video-watermark
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data in the form of media URLs and natural language user instructions, creating a vulnerability surface.
- Ingestion points: The
image_urlsparameter and usermessagesarray in the API requests documented inSKILL.md. - Boundary markers: There are no explicit delimiters or boundary markers specified in the documentation to isolate untrusted data from the agent's core instructions.
- Capability inventory: The skill communicates with the
eachsense-agent.core.eachlabs.runendpoint to perform tasks; it does not request local shell access, file system writes, or execution of local binaries. - Sanitization: The skill instructions do not describe any mechanisms for sanitizing or validating the content of the external URLs or user inputs.
Audit Metadata