form-exam

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

• Third-party content exposure detected (medium risk: 0.65). The required workflow ingests the user-provided “Surface” (URL/screenshot/description/code) and then performs the theory audit using that content; if the surface is a URL or downloaded/forwarded document, the skill may fetch/parse outsider-authored free text at runtime and include it in the LLM context (public web content / downloaded attachments).