tao-train-optical-inspection

Pass

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface.\n
  • Ingestion points: Data enters the agent context through dataset CSV files and image folders referenced in configuration keys like dataset.train_dataset.csv_path as seen in SKILL.md and references/skill_info.yaml.\n
  • Boundary markers: No explicit delimiters or boundary markers are defined to isolate untrusted data from the command structure.\n
  • Capability inventory: The skill executes CLI actions such as optical_inspection train, evaluate, and inference via the Bash tool, which perform file operations and subprocess execution.\n
  • Sanitization: There is no evidence of sanitization or content validation for the metadata provided in the external dataset CSV files.\n- [EXTERNAL_DOWNLOADS]: The documentation includes commands that fetch Docker images from the official NVIDIA Container Registry (nvcr.io).
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 30, 2026, 05:48 PM
Security Audit — agent-trust-hub — tao-train-optical-inspection