ad-creative

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime path is user-supplied prompts and (in iteration/strategy modes) user-provided performance data and uploaded “inputs” files (reviews/winning-ads/ad comments) that the LLM then reads to generate ad copy, so outsider-authored free text can directly enter the model without any prior selection step.