--- name: openclaw-image-skill description: "Generate images via OpenAI DALL-E 3 API. Use when a user asks to create, generate, or visualize an image from a text prompt." --- # OpenClaw Image Skill Generate images from text prompts using OpenAI's DALL-E 3 model. ## Requirements - `OPENAI_API_KEY` environment variable set, **or** pass the key inline - `curl` available in PATH ## Usage ```bash # Using env var (recommended) export OPENAI_API_KEY=sk-proj-... ./scripts/generate.sh "a glowing robot holding a paintbrush" output.png # Inline key OPENAI_API_KEY=sk-proj-... ./scripts/generate.sh "a cat on a laptop" cat.png ``` The script downloads the generated image to the specified output path. ## How It Works 1. Sends a POST to `https://api.openai.com/v1/images/generations` with model `dall-e-3` 2. Parses the returned image URL from the JSON response 3. Downloads the image to disk (URLs expire after ~2 hours) 4. Prints the output path on success ## Parameters | Param | Default | Notes | |---|---|---| | Model | `dall-e-3` | Hardcoded; swap to `dall-e-2` for cheaper/smaller | | Size | `1024x1024` | DALL-E 3 supports `1024x1024`, `1792x1024`, `1024x1792` | | Quality | `standard` | Use `hd` for higher detail (costs more) | ## Sample Images - `images/skill-concept.png` — AI-generated visualization of "what a skill is": a glowing modular puzzle piece slotting into a robot brain - `images/cat.png` — AI-generated orange tabby cat on a desk next to a laptop ## Agent Instructions When a user asks to generate an image: 1. Confirm you have `OPENAI_API_KEY` available 2. Craft a detailed, descriptive prompt from the user's request 3. Run `./scripts/generate.sh "<prompt>" <output_path>` 4. Report the saved file path and share the image if the environment supports it 5. Remind the user the URL expires after 2 hours — the downloaded file is permanent ## Security Note Never log, echo, or commit the API key. Always use env vars or a secrets manager.