hi-fi-design
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, including codebase context, UI kits, and design system files.
- Ingestion Points: Instructions explicitly tell the agent to 'find existing UI kits', 'copy ALL relevant components', 'read ALL relevant examples', and 'Import their codebase' or 'link a local codebase'. These files are processed to build design assumptions and prototypes.
- Boundary Markers: The instructions lack explicit directives for the agent to ignore potentially malicious instructions embedded within these external files or design system documentation.
- Capability Inventory: The agent is empowered to generate HTML/CSS/JS/SVG prototypes and execute tools to 'check, verify and iterate on the design'. While this does not involve direct system shell access, it allows the rendering of potentially malicious scripts in the designer view/canvas.
- Sanitization: There are no mentioned sanitization steps for content imported from the user's codebase or external UI kits.
Audit Metadata