visualize
Pass
Audited by Gen Agent Trust Hub on May 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The helper script
scripts/render-mermaid.tsinvokes the Mermaid CLI (mmdc) usingexecFileSync. This is the primary method for rendering visual content.\n- [EXTERNAL_DOWNLOADS]: The script utilizesnpxto execute the@mermaid-js/mermaid-clipackage, which involves downloading it from the NPM registry if it is not already available.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its handling of arbitrary conversation context and file content.\n - Ingestion points: Content is sourced from conversation history and user-specified files in
SKILL.mdandscripts/render-mermaid.ts.\n - Boundary markers: No delimiters or instructions are used to prevent the agent from acting on malicious instructions contained within the processed data.\n
- Capability inventory: The skill allows reading local files and executing commands through a subprocess.\n
- Sanitization: There is no input validation or sanitization of the content before it is passed to the renderer.
Audit Metadata