torch-geometric
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and reference materials align with official library documentation and community best practices for Graph Neural Network development. Findings related to data ingestion and dynamic execution are consistent with the skill's primary purpose.
- [EXTERNAL_DOWNLOADS]: The skill references downloads of configuration, datasets, and pre-built binaries from official project domains such as pytorch.org and data.pyg.org, as well as the pyg-team GitHub organization.
- [INDIRECT_PROMPT_INJECTION]: The skill documents common patterns for loading external graph data from CSV files and remote URLs. It correctly advises users to verify data integrity and use trusted sources, addressing the attack surface associated with untrusted data ingestion.
- [DYNAMIC_EXECUTION]: The skill includes documentation for standard performance optimization features such as
torch.compile, which is a routine and legitimate part of the PyTorch ecosystem for accelerating model execution.
Audit Metadata