Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate a Beamer slide deck from the final tutorial and approved module structure. **Trigger**: beamer scaffold, slides from tutorial, tutorial slides, 生成 beamer, 教程幻灯片. **Use when**: `source-tutorial` 的 C4，需要把 `output/TUTORIAL.md` 转成可编译的 `latex/slides/main.tex`。 **Skip if**: 还没有 tutorial 正文。 **Network**: none. **Guardrail**: slides 不能只是机械 heading dump；必须保持模块对齐并适合讲授/轻量自学。
.claude/skills/willoscar-beamer-scaffold/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -74% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -71% | 0% |
source-tutorial 的 C4，需要把 output/TUTORIAL.md 转成可编译的 latex/slides/main.tex。Goal: produce a compile-ready Beamer deck from the final tutorial.
output/TUTORIAL.md is the canonical prose input, while outline/module_plan.yml keeps the slide deck aligned with the approved module order.
output/TUTORIAL.mdoutline/module_plan.ymllatex/slides/main.texuv run python .codex/skills/beamer-scaffold/scripts/run.py --workspace <workspace>--workspace <dir> (required)--unit-id <U###>--inputs <semicolon-separated>--outputs <semicolon-separated>--checkpoint <C#>latex/slides/main.tex from the tutorial:uv run python .codex/skills/beamer-scaffold/scripts/run.py --workspace <workspace>| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 10,503 | 7,561 | -28% | 1 | 1 | 0% | 2,084 | 447 | -79% | 0 | 0 | — |
case-02 | fail→fail | 23,219 | 4,869 | -79% | 1 | 1 | 0% | 5,240 | 442 | -92% | 0 | 0 | — |
case-03 | fail→fail | 28,393 | 4,866 | -83% | 1 | 1 | 0% | 6,169 | 455 | -93% | 0 | 0 | — |
case-04 | pass→pass | 8,581 | 6,022 | -30% | 1 | 1 | 0% | 1,671 | 1,323 | -21% | 0 | 0 | — |
case-05 | pass→pass | 8,844 | 2,613 | -70% | 1 | 1 | 0% | 1,515 | 594 | -61% | 0 | 0 | — |
case-06 | pass→pass | 18,381 | 14,601 | -21% | 1 | 1 | 0% | 4,093 | 3,538 | -14% | 0 | 0 | — |
case-07 | fail→pass | 4,508 | 1,790 | -60% | 1 | 1 | 0% | 753 | 508 | -33% | 0 | 0 | — |
case-08 | fail→pass | 11,603 | 1,894 | -84% | 1 | 1 | 0% | 1,952 | 509 | -74% | 0 | 0 | — |
case-09 | fail→pass | 6,601 | 1,523 | -77% | 1 | 1 | 0% | 1,058 | 446 | -58% | 0 | 0 | — |
case-10 | pass→pass | 6,572 | 1,829 | -72% | 1 | 1 | 0% | 1,202 | 483 | -60% | 0 | 0 | — |
case-11 | fail→pass | 5,513 | 1,846 | -67% | 1 | 1 | 0% | 962 | 464 | -52% | 0 | 0 | — |
case-12 | fail→pass | 7,744 | 1,369 | -82% | 1 | 1 | 0% | 1,381 | 395 | -71% | 0 | 0 | — |
case-13 | fail→pass | 7,241 | 1,385 | -81% | 1 | 1 | 0% | 1,320 | 412 | -69% | 0 | 0 | — |
case-14 | fail→pass | 11,356 | 1,531 | -87% | 1 | 1 | 0% | 1,770 | 434 | -75% | 0 | 0 | — |
case-15 | fail→pass | 6,165 | 1,386 | -78% | 1 | 1 | 0% | 965 | 388 | -60% | 0 | 0 | — |
case-16 | fail→fail | 10,294 | 1,424 | -86% | 1 | 1 | 0% | 1,659 | 434 | -74% | 0 | 0 | — |
case-17 | fail→pass | 8,364 | 1,635 | -80% | 1 | 1 | 0% | 1,620 | 384 | -76% | 0 | 0 | — |
case-18 | fail→pass | 9,378 | 1,310 | -86% | 1 | 1 | 0% | 1,426 | 356 | -75% | 0 | 0 | — |
case-19 | fail→pass | 6,872 | 2,177 | -68% | 1 | 1 | 0% | 1,349 | 546 | -60% | 0 | 0 | — |
case-20 | fail→pass | 8,665 | 1,679 | -81% | 1 | 1 | 0% | 1,315 | 442 | -66% | 0 | 0 | — |
case-21 | fail→pass | 11,597 | 2,044 | -82% | 1 | 1 | 0% | 1,815 | 532 | -71% | 0 | 0 | — |
case-22 | fail→pass | 10,254 | 1,509 | -85% | 1 | 1 | 0% | 1,445 | 464 | -68% | 0 | 0 | — |
case-23 | fail→pass | 8,592 | 2,360 | -73% | 1 | 1 | 0% | 1,483 | 640 | -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 20 counted toward the lift figure. The other 3 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 +65 percentage points is the difference between those two pass rates over the 20 comparable cases.
The publisher has shipped newer versions since this run, so these numbers describe v1, not the version currently listed.
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.