solo-community-outreach

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs the agent to search and scrape user-generated content from open third-party sites (Reddit, Hacker News, ProductHunt) via web_search/WebSearch/WebFetch in Steps 3a–3c and then uses that content to select threads and draft outreach, so untrusted third-party content can directly influence actions.