markstream-nuxt
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the rendering of markdown data which is often sourced from untrusted external entities or model outputs.
- Ingestion points: The markdown prop within the MarkdownRender component in SKILL.md serves as the primary entry point for untrusted content.
- Boundary markers: The documentation recommends the use of html-policy="safe" and strict settings for diagram rendering (Mermaid) to limit the execution context of processed data.
- Capability inventory: No subprocess calls, network operations, or file system writes are identified in the skill's instructions or code examples.
- Sanitization: The instructions explicitly warn against using 'loose' settings or 'trusted' HTML with untrusted content, advocating for built-in library sanitization features.
Audit Metadata