get-to-know-a-client

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow ingests “client notes, meeting transcripts, documents, previous work” (Phase 1: “Read and analyze all available client information”), which are typically outsider-authored free text provided by the client/other parties, and that text is then analyzed and used in the agent’s LLM context.