sklearn-advanced
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [REMOTE_CODE_EXECUTION] (SAFE): No remote code execution or untrusted script downloads were detected. The skill correctly suggests established libraries like
skl2onnxfor specific deployment tasks. - [DATA_EXFILTRATION] (SAFE): Analysis of file and network operations shows no evidence of exfiltration. Model persistence using
joblibis demonstrated for local storage purposes, which is a standard workflow in data science. - [COMMAND_EXECUTION] (SAFE): The skill utilizes Python APIs for machine learning. There are no calls to system shells, subprocesses, or arbitrary command execution primitives.
- [DYNAMIC_EXECUTION] (LOW): The skill demonstrates the use of
joblib.load()for model deserialization. While loading untrusted pickle-based files poses a theoretical risk of code execution, it is the standard tool for this use case and is presented here in a safe, educational context for loading locally saved models.
Audit Metadata