notebooks

Pass

Audited by Gen Agent Trust Hub on May 28, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses command-line interfaces such as uv and marimo for environment management, notebook linting, and execution. These are standard operations for a data science and notebook development workflow.
  • [REMOTE_CODE_EXECUTION]: The skill is designed to execute Python code contained within notebooks using the marimo runtime and the provided scripts/execute_notebook.py utility. This capability is the primary intended function of the skill. The instructions advocate for security best practices, such as running notebooks in isolated sandboxes (--sandbox) and using fresh kernels to ensure reproducibility and limit environmental side effects.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of standard data science packages (e.g., polars, duckdb, matplotlib) and browser dependencies for PDF generation (Chromium) via well-known package managers and registries. These downloads are standard for the described use case.
Audit Metadata
Risk Level
SAFE
Analyzed
May 28, 2026, 08:10 PM
Security Audit — agent-trust-hub — notebooks