image-to-code
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a clear and professional methodology for UI implementation based on visual references, emphasizing fidelity, accessibility, and consistency.
- [NO_CODE]: The skill package consists entirely of documentation and instructional markdown files; it does not include or execute any scripts, binaries, or automated configuration files.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data (screenshots, design images, and Figma files), which serves as an attack surface for instructions hidden within visual content. 1. Ingestion points: Visual reference images and Figma design data described in SKILL.md and references/image-generation.md. 2. Boundary markers: The instructions lack explicit directives to ignore or sanitize text-based instructions that may be embedded within the processed images. 3. Capability inventory: The agent is tasked with implementing code (file system writes) and using the imagegen-frontend-web tool. 4. Sanitization: No specific mechanisms for sanitizing content extracted from images or design files are provided.
Audit Metadata