evaluate

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow reads outsider-authored free text from evals/files/predictions.csv and evals/files/regression_predictions.csv (CSV contents are user-provided evaluation data, not bundled reference material), which the agent then ingests into the LLM context to compute metrics and generate the evaluation report.