tao-mine-od-images
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill provides instructions and a validation script for running a containerized NVIDIA TAO Data Services command.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill processes local data files (parquet) and writes results to a local output directory. No evidence of data exfiltration to unauthorized remote servers was found. It references the NVIDIA Container Registry (nvcr.io), which is a well-known and trusted service.
- [INDIRECT_PROMPT_INJECTION]: The skill's input surface is limited to structured YAML configuration files and parquet data files. The validation script uses
yaml.safe_load(), which is a secure method for parsing YAML. The processed data is passed to a containerized environment, minimizing the risk of influencing the agent's own instruction context.
Audit Metadata