email-design
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references external documentation for CLI installation hosted on the inference.sh GitHub repository. This is used to set up the necessary environment for image generation tools.
- [COMMAND_EXECUTION]: The skill uses the
infshcommand-line interface via Bash to run applications on the inference.sh platform, such as HTML-to-image conversion and AI image generation (Flux). These operations involve sending data to and receiving assets from a remote service. - [CREDENTIALS_UNSAFE]: The documentation includes the
infsh logincommand. While this is a standard authentication step for the service, it indicates that the tool manages user credentials to access its features. - [PROMPT_INJECTION]: The skill processes structured HTML input to generate images. This creates a surface for indirect prompt injection where malicious instructions could theoretically be embedded in the HTML data processed by the rendering engine, although the risk is localized to the image generation tool.
Audit Metadata