lovstudio-image-creator
Pass
Audited by Gen Agent Trust Hub on Apr 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
gen_image.pyscript automatically installs thePillowandgoogle-genaiPython packages from the standard package registry usingpipif they are not already present on the user's system. - [DATA_EXFILTRATION]: The skill transmits user prompts and retrieves image data through a third-party API endpoint (
https://zenmux.ai/api/vertex-ai) which acts as a proxy for the underlying AI models. - [PROMPT_INJECTION]: The 'Code-Based Rendering' mechanism creates a vulnerability surface for indirect prompt injection.
- Ingestion points: User-provided descriptions are used to generate self-contained HTML and React code in
SKILL.md. - Boundary markers: The template provided for HTML generation does not include delimiters or instructions to the agent to disregard potential instructions embedded in the user's data.
- Capability inventory: The generated HTML is executed locally using Playwright in
scripts/render_to_png.py, which operates in a browser context with access to the local file system via thefile://protocol. The skill additionally possessesBashandWritecapabilities. - Sanitization: No validation or escaping of user-provided content is implemented before it is interpolated into the HTML/JavaScript code for rendering.
Audit Metadata