muapi-product-ad-cinematic

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 the agent to reference/condition on a user-supplied product_image URL (Phase A: "Reference the user's product_image if the model supports image conditioning"), meaning the agent fetches and interprets arbitrary third‑party image content that can materially influence prompts and downstream generation steps.