tool_config.json

{
  "description": "This is a tool that creates an image according to a prompt, which is a text description.",
  "inputs": "{'prompt': {'type': 'text', 'description': \"The image generator prompt. Don't hesitate to add details in the prompt to make the image look better, like 'high-res, photorealistic', etc.\"}}",
  "name": "image_generator",
  "output_type": "image",
  "tool_class": "tool.TextToImageTool"
}