ag2-network-discussion

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via the runtime EV_TEXT(reply.body) that is produced by other participants (e.g., human or other agents) and then fanned out through the hub/view into each participant’s Agent.ask prompt; this includes any non-operating-user-authored text contributed during the discussion.