groundwork-session

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

• Third-party content exposure detected (medium risk: 0.30). Outsider-authored free text is ingested directly from the user in the required runtime workflow (“Read context…”, then “Ask the opening question… Every assistant turn… MUST end with a question” and “Then the full transcript…as-is”), i.e., the conversation content provided by the user is LLM-readable and logged.