analyze-account-signals

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

• Third-party content exposure detected (high risk: 0.75). The required workflow ingests outsider-authored free text via the external_messaging, meeting_notes, internal_messaging, and document_store source categories (e.g., customer emails/threads, meeting notes, and documents) into the agent’s LLM context during evidence collection and signal normalization.