meta-campaign

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

• Third-party content exposure detected (high risk: 0.90). High—Stage 00/04/01 require runtime web browsing/scraping of outsider-authored content (Meta Ad Library pages and WebSearch/WebFetch of competitor reviews/forums like Reddit), and that fetched prose/quotes are then synthesized and fed back into the agent’s LLM context for downstream strategy/copy.