etsy-tony-full-engraving-photo
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data in the form of engraving images and transcribed personalization text. This content is used to construct prompts for the
image_gentool, creating a surface where malicious instructions embedded in the design could influence the output. - Ingestion points: User-uploaded 'Engraving source' images and 'Exact personalization' text as specified in
SKILL.md. - Boundary markers: The instructions in
references/product-master.mdandreferences/photo-set.mdcontain directives to transcribe text 'character-for-character' and 'not to rewrite or improve' it, providing a policy-based boundary. - Capability inventory: The agent uses the
image_gentool for content creation and basic file system operations (copy/move) to organize the output images. - Sanitization: The skill lacks technical sanitization or validation for text extracted from user-supplied images before it is interpolated into generation prompts.
Audit Metadata