docx

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly uses the third-party "rebyteai/internet-search" (see Sub-Skills) and Step 4 "Research (if needed)" to fetch public web content (industry benchmarks, template best practices, current rates/regulations) which the agent is instructed to read and use to populate documents, so untrusted external content can materially influence generation and next actions.