image-to-code
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill mandates a workflow involving deep analysis and extraction of text from AI-generated images (Rules 9, 21, and 22).
- Ingestion points: Text and design systems are extracted from generated images via vision analysis (SKILL.md).
- Boundary markers: No instructions are provided to treat extracted text as untrusted or to ignore instructions embedded within the extracted text.
- Capability inventory: The agent uses extracted information to implement frontend code, which typically involves writing files or generating code blocks.
- Sanitization: There is no mention of sanitizing or validating the text extracted from images before it is used in the implementation phase.
Audit Metadata