regression
Pass
Audited by Gen Agent Trust Hub on Apr 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and scripts implement a valid machine learning workflow for tabular regression, focusing on model accuracy and uncertainty estimation through conformalized quantile regression.
- [EXTERNAL_DOWNLOADS]: The skill references and uses standard, industry-standard Python libraries for data science, including xgboost, scikit-learn, shap, and pandas. These packages are reputable and hosted on the official Python Package Index (PyPI).
- [DATA_EXFILTRATION]: Analysis of the code and instructions reveals no network activity or access to sensitive local files (such as credentials or SSH keys). Data processing is limited to provided local datasets and synthetic data.
- [PROMPT_INJECTION]: The skill contains no instructions that attempt to manipulate the AI agent's behavior, bypass safety guardrails, or extract system prompts.
- [COMMAND_EXECUTION]: All code snippets use standard API calls for machine learning and data manipulation. There is no evidence of arbitrary shell command execution or risky system-level operations.
Audit Metadata