Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →11 latex skills. Trigger: LaTeX typesetting, formatting papers, mathematical notation, Beamer. Design: template-based guides with package recommendations and compilation tips.
.claude/skills/brycewang-stanford-latex-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -67% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -71% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | academic-writing-latex | LaTeX-based academic writing assistant for thesis and paper templates | | bibliography-management-guide | Manage references with BibLaTeX, natbib, and LaTeX bibliography styles | | latex-drawing-collection | LaTeX drawing examples for Bayesian networks, tensors, and diagrams | | latex-drawing-guide | TikZ and PGFPlots techniques for publication-quality scientific figures | | latex-ecosystem-guide | Comprehensive guide to LaTeX editors, packages, and typesetting workflows | | latex-templates-collection | Collection of LaTeX templates for papers, presentations, and CVs | | math-typesetting-guide | LaTeX math typesetting, equation formatting, and cross-referencing | | md-to-pdf-academic | Convert Markdown to publication-quality PDF with LaTeX math rendering | | overleaf-collaboration-guide | Guide to collaborative LaTeX editing with Overleaf | | tex-render-guide | Render LaTeX math expressions to images in PNG, JPEG, and SVG | | tikz-diagrams-guide | Create publication-quality scientific diagrams with TikZ in LaTeX |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-17 | fail→fail | 13,072 | 2,917 | -78% | 1 | 1 | 0% | 2,344 | 890 | -62% | 0 | 0 | — |
case-01 | fail→fail | 10,740 | 23,702 | +121% | 1 | 1 | 0% | 1,698 | 749 | -56% | 0 | 0 | — |
case-02 | fail→pass | 10,007 | 3,188 | -68% | 1 | 1 | 0% | 1,427 | 899 | -37% | 0 | 0 | — |
case-03 | fail→pass | 11,055 | 6,146 | -44% | 1 | 1 | 0% | 1,501 | 883 | -41% | 0 | 0 | — |
case-04 | fail→pass | 15,055 | 3,045 | -80% | 1 | 1 | 0% | 2,448 | 894 | -63% | 0 | 0 | — |
case-05 | fail→pass | 18,634 | 3,462 | -81% | 1 | 1 | 0% | 3,246 | 1,085 | -67% | 0 | 0 | — |
case-06 | fail→fail | 18,733 | 3,785 | -80% | 1 | 1 | 0% | 3,018 | 655 | -78% | 0 | 0 | — |
case-07 | fail→pass | 17,309 | 4,045 | -77% | 1 | 1 | 0% | 2,845 | 824 | -71% | 0 | 0 | — |
case-08 | fail→pass | 13,899 | 8,049 | -42% | 1 | 1 | 0% | 2,438 | 1,368 | -44% | 0 | 0 | — |
case-09 | fail→pass | 11,313 | 10,865 | -4% | 1 | 1 | 0% | 1,947 | 2,272 | +17% | 0 | 0 | — |
case-10 | fail→fail | 14,216 | 6,397 | -55% | 1 | 1 | 0% | 2,148 | 1,067 | -50% | 0 | 0 | — |
case-11 | fail→fail | 16,040 | 3,759 | -77% | 1 | 1 | 0% | 2,376 | 924 | -61% | 0 | 0 | — |
case-12 | fail→pass | 18,687 | 4,088 | -78% | 1 | 1 | 0% | 3,279 | 1,173 | -64% | 0 | 0 | — |
case-13 | fail→fail | 21,963 | 3,325 | -85% | 1 | 1 | 0% | 3,165 | 1,046 | -67% | 0 | 0 | — |
case-14 | fail→pass | 9,533 | 3,365 | -65% | 1 | 1 | 0% | 1,514 | 969 | -36% | 0 | 0 | — |
case-15 | fail→pass | 13,434 | 1,916 | -86% | 1 | 1 | 0% | 2,087 | 731 | -65% | 0 | 0 | — |
case-16 | fail→pass | 10,035 | 3,283 | -67% | 1 | 1 | 0% | 1,395 | 914 | -34% | 0 | 0 | — |
case-18 | fail→fail | 13,336 | 4,838 | -64% | 1 | 1 | 0% | 1,949 | 1,203 | -38% | 0 | 0 | — |
case-19 | fail→pass | 9,928 | 3,071 | -69% | 1 | 1 | 0% | 1,408 | 922 | -35% | 0 | 0 | — |
case-20 | pass→pass | 6,013 | 5,551 | -8% | 1 | 1 | 0% | 988 | 1,349 | +37% | 0 | 0 | — |
case-21 | pass→pass | 6,575 | 4,315 | -34% | 1 | 1 | 0% | 986 | 1,108 | +12% | 0 | 0 | — |
case-22 | pass→pass | 7,635 | 7,569 | -1% | 1 | 1 | 0% | 1,234 | 1,622 | +31% | 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 20 counted toward the lift figure. The other 2 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 20 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.