engineering-email-intelligence-engineer

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow ingests and LLM-processes raw email bodies by fetching RFC822 messages (including user-authored content) and then assembling query-relevant context via hybrid retrieval, so outsiders can submit poison by sending emails to the monitored/queried mailbox/source.