pytorch-lightning
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides instructional content and code templates for the PyTorch Lightning library. Analysis of the documentation and code examples found no malicious patterns, obfuscation, or unauthorized data access.- [EXTERNAL_DOWNLOADS]: The skill references established machine learning logging services (Weights & Biases, MLflow, Neptune, Comet) and the author's official platform (k-dense.ai). These are standard industry tools or vendor-owned resources used for experiment tracking and research workflows.- [SAFE]: Potential surface for indirect prompt injection was evaluated. Ingestion points: User datasets processed via LightningDataModule classes as described in SKILL.md. Boundary markers: None mentioned. Capability inventory: Use of subprocesses for distributed training strategies (DDP, FSDP, DeepSpeed) and network operations for metric logging. Sanitization: Not applicable for this documentation-only skill. The instructions follow standard framework usage patterns.
Audit Metadata