product-feedback-sim

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow samples a target user panel via sampler.sample_personas(..., source="nemotron_usa", mode="stream"), which (as a public/community persona source) can inject outsider-authored free-text persona attributes into the LLM context during panel_score/sgo evaluation.