langgraph-agent-patterns

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

• Third-party content exposure detected (medium risk: 0.30). In references/supervisor-subagent.md, the runtime flow for the researcher subagent ingests outsider-authored free text from a search results payload via TavilySearchResults (search_tool.invoke(latest_query)) and then formats result['content'] into result_text that is passed back into the agents’ messages.