modelopt-compare-results

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

• Third-party content exposure detected (low risk: 0.10). The required workflow's External Baseline Sanity Check has the agent look up external reference scores from web sources, which are outsider-authored pages that the agent must actively search for or fetch, so exposure is low; the NEL/MLflow run configs, scores, and logs are own-tenant artifacts and do not count.