,161 | 3,816 | -58% | 1 | 1 | 0% | 1,885 | 627 | -67% | 0 | 0 | — |
case-03 | pass→fail | 7,517 | 38,463 | +412% | 1 | 1 | 0% | 1,300 | 593 | -54% | 0 | 0 | — |
case-04 | pass→fail | 8,714 | 3,244 | -63% | 1 | 1 | 0% | 1,321 | 598 | -55% | 0 | 0 | — |
case-05 | pass→fail | 13,459 | 4,271 | -68% | 1 | 1 | 0% | 2,246 | 661 | -71% | 0 | 0 | — |
case-06 | fail→fail | 15,950 | 3,399 | -79% | 1 | 1 | 0% | 2,447 | 547 | -78% | 0 | 0 | — |
case-07 | pass→fail | 10,463 | 3,313 | -68% | 1 | 1 | 0% | 1,856 | 532 | -71% | 0 | 0 | — |
case-08 | pass→fail | 8,929 | 3,617 | -59% | 1 | 1 | 0% | 1,532 | 627 | -59% | 0 | 0 | — |
case-09 | fail→fail | 10,157 | 2,543 | -75% | 1 | 1 | 0% | 1,740 | 533 | -69% | 0 | 0 | — |
case-10 | pass→pass | 6,251 | 5,654 | -10% | 1 | 1 | 0% | 950 | 719 | -24% | 0 | 0 | — |
case-11 | pass→fail | 2,897 | 4,993 | +72% | 1 | 1 | 0% | 435 | 708 | +63% | 0 | 0 | — |
case-12 | pass→pass | 2,321 | 4,836 | +108% | 1 | 1 | 0% | 342 | 609 | +78% | 0 | 0 | — |
case-13 | pass→pass | 2,565 | 2,885 | +12% | 1 | 1 | 0% | 417 | 782 | +88% | 0 | 0 | — |
case-14 | pass→pass | 3,135 | 2,904 | -7% | 1 | 1 | 0% | 536 | 738 | +38% | 0 | 0 | — |
case-15 | pass→pass | 2,879 | 2,597 | -10% | 1 | 1 | 0% | 424 | 788 | +86% | 0 | 0 | — |
case-16 | fail→pass | 10,465 | 3,750 | -64% | 1 | 1 | 0% | 1,559 | 899 | -42% | 0 | 0 | — |
case-17 | fail→pass | 4,812 | 6,339 | +32% | 1 | 1 | 0% | 700 | 898 | +28% | 0 | 0 | — |
case-18 | fail→pass | 8,485 | 3,580 | -58% | 1 | 1 | 0% | 1,277 | 899 | -30% | 0 | 0 | — |
case-19 | fail→fail | 14,526 | 4,664 | -68% | 1 | 1 | 0% | 2,260 | 726 | -68% | 0 | 0 | — |
case-20 | fail→pass | 12,825 | 28,563 | +123% | 1 | 1 | 0% | 2,043 | 2,265 | +11% | 0 | 0 | — |
case-21 | pass→fail | 16,602 | 5,112 | -69% | 1 | 1 | 0% | 2,454 | 789 | -68% | 0 | 0 | — |
case-22 | fail→pass | 21,747 | 9,052 | -58% | 1 | 1 | 0% | 2,986 | 1,275 | -57% | 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. 23 cases were attempted, and 11 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 -22 percentage points is the difference between those two pass rates over the 11 comparable cases. 12 cases got worse with the skill loaded, and they are included in that figure.
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.