AgentSkillsCN

grok-imagine

通过xAI的Grok Imagine API生成图像。当用户希望利用xAI/Grok生成AI图像，或在OpenAI图像生成服务暂时不可用时，便可使用此功能。

SKILL.md
--- frontmatter
name: grok-imagine
description: Generate images via xAI's Grok Imagine API. Use when the user wants to create AI-generated images using xAI/Grok, or when OpenAI image generation is unavailable.
homepage: https://docs.x.ai/docs/api-reference#image-generation
metadata:
  {
    "openclaw":
      {
        "emoji": "🎨",
        "requires": { "bins": ["node"], "env": ["XAI_API_KEY"] },
        "primaryEnv": "XAI_API_KEY",
      },
  }

Grok Imagine

Generate images via xAI's Grok Imagine API.

Run

bash
node {baseDir}/scripts/gen.mjs --prompt "your image description"

Examples

bash
# Basic image generation
node {baseDir}/scripts/gen.mjs --prompt "a cyberpunk city at sunset"

# Multiple images
node {baseDir}/scripts/gen.mjs --prompt "a friendly robot" --count 4

# Custom output directory
node {baseDir}/scripts/gen.mjs --prompt "mountain landscape" --out-dir ./images

# Image editing (provide input image)
node {baseDir}/scripts/gen.mjs --prompt "add a rainbow to the sky" --input /path/to/image.png

Models

  • •grok-imagine-image: Text-to-image and image editing (default)
  • •grok-2-image: Legacy image generation model

Parameters

  • •--prompt, -p: Image description (required)
  • •--count, -n: Number of images to generate (default: 1)
  • •--model, -m: Model to use (default: grok-imagine-image)
  • •--input, -i: Input image path for editing tasks (optional)
  • •--out-dir, -o: Output directory (default: ./tmp/grok-imagine-<timestamp>)

Output

  • •Generated images saved as PNG files
  • •prompts.json with prompt → file mapping
  • •index.html thumbnail gallery
  • •MEDIA: lines for OpenClaw auto-attach

API Key

Set XAI_API_KEY environment variable, or configure in OpenClaw:

  • •skills."grok-imagine".apiKey in ~/.openclaw/openclaw.json