visual-style
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the extraction of design system specifications from untrusted external sources, creating a surface for indirect prompt injection.
- Ingestion points: Untrusted data enters the agent's context through the processing of website URLs (references/extractors/from-website.md), PDF brand guides (references/extractors/from-pdf.md), and video keyframes (references/extractors/from-video.md).
- Boundary markers: The instructions lack explicit boundary markers or directives for the agent to ignore embedded instructions within the raw source content, relying instead on a target schema.
- Capability inventory: The skill is capable of generating executable HTML/CSS presentation slides (references/connectors/html-slides.md) and constructing natural language prompts for external AI services like the HeyGen Video Agent (references/connectors/heygen-video-agent.md).
- Sanitization: There are no requirements or instructions for sanitizing text values extracted from external sources before they are interpolated into critical prompt fields or rendered into code.
Audit Metadata