himalaya

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

• Third-party content exposure detected (high risk: 0.85). Outsider-authored email content (e.g., message bodies/headers from the user’s IMAP account) is fetched at runtime and then ingested into the agent/LLM context when the workflow runs commands like himalaya message read / message export / attachment download, which return readable prose derived from other senders.