team-composition-patterns

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

• Third-party content exposure detected (high risk: 1.00). This skill's Research Team preset explicitly gives agents WebSearch/WebFetch and instructs researchers to search the web and fetch page content (e.g., "WebSearch, WebFetch" and researcher tasks asking to "search for comparisons" and "fetch page content"), so the agent will read and interpret untrusted public web pages/URLs.