analyzing-mlflow-session

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

• Third-party content exposure detected (medium risk: 0.30). The skill describes an MLflow CLI workflow that searches and fetches traces belonging to a given mlflow.trace.session, and those traces can include outsider-authored chat inputs/outputs via extracted fields like mlflow.traceInputs/mlflow.traceOutputs and via mlflow traces get full JSON for specific turns.