generative-ai-design

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs using arbitrary external image URLs as style references (e.g., Midjourney "--sref [image URL]" and the example "/imagine ... --sref https://[your-reference-image-url]"), which causes the agent/image model to ingest untrusted third-party content that can materially influence generation behavior.