Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →動画スクリプトからAI画像生成で絵コンテ（ストーリーボード）を作成するスキル。 「絵コンテを作って」「ストーリーボード生成」「スクリプトから画像を作成」等で発動。
.claude/skills/minicoohei-video-storyboard/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 59% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 196% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -66% | 0% |
「絵コンテ」「ストーリーボード」「動画スクリプトから画像生成」「動画企画」
Convert video scripts into visual storyboards with AI-generated images.
GEMINI_API_KEY を設定するか、$KEY として渡す）frames/ がプロジェクト内に存在すること（なければ自動作成）Input: A script from video-scriptwriter with scene descriptions.
responseModalities: ["TEXT", "IMAGE"]scene_XX.pngbashcurl -s "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash-exp:generateContent?key=$KEY" \ -H "Content-Type: application/json" \ -d '{"contents":[{"parts":[{"text":"Generate an image: ..."}]}],"generationConfig":{"responseModalities":["TEXT","IMAGE"]}}'
frames/scene_01.png ... scene_NN.pngOutput feeds into fal.ai i2V for video clip generation.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 12,080 | 14,833 | +23% | 1 | 1 | 0% | 1,866 | 2,967 | +59% | 0 | 0 | — |
case-02 | fail→pass | 7,242 | 16,657 | +130% | 1 | 1 | 0% | 1,176 | 3,483 | +196% | 0 | 0 | — |
case-03 | fail→fail | 10,075 | 3,945 | -61% | 1 | 1 | 0% | 1,552 | 996 | -36% | 0 | 0 | — |
case-04 | pass→pass | 13,107 | 12,529 | -4% | 1 | 1 | 0% | 2,044 | 2,633 | +29% | 0 | 0 | — |
case-05 | pass→pass | 14,426 | 12,140 | -16% | 1 | 1 | 0% | 2,914 | 2,876 | -1% | 0 | 0 | — |
case-06 | pass→pass | 15,322 | 13,623 | -11% | 1 | 1 | 0% | 2,639 | 2,970 | +13% | 0 | 0 | — |
case-07 | fail→pass | 11,095 | 4,218 | -62% | 1 | 1 | 0% | 2,076 | 1,118 | -46% | 0 | 0 | — |
case-08 | pass→pass | 7,345 | 3,305 | -55% | 1 | 1 | 0% | 1,331 | 927 | -30% | 0 | 0 | — |
case-09 | fail→pass | 10,815 | 5,108 | -53% | 1 | 1 | 0% | 1,637 | 1,207 | -26% | 0 | 0 | — |
case-10 | fail→pass | 13,073 | 2,930 | -78% | 1 | 1 | 0% | 2,277 | 765 | -66% | 0 | 0 | — |
case-11 | fail→pass | 7,444 | 2,659 | -64% | 1 | 1 | 0% | 1,305 | 882 | -32% | 0 | 0 | — |
case-12 | pass→pass | 14,711 | 13,625 | -7% | 1 | 1 | 0% | 2,375 | 2,510 | +6% | 0 | 0 | — |
case-13 | fail→pass | 11,292 | 6,332 | -44% | 1 | 1 | 0% | 2,049 | 1,501 | -27% | 0 | 0 | — |
case-14 | pass→pass | 9,477 | 1,851 | -80% | 1 | 1 | 0% | 1,579 | 625 | -60% | 0 | 0 | — |
case-15 | pass→pass | 10,681 | 2,949 | -72% | 1 | 1 | 0% | 1,597 | 859 | -46% | 0 | 0 | — |
case-16 | fail→pass | 13,505 | 5,727 | -58% | 1 | 1 | 0% | 2,171 | 1,260 | -42% | 0 | 0 | — |
case-17 | pass→pass | 12,882 | 10,389 | -19% | 1 | 1 | 0% | 2,203 | 2,143 | -3% | 0 | 0 | — |
case-18 | pass→pass | 11,710 | 6,537 | -44% | 1 | 1 | 0% | 1,772 | 1,370 | -23% | 0 | 0 | — |
case-19 | fail→fail | 5,966 | 2,607 | -56% | 1 | 1 | 0% | 852 | 807 | -5% | 0 | 0 | — |
case-20 | fail→pass | 16,135 | 9,710 | -40% | 1 | 1 | 0% | 2,811 | 2,078 | -26% | 0 | 0 | — |
case-21 | pass→pass | 12,684 | 3,801 | -70% | 1 | 1 | 0% | 2,160 | 1,048 | -51% | 0 | 0 | — |
case-22 | pass→pass | 14,680 | 11,583 | -21% | 1 | 1 | 0% | 2,289 | 2,311 | +1% | 0 | 0 | — |
case-23 | fail→pass | 13,059 | 9,051 | -31% | 1 | 1 | 0% | 2,019 | 1,843 | -9% | 0 | 0 | — |
case-24 | fail→pass | 7,532 | 4,805 | -36% | 1 | 1 | 0% | 1,352 | 1,251 | -7% | 0 | 0 | — |
case-25 | pass→pass | 10,417 | 7,160 | -31% | 1 | 1 | 0% | 1,667 | 1,537 | -8% | 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. 25 cases were attempted. The headline lift of +44 percentage points is the difference between those two pass rates over the 25 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.