tao-validate-dataset-format
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified. The skill's operations are confined to dataset validation using vendor-provided tools.
- [COMMAND_EXECUTION]: The skill uses the
tao-daftcommand-line utility to perform structural and schema validation on datasets. This is the primary function of the skill and utilizes an official tool from the author (NVIDIA). - [EXTERNAL_DOWNLOADS]: The instructions describe installing the
nvidia-tao-daftPython package from a standard registry. This package is a vendor-owned resource required for the skill's operation. - [DATA_EXPOSURE]: The skill processes dataset files at a user-specified path. This ingestion is necessary for the validation task and does not involve unauthorized data access or exfiltration.
Audit Metadata