skills/nvidia/skills/cuequivariance/Gen Agent Trust Hub

cuequivariance

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill pins its environment to a specific revision of the official NVIDIA cuEquivariance library hosted on GitHub in scripts/requirements.txt. This is a documented, well-known service from a trusted organization for development purposes.
  • [COMMAND_EXECUTION]: The skill provides Python code snippets for implementing group theory mathematics and performing NumPy-based numerical checks. These operations are standard for scientific computing and do not involve dangerous system calls or shell injection.
  • [DATA_EXPOSURE]: Analysis of the instructions and examples shows no access to sensitive file paths, environment variables, or hardcoded credentials. All data processing is local and mathematical in nature.
  • [REMOTE_CODE_EXECUTION]: No patterns of remote script execution (e.g., piping curl to bash) or dynamic code execution from untrusted sources were found. The use of a requirements file to pin a specific git commit hash is a security best practice for reproducible research environments.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 08:54 AM