workflow-permission-web

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

• Third-party content exposure detected (high risk: 0.75). Outsider-authored free text can enter the LLM context via Step 4’s WebSearch results (public web content) that are then used to assess/categorize unknown domains and presented in the TOON report; the workflow explicitly performs web research for domains not in trusted lists.