NYC

pytorch-lightning

Pass

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: LOWSAFE
Full Analysis
  • SAFE (SAFE): No malicious patterns, obfuscation, persistence mechanisms, or unauthorized access attempts were detected in the skill files. The code consists of established boilerplate for PyTorch Lightning model development and data management.\n- Indirect Prompt Injection (LOW): The skill defines a data ingestion surface common to machine learning workflows. 1. Ingestion points: scripts/template_datamodule.py (CustomDataset class) loads data via path-based arguments. 2. Boundary markers: Absent (typical for data-loading templates). 3. Capability inventory: The skill is limited to local tensor operations and logging via standard PyTorch/Lightning libraries; no file-writing, network access, or command execution is present across any script. 4. Sanitization: Absent (not applicable for numeric tensor data). Severity is low as the skill's capabilities are restricted to internal model training without external side effects.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 15, 2026, 11:12 PM