product-demo
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided product descriptions, uploaded screenshots, and captures content from public web pages to generate video scripts. This creates a surface where text embedded in screenshots or external web pages could be interpreted as instructions by the agent.
- Ingestion points: User input for product details and screenshots uploaded by the user or captured from public URLs (SKILL.md).
- Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard commands found within the processed data.
- Capability inventory: The skill utilizes ElevenLabs MCP for voiceover generation and references various internal film-generation tools (SKILL.md).
- Sanitization: The skill does not define any methods for sanitizing or escaping text extracted from visual assets before processing it.
Audit Metadata