follow-up-chaser

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

• Third-party content exposure detected (high risk: 0.85). This skill’s required workflow reads the user’s email threads from the Email connector (Gmail/Microsoft 365) and then has the agent ingest that thread text/subject/body to determine what was owed; that email content is outsider-authored free text (other people’s messages) and is fed into the LLM context for drafting decisions.