Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Gemini画像生成APIでテキストから画像生成、既存画像の編集を行うスキル。 「画像を生成して」「イラストを作って」「写真を編集して」等のリクエストで発動。
.claude/skills/minicoohei-nanobanana/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 95% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -64% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -79% | 0% |
Generate or edit images using Gemini Image Generation API.
bash# Text-to-image python scripts/nanobanana.py "{prompt}" --aspect-ratio "{ratio}" # Image editing python scripts/nanobanana.py "{prompt}" --input "{image_path}" # Multiple reference images python scripts/nanobanana.py "{prompt}" --input "{image1}" --input "{image2}"
| Parameter | Required | Default | Description | |-----------|----------|---------|-------------| | prompt | Yes | - | Generation/edit instruction | | --input, -i | No | - | Input image(s) for editing | | --output, -o | No | auto | Output path | | --aspect-ratio, -ar | No | 16:9 | Output ratio | | --session, -s | No | - | Session name for organizing output | | --force-ar | No | false | Force aspect ratio in edit mode |
1:1, 4:3, 3:4, 16:9, 9:16, 21:9
bash# Generate landscape python scripts/nanobanana.py "Mt. Fuji at sunset, photorealistic" # Edit image python scripts/nanobanana.py "Remove the background" --input photo.jpg # Combine images python scripts/nanobanana.py "Merge these into one composition" --input img1.png --input img2.png # With session organization python scripts/nanobanana.py "Company logo design" --session "brand_assets" --aspect-ratio 1:1
Gemini Image Generation APIを使って、テキストプロンプトからの画像生成や、既存画像の編集を行う汎用画像生成スキルです。複数参照画像の合成にも対応。
| エラー | 解決方法 | |--------|---------| | API key not found | GEMINI_API_KEY または GOOGLE_API_KEY を環境変数に設定 | | Input image not found | --input に正しいファイルパスを指定。相対パスは実行ディレクトリ基準 |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-12 | fail→pass | 8,410 | 2,385 | -72% | 1 | 1 | 0% | 1,464 | 963 | -34% | 0 | 0 | — |
case-01 | fail→pass | 3,269 | 2,081 | -36% | 1 | 1 | 0% | 472 | 919 | +95% | 0 | 0 | — |
case-13 | pass→pass | 6,323 | 2,724 | -57% | 1 | 1 | 0% | 1,028 | 1,051 | +2% | 0 | 0 | — |
case-02 | fail→pass | 13,346 | 2,082 | -84% | 1 | 1 | 0% | 2,625 | 953 | -64% | 0 | 0 | — |
case-03 | fail→fail | 16,866 | 5,588 | -67% | 1 | 1 | 0% | 2,244 | 1,001 | -55% | 0 | 0 | — |
case-04 | fail→pass | 7,111 | 2,329 | -67% | 1 | 1 | 0% | 1,417 | 945 | -33% | 0 | 0 | — |
case-05 | fail→pass | 24,953 | 2,399 | -90% | 1 | 1 | 0% | 4,950 | 1,024 | -79% | 0 | 0 | — |
case-11 | pass→pass | 10,723 | 2,193 | -80% | 1 | 1 | 0% | 1,649 | 929 | -44% | 0 | 0 | — |
case-06 | fail→pass | 10,281 | 1,749 | -83% | 1 | 1 | 0% | 1,791 | 856 | -52% | 0 | 0 | — |
case-07 | pass→pass | 10,235 | 2,333 | -77% | 1 | 1 | 0% | 1,695 | 985 | -42% | 0 | 0 | — |
case-08 | pass→pass | 11,500 | 2,733 | -76% | 1 | 1 | 0% | 1,865 | 1,062 | -43% | 0 | 0 | — |
case-09 | pass→pass | 7,019 | 2,332 | -67% | 1 | 1 | 0% | 1,277 | 941 | -26% | 0 | 0 | — |
case-10 | pass→pass | 5,651 | 2,255 | -60% | 1 | 1 | 0% | 939 | 956 | +2% | 0 | 0 | — |
case-14 | fail→pass | 8,801 | 2,821 | -68% | 1 | 1 | 0% | 1,527 | 1,103 | -28% | 0 | 0 | — |
case-15 | fail→pass | 9,622 | 1,544 | -84% | 1 | 1 | 0% | 1,347 | 803 | -40% | 0 | 0 | — |
case-16 | fail→pass | 7,681 | 3,691 | -52% | 1 | 1 | 0% | 1,225 | 1,117 | -9% | 0 | 0 | — |
case-17 | pass→pass | 6,198 | 2,862 | -54% | 1 | 1 | 0% | 939 | 1,037 | +10% | 0 | 0 | — |
case-18 | pass→pass | 9,757 | 2,278 | -77% | 1 | 1 | 0% | 1,574 | 982 | -38% | 0 | 0 | — |
case-19 | pass→pass | 5,610 | 2,340 | -58% | 1 | 1 | 0% | 938 | 1,030 | +10% | 0 | 0 | — |
case-20 | pass→pass | 7,896 | 5,748 | -27% | 1 | 1 | 0% | 1,705 | 1,771 | +4% | 0 | 0 | — |
case-21 | pass→pass | 5,028 | 2,602 | -48% | 1 | 1 | 0% | 921 | 1,073 | +17% | 0 | 0 | — |
case-22 | pass→pass | 4,074 | 2,855 | -30% | 1 | 1 | 0% | 778 | 1,065 | +37% | 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 +41 percentage points is the difference between those two pass rates over the 21 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.