evaluate-ad

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

• Third-party content exposure detected (medium risk: 0.65). Outsider free text can enter the LLM context via the “Current evidence” inputs to the Diagnosis Agent (e.g., “comments, sentiment, click-quality, support tickets, customer DMs referencing the ad, screenshot of comments thread”), which are not authored by the operating user and are read into the agent prompt at runtime.