Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →動画分析結果からタイプ別Playbookに知見を蓄積・活用するスキル。 video-analyzerの出力template.jsonを入力として使う。 「Playbook更新」「動画タイプ別知見」「Playbook確認」等で発動。
.claude/skills/minicoohei-video-playbook/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-17 | ✗→✓ | ▲ Improved | -9% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -44% | 0% |
動画分析結果（template.json）から動画タイプを判定し、タイプ別Playbookに制作知見を蓄積・活用する。
| タイプ | 説明 | |--------|------| | intro | 紹介・レビュー（商品紹介、サービス紹介、人物紹介） | | teaching | ティーチング・解説（ハウツー、知識共有、tips、ノウハウ） | | template | テンプレート・トレンド（流行りのフォーマット、音源同期、チャレンジ） | | meme | Meme・ネタ（オチ重視、ユーモア、パロディ） | | dance | ダンス・パフォーマンス（振付、BPM同期、カバー） | | mv | MV・シネマティック（音楽映像、エフェクト重視、映画的演出） | | clip | 切り抜き・ハイライト（長尺→短尺、名場面、配信切り抜き） |
bash# video-analyzerで分析後、Playbookに知見を追加 python skills/video-playbook/scripts/manage_playbook.py \ --add -t output/templates/video_001/template.json # タイプ別Playbook一覧 python skills/video-playbook/scripts/manage_playbook.py --list # 特定タイプのPlaybook表示 python skills/video-playbook/scripts/manage_playbook.py --show teaching # Markdown形式でエクスポート python skills/video-playbook/scripts/manage_playbook.py --export teaching
1. video-analyzer で動画分析 → template.json
2. manage_playbook.py --add -t template.json
→ 動画タイプ自動判定
→ タイミング・構成・テロップ等の知見抽出
→ タイプ別playbook JSONに追加
→ 集計データ自動更新
3. manage_playbook.py --show TYPE で蓄積知見を確認
4. 新動画制作時にPlaybookを参照各分析結果から以下の知見を抽出し、タイプ別に蓄積：
サンプルが増えるほど集計データ（aggregated）の精度が上がり、 「このタイプの動画はこう作るべき」という制作ガイドラインが自動生成される。
Playbookの知見を使って新動画を作る際：
--show TYPE で対象タイプの知見を確認--export TYPE でMarkdownサマリーを生成skills/video-playbook/playbooks/
├── teaching.json # ティーチング系の知見
├── intro.json # 紹介系の知見
├── meme.json # Meme系の知見
└── ...| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-08 | pass→pass | 16,363 | 1,861 | -89% | 1 | 1 | 0% | 2,675 | 985 | -63% | 0 | 0 | — |
case-09 | pass→pass | 15,568 | 1,974 | -87% | 1 | 1 | 0% | 2,488 | 1,056 | -58% | 0 | 0 | — |
case-10 | pass→pass | 14,687 | 1,660 | -89% | 1 | 1 | 0% | 2,362 | 1,011 | -57% | 0 | 0 | — |
case-11 | pass→pass | 13,143 | 1,945 | -85% | 1 | 1 | 0% | 2,102 | 1,101 | -48% | 0 | 0 | — |
case-17 | fail→pass | 7,050 | 1,591 | -77% | 1 | 1 | 0% | 1,077 | 982 | -9% | 0 | 0 | — |
case-18 | pass→pass | 12,435 | 1,983 | -84% | 1 | 1 | 0% | 1,982 | 1,046 | -47% | 0 | 0 | — |
case-01 | fail→pass | 11,374 | 2,662 | -77% | 1 | 1 | 0% | 1,896 | 1,141 | -40% | 0 | 0 | — |
case-07 | pass→pass | 19,516 | 2,338 | -88% | 1 | 1 | 0% | 3,013 | 1,125 | -63% | 0 | 0 | — |
case-02 | pass→fail | 16,569 | 4,557 | -72% | 1 | 1 | 0% | 2,392 | 1,007 | -58% | 0 | 0 | — |
case-03 | pass→pass | 17,641 | 9,655 | -45% | 1 | 1 | 0% | 2,549 | 2,175 | -15% | 0 | 0 | — |
case-04 | fail→pass | 8,354 | 1,521 | -82% | 1 | 1 | 0% | 1,419 | 971 | -32% | 0 | 0 | — |
case-05 | pass→pass | 17,664 | 2,936 | -83% | 1 | 1 | 0% | 2,964 | 1,228 | -59% | 0 | 0 | — |
case-06 | pass→pass | 14,653 | 1,830 | -88% | 1 | 1 | 0% | 2,271 | 1,089 | -52% | 0 | 0 | — |
case-12 | fail→pass | 10,128 | 1,473 | -85% | 1 | 1 | 0% | 1,648 | 963 | -42% | 0 | 0 | — |
case-13 | fail→pass | 11,367 | 1,355 | -88% | 1 | 1 | 0% | 1,687 | 946 | -44% | 0 | 0 | — |
case-14 | pass→pass | 15,359 | 8,821 | -43% | 1 | 1 | 0% | 2,345 | 2,091 | -11% | 0 | 0 | — |
case-15 | pass→pass | 8,902 | 2,447 | -73% | 1 | 1 | 0% | 1,476 | 1,060 | -28% | 0 | 0 | — |
case-16 | pass→pass | 12,580 | 4,668 | -63% | 1 | 1 | 0% | 1,962 | 1,430 | -27% | 0 | 0 | — |
case-19 | pass→pass | 8,865 | 2,298 | -74% | 1 | 1 | 0% | 1,389 | 1,061 | -24% | 0 | 0 | — |
case-20 | pass→pass | 2,845 | 2,169 | -24% | 1 | 1 | 0% | 458 | 1,049 | +129% | 0 | 0 | — |
case-21 | pass→pass | 4,480 | 2,046 | -54% | 1 | 1 | 0% | 766 | 1,072 | +40% | 0 | 0 | — |
case-22 | pass→pass | 2,643 | 1,951 | -26% | 1 | 1 | 0% | 421 | 1,067 | +153% | 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 21 counted toward the lift figure. The other 1 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 +18 percentage points is the difference between those two pass rates over the 21 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.