create-table-classifier

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The run.sh script provides instructions to install the uv tool from Astral's official domain (astral.sh), which is a well-known and trusted service for Python package management.\n- [EXTERNAL_DOWNLOADS]: Various training and utility scripts dynamically check for and install required libraries such as torch, transformers, timm, and pillow using pip from the official Python Package Index (PyPI).\n- [COMMAND_EXECUTION]: The skill makes extensive use of subprocess.run to orchestrate pipeline stages, including data collection from PDF corpuses, executing training scripts, and running evaluation benchmarks.\n- [REMOTE_CODE_EXECUTION]: Model loading in merge_inference.py and shadow_s00_inference.py uses pickle and joblib formats. While these formats can execute code during deserialization, they are standard in the Python ML ecosystem and are used here to load models from the skill's own directory structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 06:35 AM
Security Audit — agent-trust-hub — create-table-classifier