tao-finetune-nv-tesseract-ad-diffusion

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches the official source repository from NVIDIA's GitHub account and downloads model weights from the NVIDIA organization on Hugging Face.
  • [COMMAND_EXECUTION]: Utilizes 'git clone' for codebase retrieval and 'uv run python' to execute the provided examples and fine-tuning scripts.
  • [INDIRECT_PROMPT_INJECTION]: As the skill processes user-provided CSV data for time-series analysis, it possesses a surface for indirect prompt injection; however, it instructs the agent to drop non-numeric columns before processing, which acts as a data sanitization step.
  • [DYNAMIC_EXECUTION]: The skill provides a template for generating an inference trial script at runtime for AutoML and dynamically modifies the Python system path to load the model SDK from the local file system.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 02:56 AM
Security Audit — agent-trust-hub — tao-finetune-nv-tesseract-ad-diffusion