ugc-supplements
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill takes user-supplied details (such as product names, brand names, and supplement lists) and interpolates them into prompt templates for a video generation API (
POST /videos/generate). This creates a surface for indirect prompt injection where a malicious user could provide instructions that override the intended video generation behavior. - Ingestion points:
SKILL.mdcontains multiple templates with placeholders like[PRODUCT],[BRAND], and[SUPPLEMENT LIST]. - Boundary markers: The templates do not use explicit delimiters or instructions to the model to ignore potential commands within the user-provided data.
- Capability inventory: The skill is designed to trigger video generation via a specific API endpoint mentioned in the instructions.
- Sanitization: There are no instructions for the agent to sanitize or validate the user-provided content before inserting it into the prompt.
Audit Metadata