tao-train-deformable-detr
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE]: The skill includes configuration templates with placeholder values for S3 buckets and encryption keys. These are standard documentation practices for the TAO Toolkit and do not expose sensitive credentials.
- [EXTERNAL_DOWNLOADS]: The skill configuration references official NVIDIA Docker images from the NVIDIA GPU Cloud (nvcr.io). These are verified sources for the TAO Toolkit environment.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied dataset paths and configuration parameters that are eventually passed to a command-line interface. While this represents a data ingestion surface, the risk is minimized by the specialized nature of the target CLI and the containerized execution environment.
- [COMMAND_EXECUTION]: The skill defines specific commands for the
deformable_detrtool. These commands are restricted to the model training and deployment lifecycle and are executed within a controlled Docker environment.
Audit Metadata