Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →マーケティングコンテンツ制作スキル。X/Instagram投稿、Note/Medium記事、 バナー画像、動画スクリプト等を作成する。 「投稿作って」「バナー作成」「記事書いて」「コピー作成」等で発動。 product-context.mdを参照してブランド一貫性を保つ。
.claude/skills/minicoohei-content-creator/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 36% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 49% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -13% | 0% |
マーケティングコンテンツを制作するスキル。プロダクトコンテキストに基づいてブランド一貫性のあるコンテンツを生成する。
marketing/product-context.md が存在すること。なければ marketing-planner スキルで先に作成を促す。
入力: テーマ or プロンプト 出力: 投稿テキスト + ハッシュタグ + （必要なら）画像プロンプト
フォーマット → references/post-formats.md
入力: テーマ、ターゲット、目的 出力: 構成案 → 本文 → メタ情報
プロセス:
入力: 用途、テキスト、スタイル 出力: Gemini画像生成プロンプト → 生成 → 確認
サイズガイド:
入力: テーマ、尺、プラットフォーム 出力: スクリプト（セリフ + 画面指示）
入力: 目的、ターゲット、シーケンス位置 出力: 件名 + 本文 + CTA
詳細 → references/email-templates.md
生成後に必ず確認:
「1週間分の投稿作って」等のリクエストに対応:
marketing/drafts/ に保存marketing-planner — コンテキスト文書作成post-publisher — 制作コンテンツの投稿実行copywriting, copy-editing, social-content, email-sequence| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 20,503 | 12,147 | -41% | 1 | 1 | 0% | 2,907 | 2,449 | -16% | 0 | 0 | — |
case-02 | fail→fail | 11,747 | 7,799 | -34% | 1 | 1 | 0% | 1,766 | 1,833 | +4% | 0 | 0 | — |
case-03 | fail→fail | 14,579 | 10,331 | -29% | 1 | 1 | 0% | 2,181 | 2,196 | +1% | 0 | 0 | — |
case-04 | fail→pass | 8,943 | 8,040 | -10% | 1 | 1 | 0% | 1,470 | 1,997 | +36% | 0 | 0 | — |
case-05 | pass→pass | 8,818 | 9,241 | +5% | 1 | 1 | 0% | 1,401 | 2,100 | +50% | 0 | 0 | — |
case-06 | fail→pass | 8,545 | 7,528 | -12% | 1 | 1 | 0% | 1,283 | 1,916 | +49% | 0 | 0 | — |
case-07 | fail→fail | 5,318 | 2,105 | -60% | 1 | 1 | 0% | 954 | 1,081 | +13% | 0 | 0 | — |
case-08 | pass→pass | 8,387 | 3,440 | -59% | 1 | 1 | 0% | 1,482 | 1,349 | -9% | 0 | 0 | — |
case-09 | pass→pass | 4,241 | 2,479 | -42% | 1 | 1 | 0% | 644 | 1,022 | +59% | 0 | 0 | — |
case-10 | fail→pass | 4,924 | 1,820 | -63% | 1 | 1 | 0% | 798 | 944 | +18% | 0 | 0 | — |
case-11 | pass→pass | 13,426 | 8,450 | -37% | 1 | 1 | 0% | 2,223 | 1,828 | -18% | 0 | 0 | — |
case-12 | pass→pass | 15,201 | 9,328 | -39% | 1 | 1 | 0% | 2,579 | 2,281 | -12% | 0 | 0 | — |
case-13 | fail→pass | 6,759 | 1,946 | -71% | 1 | 1 | 0% | 1,170 | 1,018 | -13% | 0 | 0 | — |
case-14 | pass→fail | 9,301 | 6,201 | -33% | 1 | 1 | 0% | 1,452 | 1,767 | +22% | 0 | 0 | — |
case-15 | fail→fail | 9,237 | 2,017 | -78% | 1 | 1 | 0% | 1,478 | 1,007 | -32% | 0 | 0 | — |
case-16 | pass→pass | 6,899 | 2,602 | -62% | 1 | 1 | 0% | 979 | 1,066 | +9% | 0 | 0 | — |
case-17 | fail→fail | 6,748 | 4,761 | -29% | 1 | 1 | 0% | 1,055 | 1,451 | +38% | 0 | 0 | — |
case-18 | pass→pass | 11,622 | 3,364 | -71% | 1 | 1 | 0% | 1,794 | 1,243 | -31% | 0 | 0 | — |
case-19 | fail→pass | 4,959 | 2,254 | -55% | 1 | 1 | 0% | 733 | 1,055 | +44% | 0 | 0 | — |
case-20 | pass→pass | 7,159 | 6,031 | -16% | 1 | 1 | 0% | 1,226 | 1,597 | +30% | 0 | 0 | — |
case-21 | fail→fail | 26,615 | 6,675 | -75% | 1 | 1 | 0% | 3,882 | 1,840 | -53% | 0 | 0 | — |
case-22 | fail→fail | 13,556 | 13,868 | +2% | 1 | 1 | 0% | 2,348 | 2,416 | +3% | 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 +23 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.