Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →11 paper templates skills. Trigger: starting a new paper, formatting for submission, venue-specific layouts. Design: ready-to-use templates for arXiv preprint, conferences, thesis, and posters.
.claude/skills/brycewang-stanford-templates-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -59% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | arxiv-preprint-template | LaTeX template for arXiv preprints in NIPS/NeurIPS style format | | beamer-presentation-guide | Guide to creating academic presentations with LaTeX Beamer | | conference-paper-template | Templates and formatting guides for major academic conference submissions | | elegant-paper-template | Beautiful LaTeX template for working papers and technical reports | | graphical-abstract-guide | Create SVG graphical abstracts for journal paper submissions | | novathesis-guide | LaTeX thesis template supporting multiple universities and formats | | scientific-article-pdf | Generate publication-ready scientific article PDFs from templates | | sjtuthesis-guide | Write SJTU theses using the SJTUThesis LaTeX template with full compliance | | thesis-template-guide | Set up LaTeX templates for PhD and Master's thesis documents | | thesis-writing-guide | Templates, formatting rules, and strategies for thesis and dissertation writing | | thuthesis-guide | Write Tsinghua University theses using the ThuThesis LaTeX template |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 14,276 | 14,086 | -1% | 1 | 1 | 0% | 2,255 | 1,786 | -21% | 0 | 0 | — |
case-02 | pass→pass | 11,352 | 10,784 | -5% | 1 | 1 | 0% | 2,107 | 2,315 | +10% | 0 | 0 | — |
case-03 | fail→pass | 13,482 | 7,957 | -41% | 1 | 1 | 0% | 2,379 | 1,990 | -16% | 0 | 0 | — |
case-04 | fail→pass | 8,522 | 2,288 | -73% | 1 | 1 | 0% | 1,356 | 898 | -34% | 0 | 0 | — |
case-05 | fail→pass | 10,850 | 2,674 | -75% | 1 | 1 | 0% | 1,781 | 790 | -56% | 0 | 0 | — |
case-06 | pass→pass | 14,301 | 5,549 | -61% | 1 | 1 | 0% | 2,241 | 1,377 | -39% | 0 | 0 | — |
case-07 | pass→fail | 11,933 | 4,388 | -63% | 1 | 1 | 0% | 1,951 | 767 | -61% | 0 | 0 | — |
case-08 | fail→fail | 13,725 | 4,412 | -68% | 1 | 1 | 0% | 2,238 | 738 | -67% | 0 | 0 | — |
case-09 | pass→pass | 14,519 | 9,624 | -34% | 1 | 1 | 0% | 2,749 | 2,259 | -18% | 0 | 0 | — |
case-10 | pass→pass | 13,992 | 5,034 | -64% | 1 | 1 | 0% | 2,144 | 1,164 | -46% | 0 | 0 | — |
case-11 | pass→pass | 12,983 | 6,590 | -49% | 1 | 1 | 0% | 2,052 | 938 | -54% | 0 | 0 | — |
case-12 | fail→pass | 8,827 | 1,605 | -82% | 1 | 1 | 0% | 1,605 | 664 | -59% | 0 | 0 | — |
case-13 | fail→pass | 17,806 | 1,532 | -91% | 1 | 1 | 0% | 3,292 | 661 | -80% | 0 | 0 | — |
case-14 | fail→pass | 7,200 | 1,510 | -79% | 1 | 1 | 0% | 1,370 | 669 | -51% | 0 | 0 | — |
case-15 | fail→pass | 8,558 | 1,562 | -82% | 1 | 1 | 0% | 1,424 | 652 | -54% | 0 | 0 | — |
case-16 | fail→pass | 14,344 | 1,436 | -90% | 1 | 1 | 0% | 2,676 | 627 | -77% | 0 | 0 | — |
case-17 | fail→pass | 10,143 | 2,342 | -77% | 1 | 1 | 0% | 1,677 | 794 | -53% | 0 | 0 | — |
case-18 | fail→pass | 12,863 | 3,040 | -76% | 1 | 1 | 0% | 2,230 | 1,015 | -54% | 0 | 0 | — |
case-19 | fail→pass | 11,078 | 2,149 | -81% | 1 | 1 | 0% | 1,727 | 799 | -54% | 0 | 0 | — |
case-20 | pass→pass | 15,152 | 35,572 | +135% | 1 | 1 | 0% | 2,334 | 3,031 | +30% | 0 | 0 | — |
case-21 | pass→fail | 16,832 | 3,531 | -79% | 1 | 1 | 0% | 2,677 | 688 | -74% | 0 | 0 | — |
case-22 | pass→fail | 17,317 | 4,812 | -72% | 1 | 1 | 0% | 2,985 | 827 | -72% | 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 18 counted toward the lift figure. The other 4 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 +41 percentage points is the difference between those two pass rates over the 18 comparable cases. 3 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.