illustration-halftone-line
Warn
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The script 'scripts/halftone.mjs' uses 'new Function()' and 'eval()' to interpret and execute JavaScript logic embedded within tags like and placeholders like ${} in input SVG files. Additionally, 'scripts/render.mjs' dynamically imports the 'playwright' or 'playwright-core' library using computed paths and URL-based imports.
- [INDIRECT_PROMPT_INJECTION]: The skill relies on a workflow that processes user-controllable data files containing executable logic, creating a significant attack surface for local code execution.
- Ingestion points: The compiler script 'scripts/halftone.mjs' processes SVG source files provided as command-line arguments.
- Boundary markers: Absent. There are no mechanisms to isolate the execution of embedded JavaScript or restrict the scope of the evaluated expressions.
- Capability inventory: The skill has the capability to execute arbitrary code via Node.js ('eval', 'new Function'), write to the local file system ('fs.writeFileSync'), and control a headless browser ('playwright').
- Sanitization: Absent. The scripts do not sanitize or validate the content of the JavaScript expressions extracted from the SVG files before evaluation.
Audit Metadata