tao-train-centerpose

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface where instructions could be embedded in data processed at runtime.
  • Ingestion points: The skill ingests external datasets through parameters such as dataset.train_data, dataset.val_data, dataset.test_data, and dataset.inference_data as described in SKILL.md and references/skill_info.yaml.
  • Boundary markers: No specific delimiters or instructions to disregard embedded commands in the datasets are specified in the provided instruction files.
  • Capability inventory: The skill utilizes command execution within Docker containers (e.g., centerpose train, centerpose evaluate) to process these datasets, as defined in references/skill_info.yaml and references/tao-deploy-centerpose.md.
  • Sanitization: There are no explicit sanitization or validation steps mentioned for the content of the external datasets before they are processed by the tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 02:56 AM
Security Audit — agent-trust-hub — tao-train-centerpose