hugging-face-trackio
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- Category 8: Indirect Prompt Injection (LOW): The skill handles untrusted external data in the form of ML experiment metrics.
- Ingestion points: Metric data is retrieved via the CLI (
trackio get metric) or the dashboard. - Boundary markers: The skill documentation explicitly recommends the
--jsonflag for programmatic use, which provides structured boundaries for data parsing. - Capability inventory: The skill is primarily focused on data visualization and retrieval; it lacks high-privilege capabilities such as arbitrary file modification or system administration.
- Sanitization: No explicit sanitization is described, but the risk is minimal given the display-only nature of the tool's interaction with the agent.
Audit Metadata