tao-train-foundation-stereo

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads pretrained model weights from Hugging Face's official repository (huggingface.co). Hugging Face is a recognized and established service for hosting machine learning artifacts.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied annotation files (annotations.txt) which are read into the agent's context to determine file paths for images and ground truth data. Maliciously crafted annotation files could attempt to inject instructions into the agent's logic.
  • Ingestion points: Data sources configured in SKILL.md and references/skill_info.yaml (specifically annotations.txt).
  • Boundary markers: Absent. The skill does not define specific delimiters or instructions to ignore embedded content within these data files.
  • Capability inventory: The skill uses the Bash tool to execute docker run commands which trigger training and evaluation processes.
  • Sanitization: Absent. The skill does not mention validation or escaping of the annotation file contents before they are processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 02:56 AM
Security Audit — agent-trust-hub — tao-train-foundation-stereo