Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate and edit images using Gemini Flash Image, and generate videos using Veo. Supports text-to-image, image editing, text-to-video, and image-to-video.
.claude/skills/sonichi-image-generation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -83% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -9% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -43% | 0% |
Generate images and videos using Gemini APIs.
bash# Text-to-image python3 "$SKILL_DIR/scripts/generate.py" --prompt "A futuristic city skyline at night" # Edit an existing image python3 "$SKILL_DIR/scripts/generate.py" --input photo.jpg --prompt "Add dramatic clouds" # Specify output path python3 "$SKILL_DIR/scripts/generate.py" --prompt "A cute robot mascot" --output mascot.png # Text-to-video python3 "$SKILL_DIR/scripts/generate.py" --video --prompt "A timelapse of a city at sunset" # Video with portrait aspect ratio python3 "$SKILL_DIR/scripts/generate.py" --video --prompt "Ocean waves" --aspect 9:16 # Image-to-video (animate a reference image) python3 "$SKILL_DIR/scripts/generate.py" --video --input scene.jpg --prompt "Animate this scene with gentle wind" # Specify output python3 "$SKILL_DIR/scripts/generate.py" --video --prompt "Dancing robot" --output robot.mp4
| Flag | Description | Default | |------|-------------|---------| | --prompt | Text prompt describing what to generate | (required) | | --input | Input image path(s) for editing/reference | None | | --output | Output file path | generated-{timestamp}.png or .mp4 | | --model | Gemini model to use | gemini-2.5-flash-image / veo-3.1-generate-preview | | --video | Generate video instead of image | false | | --aspect | Video aspect ratio | 16:9 | | --quality | JPEG quality (1-100, images only) | 90 |
google-genai Python package (pip3 install google-genai)GEMINI_API_KEY in .env or environmentpip3 install Pillow) — for image generation/editing| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 6,617 | 10,554 | +59% | 1 | 1 | 0% | 1,121 | 1,516 | +35% | 0 | 0 | — |
case-02 | fail→pass | 34,733 | 1,812 | -95% | 1 | 1 | 0% | 6,178 | 1,035 | -83% | 0 | 0 | — |
case-03 | fail→fail | 6,266 | 6,298 | +1% | 1 | 1 | 0% | 908 | 1,107 | +22% | 0 | 0 | — |
case-04 | fail→pass | 8,792 | 2,062 | -77% | 1 | 1 | 0% | 1,570 | 1,060 | -32% | 0 | 0 | — |
case-05 | fail→pass | 7,261 | 2,146 | -70% | 1 | 1 | 0% | 1,175 | 1,075 | -9% | 0 | 0 | — |
case-06 | fail→pass | 11,572 | 1,585 | -86% | 1 | 1 | 0% | 1,978 | 1,018 | -49% | 0 | 0 | — |
case-07 | fail→pass | 11,124 | 1,910 | -83% | 1 | 1 | 0% | 1,912 | 1,098 | -43% | 0 | 0 | — |
case-08 | pass→pass | 4,011 | 1,174 | -71% | 1 | 1 | 0% | 683 | 868 | +27% | 0 | 0 | — |
case-09 | pass→pass | 9,516 | 1,291 | -86% | 1 | 1 | 0% | 1,498 | 902 | -40% | 0 | 0 | — |
case-10 | fail→pass | 12,682 | 1,649 | -87% | 1 | 1 | 0% | 2,129 | 1,002 | -53% | 0 | 0 | — |
case-11 | pass→pass | 9,872 | 3,935 | -60% | 1 | 1 | 0% | 1,724 | 1,364 | -21% | 0 | 0 | — |
case-12 | fail→pass | 4,316 | 977 | -77% | 1 | 1 | 0% | 674 | 860 | +28% | 0 | 0 | — |
case-13 | fail→pass | 11,193 | 2,199 | -80% | 1 | 1 | 0% | 1,692 | 1,057 | -38% | 0 | 0 | — |
case-14 | pass→pass | 3,527 | 1,230 | -65% | 1 | 1 | 0% | 526 | 876 | +67% | 0 | 0 | — |
case-15 | fail→pass | 8,955 | 1,466 | -84% | 1 | 1 | 0% | 1,466 | 914 | -38% | 0 | 0 | — |
case-16 | pass→pass | 14,728 | 7,016 | -52% | 1 | 1 | 0% | 2,299 | 1,763 | -23% | 0 | 0 | — |
case-17 | fail→pass | 9,644 | 2,120 | -78% | 1 | 1 | 0% | 1,621 | 1,096 | -32% | 0 | 0 | — |
case-18 | fail→pass | 9,681 | 1,546 | -84% | 1 | 1 | 0% | 1,546 | 971 | -37% | 0 | 0 | — |
case-19 | fail→pass | 8,979 | 2,278 | -75% | 1 | 1 | 0% | 1,331 | 1,119 | -16% | 0 | 0 | — |
case-20 | fail→pass | 9,961 | 2,494 | -75% | 1 | 1 | 0% | 1,545 | 1,150 | -26% | 0 | 0 | — |
case-21 | fail→pass | 13,104 | 10,247 | -22% | 1 | 1 | 0% | 2,331 | 2,560 | +10% | 0 | 0 | — |
case-22 | pass→pass | 8,629 | 3,606 | -58% | 1 | 1 | 0% | 1,284 | 1,364 | +6% | 0 | 0 | — |
case-23 | fail→pass | 10,689 | 8,396 | -21% | 1 | 1 | 0% | 1,725 | 2,237 | +30% | 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. 23 cases were attempted, and 21 counted toward the lift figure. The other 2 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 +65 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.