agentmail-cli

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly includes commands to list and retrieve inbox messages (e.g., "agentmail inboxes:messages list" and "agentmail inboxes:messages retrieve") and to create webhooks for incoming messages, so the agent will ingest untrusted, user-generated email content from third parties that could contain instructions influencing replies/forwards.