When To Use

Use this skill when the user wants image generation, image editing, or reference-image-based image creation through GoAI, especially for prompts like 生成图片, 画图, 出图, 做图, 修图, and 改图.

Pass the user's image prompt through as-is unless the user explicitly asks for prompt optimization, prompt polishing, rewriting, translation, or additional creative detail. Do not silently embellish short prompts such as 画一个老虎.

Do not translate the user's prompt to another language. If the user writes in Chinese, pass Chinese to --prompt. If the user writes in English, pass English to --prompt. You may strip only lightweight request wrappers such as 帮我, 给我, 请, 生成, or 画一张, but do not rewrite the core prompt content.

Examples:

• - User: 给我生成一只狮子 -> --prompt "一只狮子" or INLINECODE10
• User: 画一个老虎 -> --prompt "画一个老虎" or INLINECODE13
• Never turn 给我生成一只狮子 into a lion unless the user explicitly asks for English translation.

Default Behavior

This package currently targets the production environment at https://mustgoai.com. Override GOAI_BASE_URL only when you intentionally need a different endpoint.

This skill now uses a single cross-platform Python entrypoint through uv. If the user does not specify a model, the script queries /api/v1/models/capabilities, sorts enabled image models by display.priority, and picks the highest-priority model.

Treat uv as the only runtime dependency the user needs to install manually. On first run, uv may create a local environment, install httpx, and download Python if the machine does not already have a usable interpreter. That first-run setup is expected and should not be described as an error.

If the user provides local image paths, the script uploads them first and sends the resulting URLs as reference_images. If the user provides remote URLs, the script passes them through unchanged.

Script Rules

Always use the Python entrypoint through uv:

CODEBLOCK0

The Python path validates GOAI_API_KEY, defaults to https://mustgoai.com unless GOAI_BASE_URL is set, and treats 401, 402, 429, content blocking, missing task_id, missing image_url, and backend terminal failure states as hard failures. Polling follows the web client behavior: it checks every 3 seconds, retries transient polling errors, and waits for a backend terminal status instead of enforcing a client-side timeout.

The Python entrypoint self-heals into uv run if it is accidentally invoked as python scripts/generate_image.py ... or python3 scripts/generate_image.py .... Do not try to repair the system Python environment for this skill.

When invoking this skill through OpenClaw's exec tool, always use timeout=600 to allow up to 10 minutes for long-running image generation. Do not describe the run as timed out or killed unless the script itself exits non-zero or the backend reports a terminal failure state.

If uv is missing, explicitly guide the user to install it first:

• - macOS: INLINECODE40
• Windows: INLINECODE41

After uv is installed, rerun the same command. Do not ask the user to install Python packages by hand; uv is responsible for preparing Python and the skill dependencies.

If execution reports a missing Python module such as httpx, do not run pip install, pip install --user, or pip install --break-system-packages. Re-run the skill through uv; the dependency set comes only from pyproject.toml and uv.lock.

If GOAI_API_KEY is missing, explicitly guide the user to visit https://mustgoai.com, register or log in, open Settings -> API Key, create a key, and then configure the skill env in ~/.openclaw/openclaw.json.

Output Contract

On success, print all four lines below in this order:

• - INLINECODE55
• INLINECODE56
• INLINECODE57
• INLINECODE58

When responding to the user after a successful run, always include both the exact local file path and the exact public URL in plain text, even if OpenClaw already rendered or read the local media file. Do not omit the URL just because the image preview succeeded. On failure, exit non-zero and print a concise error message. Do not inline the binary output back into the conversation.