restaurant-online

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow uses WebSearch(...) to fetch and audit public web content (e.g., Google Business Profile, Yelp, delivery/reservation pages), which can include outsider-authored free text (reviews, Q&A, descriptions) that is then ingested into the agent’s LLM context for scoring and summarization.