Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →7 education research skills. Trigger: pedagogical research, course design, learning analytics, assessment. Design: evidence-based teaching methods and educational measurement tools.
.claude/skills/brycewang-stanford-education-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-22 | ✗→✓ | ▲ Improved | -29% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -58% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | academic-study-methods | Evidence-based study techniques for academic learning and retention | | assessment-design-guide | Psychometrics and educational assessment design for researchers | | curriculum-design-guide | Systematic approaches to curriculum design using backward design and alignment | | educational-research-methods | Quantitative and qualitative research methods for education studies | | learning-science-guide | Evidence-based learning science principles for educational research and practice | | mooc-analytics-guide | Analyzing MOOC data, learning analytics, and online education metrics | | open-syllabus-api | Analyze most-taught books and texts via Open Syllabus analytics |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-22 | fail→pass | 9,905 | 4,938 | -50% | 1 | 1 | 0% | 1,482 | 1,051 | -29% | 0 | 0 | — |
case-01 | fail→fail | 40,098 | 63,675 | +59% | 1 | 1 | 0% | 1,757 | 553 | -69% | 0 | 0 | — |
case-02 | fail→fail | 49,038 | 5,821 | -88% | 1 | 1 | 0% | 3,085 | 643 | -79% | 0 | 0 | — |
case-03 | pass→fail | 12,638 | 54,342 | +330% | 1 | 1 | 0% | 2,112 | 2,255 | +7% | 0 | 0 | — |
case-04 | fail→pass | 9,749 | 5,036 | -48% | 1 | 1 | 0% | 1,595 | 985 | -38% | 0 | 0 | — |
case-05 | fail→pass | 11,956 | 33,341 | +179% | 1 | 1 | 0% | 1,634 | 768 | -53% | 0 | 0 | — |
case-06 | fail→pass | 8,617 | 3,568 | -59% | 1 | 1 | 0% | 1,375 | 737 | -46% | 0 | 0 | — |
case-07 | fail→pass | 10,058 | 2,407 | -76% | 1 | 1 | 0% | 1,735 | 731 | -58% | 0 | 0 | — |
case-08 | fail→pass | 8,728 | 2,152 | -75% | 1 | 1 | 0% | 1,615 | 691 | -57% | 0 | 0 | — |
case-09 | pass→fail | 10,785 | 4,729 | -56% | 1 | 1 | 0% | 1,763 | 671 | -62% | 0 | 0 | — |
case-10 | pass→pass | 18,222 | 10,061 | -45% | 1 | 1 | 0% | 3,020 | 1,955 | -35% | 0 | 0 | — |
case-11 | pass→fail | 13,000 | 48,980 | +277% | 1 | 1 | 0% | 2,176 | 705 | -68% | 0 | 0 | — |
case-12 | fail→pass | 10,467 | 2,082 | -80% | 1 | 1 | 0% | 1,717 | 665 | -61% | 0 | 0 | — |
case-13 | pass→pass | 13,544 | 4,024 | -70% | 1 | 1 | 0% | 2,032 | 930 | -54% | 0 | 0 | — |
case-14 | pass→pass | 8,408 | 3,085 | -63% | 1 | 1 | 0% | 1,369 | 804 | -41% | 0 | 0 | — |
case-15 | fail→pass | 5,552 | 3,258 | -41% | 1 | 1 | 0% | 912 | 723 | -21% | 0 | 0 | — |
case-16 | pass→pass | 11,184 | 3,808 | -66% | 1 | 1 | 0% | 1,801 | 714 | -60% | 0 | 0 | — |
case-17 | pass→fail | 11,229 | 5,261 | -53% | 1 | 1 | 0% | 1,753 | 693 | -60% | 0 | 0 | — |
case-18 | fail→pass | 14,279 | 35,705 | +150% | 1 | 1 | 0% | 1,925 | 679 | -65% | 0 | 0 | — |
case-19 | pass→pass | 11,518 | 3,533 | -69% | 1 | 1 | 0% | 1,724 | 895 | -48% | 0 | 0 | — |
case-20 | fail→pass | 10,937 | 4,268 | -61% | 1 | 1 | 0% | 1,428 | 927 | -35% | 0 | 0 | — |
case-21 | fail→pass | 10,316 | 5,176 | -50% | 1 | 1 | 0% | 1,512 | 1,107 | -27% | 0 | 0 | — |
case-23 | fail→fail | 5,015 | 9,516 | +90% | 1 | 1 | 0% | 750 | 1,682 | +124% | 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 17 counted toward the lift figure. The other 6 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 +30 percentage points is the difference between those two pass rates over the 17 comparable cases. 4 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.