torch-geometric

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references binary wheels for specialized operations (pyg_lib, torch_scatter, etc.) from data.pyg.org, which is the official distribution host for the PyTorch Geometric ecosystem.
  • [COMMAND_EXECUTION]: Instructions are provided for setting up the environment using the uv package manager to install required libraries.
  • [DYNAMIC_EXECUTION]: The script scripts/create_gnn_template.py generates boilerplate Python code for various GNN architectures and saves it to a user-specified file path. The generated code is sourced from static strings defined within the script.
  • [INDIRECT_PROMPT_INJECTION]: The skill contains logic for ingesting external data from sources like CSV files and various benchmark datasets. It possesses capabilities to write files, such as code templates and visualization images, to the disk. Ingestion points: pd.read_csv and PyG dataset loaders. Boundary markers: None. Capability inventory: File writing in scripts/create_gnn_template.py and scripts/visualize_graph.py. Sanitization: None. This establishes a data-processing surface standard for machine learning workflows.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 12:15 PM