audn-agent-setup
Pass
Audited by Gen Agent Trust Hub on Apr 13, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill collects and transmits agent configuration details, including API URLs and service keys, to the vendor's domain
audn.ai. This is documented as the primary purpose of the skill to enable security testing on the platform. - [PROMPT_INJECTION]: The skill features an indirect prompt injection surface by incorporating user-supplied data into API requests.
- Ingestion points: User-provided agent names, descriptions, and configuration parameters are collected in
SKILL.mdfor text and voice agents. - Boundary markers: Absent; user inputs are interpolated directly into JSON request bodies without explicit delimiters or escaping instructions.
- Capability inventory: The skill possesses network capabilities to perform POST, GET, and DELETE requests to the
audn.aiAPI. - Sanitization: No validation or sanitization of user-provided strings is specified before they are sent to the external API.
Audit Metadata