restaurant-ads

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 includes WebSearch(...) calls to fetch “[name] [city] menu specials/events/catering” and “existing review quotes,” which are public web content authored by outsiders and would be ingested as readable text into the agent’s LLM context for ad generation.