moai-domain-svg-infographic
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The
scripts/render.mjsscript executes a Chromium-family browser usingchild_process.spawnSync. The browser path is resolved from environment variables likeCHROME_PATHor common system locations. This enables the skill to run external binary processes on the host system. - [INDIRECT_PROMPT_INJECTION]: The skill processes SVG files that are rendered in a browser context by
scripts/render.mjs. Because the rendering script does not use flags like--disable-javascript, any malicious scripts or event handlers embedded in an SVG could be executed during rendering. - Ingestion points: SVG source files passed to the linting and rendering scripts.
- Boundary markers: The
SKILL.mdinstructions include a manual checklist and workflow that advises against scripting, but these are not programmatically enforced. - Capability inventory:
scripts/render.mjscreates temporary files and directories and launches a headless browser process. - Sanitization: The script removes XML and DOCTYPE declarations from the input SVG but does not filter out
<script>tags or JavaScript event handlers. - [DYNAMIC_EXECUTION]: The
scripts/render.mjsscript dynamically generates a temporary HTML wrapper and constructs command-line arguments for the browser to perform the rendering and screenshot capture.
Audit Metadata