interactive-flow-diagram
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill generates code that fetches the D3.js library from the jsDelivr CDN and imports the Inter font family from Google Fonts. Both are well-known, established services for web utility and design.
- [INDIRECT_PROMPT_INJECTION]: The skill accepts arbitrary user text or Mermaid diagrams as input to populate the
flowDataobject in the generated HTML. This creates a surface for indirect prompt injection where malicious instructions or script tags could be embedded in the user input and subsequently rendered in the final visualization. - Ingestion points: User-provided descriptions, lists, or Mermaid code processed in
SKILL.mdto build the diagram. - Boundary markers: Content is delimited within the generated JavaScript
flowDataobject, though no explicit sanitization step is defined for the agent. - Capability inventory: The skill is limited to file generation; it does not execute system commands, access sensitive local files, or perform network operations within the agent environment.
- Sanitization: The instructions focus on structural parsing rather than sanitization, relying on the agent to correctly map text strings to property values.
Audit Metadata