ugc-skincare
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The prompt templates in
SKILL.mdare designed to interpolate user-provided data directly into instructions destined for an AI video generation model. - Ingestion points: The skill utilizes several placeholders in
SKILL.md(e.g.,[PRODUCT],[PRODUCT TEXTURE],[ADDITIONAL DETAILS]) where user-controlled strings are inserted into the final prompt. - Boundary markers: No specific delimiters or "ignore instructions" directives are present to separate user-provided content from the template instructions.
- Capability inventory: The templates are intended for use with a video generation API (
rawugc-api), which involves network-based processing of the generated strings. - Sanitization: The skill lacks instructions or mechanisms for validating or sanitizing user-provided strings prior to their interpolation.
Audit Metadata