tao-train-mask-grounding-dino

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references official container images from the NVIDIA Container Registry (nvcr.io) for both training and deployment tasks. These are legitimate resources provided by the vendor.
  • [COMMAND_EXECUTION]: The skill instructs the agent to run the mask_grounding_dino suite of commands within Docker containers to perform model-related actions like training and quantization. This is necessary for its stated purpose.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill processes external data from user-supplied dataset files (JSON and JSONL) as defined in references/skill_info.yaml.
  • Boundary markers: None are specified in the configuration templates to separate data from instructions.
  • Capability inventory: The skill has the capability to execute shell commands via the TAO CLI.
  • Sanitization: Static analysis does not show explicit sanitization or validation of the contents of the dataset files.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 02:56 AM
Security Audit — agent-trust-hub — tao-train-mask-grounding-dino