frame-data-chart-nyt
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data from the user (in the form of CSV, JSON, or text conclusions) and interpolate this content directly into a generated HTML/SVG document.
- Ingestion points: User-provided data inputs defined in the SKILL.md processing instructions.
- Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded commands within the ingested data.
- Capability inventory: The skill is capable of generating and executing HTML and SVG content within the agent's display context.
- Sanitization: The instructions lack specific requirements for escaping or sanitizing user-provided data before rendering, which could allow maliciously crafted data to influence the agent's output behavior.
Audit Metadata