ai-image-generation

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.70). The skill's examples and required workflow include passing arbitrary public image URLs to the inference.sh apps (e.g., the "Image Upscaling" example with "image_url": "https://..." and "Stitch Multiple Images" with "images": ["https://img1.jpg", "https://img2.jpg"]), which the CLI will fetch and the models will read/interpret as part of processing, exposing the agent to untrusted third‑party content that could embed instructions or influences.