google-chat

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

• Third-party content exposure detected (high risk: 0.85). Outsider-authored free text from Google Chat messages/attachments is ingested into the agent’s LLM context via the tool’s runtime read/search actions (e.g., list_messages, get_message, search_messages, list_message_attachments, search_attachments) which return message text/fields originating from other users in the space.