paw-cra-content-research

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs the agent to perform web searches and deep page extraction (web_search_exa, crawling_exa) and optionally use an agent-browser to pull social media feeds (Instagram, TikTok, LinkedIn) for competitor and trend analysis, then uses those public URLs as cited sources to drive production recommendations — exposing the agent to untrusted, user-generated third-party content that can influence actions.