mlflow-data-pipelines

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt includes examples that embed database credentials directly in command-line URIs (e.g., postgresql+psycopg2://mlflow:secret@...) and shows environment variable placeholders for secrets used in startup commands, which encourages or requires echoing secret values verbatim in commands/compose files (high exfiltration risk).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The example "mlflow run https://github.com/myorg/churn-model.git" causes mlflow to fetch (git clone) and execute code from a remote repository at runtime, so this URL is a runtime external dependency that can execute remote code.