instrumenting-with-mlflow-tracing
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXFILTRATION] (SAFE): The skill provides patterns for observability while emphasizing data privacy. It includes a specific guide for PII redaction using SpanProcessors to prevent sensitive information (emails, phone numbers, credit cards, etc.) from being sent to the tracking server.
- [EXTERNAL_DOWNLOADS] (SAFE): Installation instructions point to official packages (mlflow, mlflow-tracing) via standard package managers (pip, npm). No untrusted third-party scripts or URLs are used.
- [CREDENTIALS_UNSAFE] (SAFE): The skill correctly demonstrates using environment variables for authentication (MLFLOW_TRACKING_TOKEN, MLFLOW_TRACKING_PASSWORD), which is a secure practice compared to hardcoding secrets.
- [PROMPT_INJECTION] (SAFE): No malicious instructions or bypass attempts were detected. The use of an LLM for evaluation is handled safely by constraining the expected output to a numeric score via integer casting.
Audit Metadata