vobiz-ai-voice-agents
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as technical documentation for connecting AI voice agents (like Vapi, ElevenLabs, and OpenAI) to telephony services.
- [CREDENTIALS_UNSAFE]: While the document discusses SIP credentials and gateway authentication, it provides placeholders (e.g.,
<unique>.sip.vobiz.ai,<trunk_id>) and specific public gateway IPs (13.233.44.61) for whitelisting purposes. No sensitive secrets, private keys, or actual user credentials are hardcoded. - [EXTERNAL_DOWNLOADS]: The skill references a GitHub repository (
vobiz-ai/Vobiz-X-Pipecat) which belongs to the skill's author (vobiz-ai). It also points to official domains of well-known services likevapi.aiandlivekit.cloud. These are legitimate references for the stated purpose of the skill. - [PROMPT_INJECTION]: The content is focused on infrastructure configuration. There are no attempts to override agent safety filters or inject malicious system-level instructions.
- [INDIRECT_PROMPT_INJECTION]: The skill describes handling external data (audio streams via WebSockets), which is a standard surface for voice AI. However, the skill only provides configuration logic and does not implement unsafe data processing or command execution based on that input.
- [COMMAND_EXECUTION]: There are no shell commands, script generation, or dynamic code execution patterns detected in the skill instructions.
Audit Metadata