pufferlib
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill provides legitimate tools and documentation for reinforcement learning research.
- [COMMAND_EXECUTION]: Provides command-line interfaces for training (
puffer train) and distributed execution (torchrun), which are standard practices for this domain. - [EXTERNAL_DOWNLOADS]: References the installation of "pufferlib" via standard package managers and integration with well-known logging services like Weights & Biases and Neptune.
Audit Metadata