Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →日本語翻訳：このファイルは manim-video 用の日本語翻訳が必要です
.claude/skills/affaan-m-manim-video/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 61% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 41% | 0% |
在运动、结构和清晰度比逼真度更重要的技术讲解中，使用 Manim。
manim 命令行用于场景渲染ffmpeg 用于后期处理（如需）video-editing 用于最终合成或润色remotion-video-creation 当最终成品需要合成 UI、字幕或额外运动层时对于社交图谱和网络优化讲解：
使用 assets/network\_graph\_scene.py 作为网络图讲解的起点。
烟雾测试示例：
bashmanim -ql assets/network_graph_scene.py NetworkGraphExplainer
返回：
video-editing 用于最终润色remotion-video-creation 用于运动密集型后期处理或合成content-engine 当动画是更广泛发布的一部分时| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 41,925 | 23,928 | -43% | 1 | 1 | 0% | 4,085 | 4,936 | +21% | 0 | 0 | — |
case-02 | fail→pass | 23,178 | 22,616 | -2% | 1 | 1 | 0% | 3,202 | 4,698 | +47% | 0 | 0 | — |
case-03 | fail→fail | 18,726 | 19,048 | +2% | 1 | 1 | 0% | 2,791 | 4,164 | +49% | 0 | 0 | — |
case-04 | fail→fail | 14,645 | 16,953 | +16% | 1 | 1 | 0% | 2,433 | 3,940 | +62% | 0 | 0 | — |
case-05 | fail→pass | 18,173 | 13,814 | -24% | 1 | 1 | 0% | 3,077 | 2,575 | -16% | 0 | 0 | — |
case-06 | fail→pass | 12,166 | 9,056 | -26% | 1 | 1 | 0% | 1,791 | 2,016 | +13% | 0 | 0 | — |
case-07 | pass→pass | 14,764 | 19,591 | +33% | 1 | 1 | 0% | 2,112 | 4,195 | +99% | 0 | 0 | — |
case-08 | pass→pass | 12,486 | 14,896 | +19% | 1 | 1 | 0% | 1,867 | 2,814 | +51% | 0 | 0 | — |
case-09 | pass→pass | 11,945 | 7,745 | -35% | 1 | 1 | 0% | 1,817 | 1,887 | +4% | 0 | 0 | — |
case-20 | pass→pass | 23,083 | 18,683 | -19% | 1 | 1 | 0% | 3,200 | 3,203 | +0% | 0 | 0 | — |
case-10 | pass→pass | 13,650 | 10,963 | -20% | 1 | 1 | 0% | 2,244 | 2,315 | +3% | 0 | 0 | — |
case-11 | pass→pass | 15,795 | 23,112 | +46% | 1 | 1 | 0% | 2,172 | 4,610 | +112% | 0 | 0 | — |
case-12 | fail→pass | 28,642 | 18,518 | -35% | 1 | 1 | 0% | 2,260 | 3,649 | +61% | 0 | 0 | — |
case-13 | fail→pass | 6,938 | 5,785 | -17% | 1 | 1 | 0% | 1,082 | 1,521 | +41% | 0 | 0 | — |
case-14 | fail→pass | 10,842 | 8,497 | -22% | 1 | 1 | 0% | 1,728 | 2,094 | +21% | 0 | 0 | — |
case-15 | pass→pass | 9,712 | 6,321 | -35% | 1 | 1 | 0% | 1,516 | 1,664 | +10% | 0 | 0 | — |
case-16 | pass→pass | 17,156 | 8,582 | -50% | 1 | 1 | 0% | 2,282 | 2,050 | -10% | 0 | 0 | — |
case-17 | pass→pass | 7,970 | 4,364 | -45% | 1 | 1 | 0% | 1,182 | 1,241 | +5% | 0 | 0 | — |
case-18 | fail→pass | 10,464 | 2,833 | -73% | 1 | 1 | 0% | 1,676 | 1,097 | -35% | 0 | 0 | — |
case-19 | pass→pass | 13,470 | 12,260 | -9% | 1 | 1 | 0% | 1,890 | 2,478 | +31% | 0 | 0 | — |
case-21 | pass→fail | 31,338 | 26,342 | -16% | 1 | 1 | 0% | 2,761 | 5,441 | +97% | 0 | 0 | — |
case-22 | pass→pass | 18,753 | 17,426 | -7% | 1 | 1 | 0% | 3,160 | 3,590 | +14% | 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. The headline lift of +27 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.