Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate an image from a text prompt using an OpenAI-compatible image generation API (gpt-image-1-mini or compatible). The image is uploaded to the gofile.io public file sharing service and ONLY the public download page URL is returned. Trigger when user asks to draw, paint, generate, or create an image.
.claude/skills/ai-sns-draw-image-0dae1c/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -13% | 0% |
Generate images via an OpenAI-compatible images/generations endpoint. The script has a built-in API base and key, so no extra configuration is needed.
Generate an image from a text prompt:
bashpython3 ~/.hermes/skills/ai-sns/draw-image/scripts/draw.py --prompt "A futuristic cityscape at sunset with flying cars"
All parameters are optional except --prompt:
bashpython3 ~/.hermes/skills/ai-sns/draw-image/scripts/draw.py \ --prompt "A cute cat wearing a spacesuit" \ --size 1024x1024 \ --model gpt-image-1-mini\ --quality low \ --n 1
\n- --prompt: Text description of the image (required).
--size: Image dimensions (optional). Default is 1024x1024.--model: Image generation model (optional). Default is gpt-image-1-mini.--quality: Image quality, either low,medium,high (optional). Default is low.--n: Number of images to generate (optional). Default is 1.--out-dir (string, optional): Directory to save images. Default: <projectRoot>/uploads/generated_images/.--api-base (string, optional): Override the API base URL.--api-key (string, optional): Override the API key.Alternatively, pass parameters as JSON on stdin:
bashecho '{"prompt": "A serene mountain landscape at dawn"}' | python3 ~/.hermes/skills/ai-sns/draw-image/scripts/draw.py --stdin
DRAW_IMAGE_API_BASE: Override the API base URL.DRAW_IMAGE_API_KEY: Override the API key.Generate an image of a cow with the following command:
bashpython3 ~/.hermes/skills/ai-sns/draw-image/scripts/draw.py --prompt 'A cow' --size 1024x1024 --model gpt-image-1-mini --quality low
The image is generated, uploaded to gofile.io, and ONLY the public download page URL is returned. The script prints a JSON object to stdout:
ok (bool): Whether the generation and upload succeeded.download_page (string): The public gofile.io download page URL of the generated image. Only this URL is returned as the result. Open this URL in a browser to download the generated image.note (string): A short hint explaining that download_page is used to download the generated image.error (string): Error message on failure.Example success output:
json{ "ok": true, "download_page": "https://gofile.io/d/WcWZKt", "note": "Use this URL to download the generated image." }
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,899 | 6,025 | +2% | 1 | 1 | 0% | 855 | 1,158 | +35% | 0 | 0 | — |
case-02 | fail→fail | 8,983 | 5,452 | -39% | 1 | 1 | 0% | 1,287 | 912 | -29% | 0 | 0 | — |
case-03 | fail→fail | 4,307 | 5,032 | +17% | 1 | 1 | 0% | 293 | 958 | +227% | 0 | 0 | — |
case-04 | fail→pass | 6,365 | 1,583 | -75% | 1 | 1 | 0% | 1,086 | 922 | -15% | 0 | 0 | — |
case-05 | fail→pass | 7,811 | 1,659 | -79% | 1 | 1 | 0% | 1,241 | 897 | -28% | 0 | 0 | — |
case-06 | pass→pass | 5,931 | 1,396 | -76% | 1 | 1 | 0% | 934 | 861 | -8% | 0 | 0 | — |
case-07 | pass→pass | 6,393 | 1,475 | -77% | 1 | 1 | 0% | 831 | 859 | +3% | 0 | 0 | — |
case-08 | fail→pass | 15,963 | 1,710 | -89% | 1 | 1 | 0% | 2,573 | 956 | -63% | 0 | 0 | — |
case-09 | fail→pass | 7,540 | 1,826 | -76% | 1 | 1 | 0% | 1,263 | 966 | -24% | 0 | 0 | — |
case-10 | fail→pass | 6,471 | 1,564 | -76% | 1 | 1 | 0% | 1,038 | 899 | -13% | 0 | 0 | — |
case-11 | fail→pass | 9,363 | 1,546 | -83% | 1 | 1 | 0% | 1,517 | 913 | -40% | 0 | 0 | — |
case-12 | pass→pass | 12,767 | 1,796 | -86% | 1 | 1 | 0% | 1,943 | 970 | -50% | 0 | 0 | — |
case-13 | fail→pass | 6,947 | 2,092 | -70% | 1 | 1 | 0% | 1,214 | 1,058 | -13% | 0 | 0 | — |
case-14 | fail→pass | 6,349 | 4,486 | -29% | 1 | 1 | 0% | 1,199 | 1,461 | +22% | 0 | 0 | — |
case-15 | fail→pass | 12,374 | 3,148 | -75% | 1 | 1 | 0% | 2,097 | 1,236 | -41% | 0 | 0 | — |
case-16 | fail→pass | 8,414 | 1,615 | -81% | 1 | 1 | 0% | 1,290 | 865 | -33% | 0 | 0 | — |
case-17 | fail→pass | 8,988 | 2,344 | -74% | 1 | 1 | 0% | 1,329 | 1,129 | -15% | 0 | 0 | — |
case-18 | fail→pass | 8,971 | 4,392 | -51% | 1 | 1 | 0% | 1,436 | 1,426 | -1% | 0 | 0 | — |
case-23 | pass→fail | 8,408 | 4,690 | -44% | 1 | 1 | 0% | 1,459 | 890 | -39% | 0 | 0 | — |
case-19 | pass→pass | 5,950 | 2,446 | -59% | 1 | 1 | 0% | 876 | 1,111 | +27% | 0 | 0 | — |
case-20 | pass→pass | 9,724 | 1,959 | -80% | 1 | 1 | 0% | 1,760 | 999 | -43% | 0 | 0 | — |
case-21 | pass→pass | 8,501 | 7,729 | -9% | 1 | 1 | 0% | 1,536 | 2,132 | +39% | 0 | 0 | — |
case-22 | fail→fail | 3,428 | 8,811 | +157% | 1 | 1 | 0% | 514 | 1,529 | +197% | 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 18 counted toward the lift figure. The other 5 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 +48 percentage points is the difference between those two pass rates over the 18 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.