weights-biases
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is purely informational, providing expert guidance, documentation, and best practices for the Weights & Biases (W&B) library.
- [SAFE]: Code examples demonstrate standard and expected API usage for experiment tracking (
wandb.init,wandb.log), artifact versioning (wandb.Artifact,run.log_artifact), and hyperparameter sweeps (wandb.sweep). - [SAFE]: The skill explicitly includes security best practices, such as advising the user to configure the
WANDB_API_KEYvia environment variables rather than hardcoding it, and warning against logging sensitive credentials or PII to dashboards. - [SAFE]: All external references target the official Weights & Biases domain (wandb.ai), which is a well-known service in the machine learning ecosystem.
Audit Metadata