code-snippet-images
Pass
Audited by Gen Agent Trust Hub on May 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a shell script
scripts/render.shto executegoogle-chromefor image generation. The command is run with the--no-sandboxflag, which disables the browser's primary security sandbox and increases the risk of system compromise if a browser-based exploit is triggered. - [PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection. It interpolates unvalidated content, such as code snippets and filenames, directly into HTML templates which are subsequently rendered by a headless browser.
- Ingestion points: Untrusted data is injected into
<!-- FILENAME -->and<!-- CODE CONTENT -->placeholders inassets/template.htmlandassets/grid-2x2.html. - Boundary markers: No markers are used to delimit or sanitize the injected content.
- Capability inventory: The rendered content is processed by
google-chromeviascripts/render.sh, which can execute JavaScript and, with the sandbox disabled, may interact with the host system. - Sanitization: There is no evidence of HTML escaping or validation of the input strings before they are placed in the templates.
- [EXTERNAL_DOWNLOADS]: The skill's HTML templates fetch font resources from Google Fonts (fonts.googleapis.com).
Audit Metadata