setup-agent

Configuring an AI agent for evaluation requires providing a valid endpoint URL or phone number, assigning a system prompt, and attaching default test sets and metrics to ensure immediate readiness for evaluation runs.

Agent endpoints require strict formatting standards during evaluation setup: telephone numbers must adhere to E.164 formatting, and digital communication channels must use valid URL formatting for successful endpoint validation.

Yes, you can connect voice, SMS, chat, and WebSocket agents to the Coval evaluation platform. The setup process validates endpoints across these various communication protocols and links necessary testing resources.

Yes, valid Coval CLI authentication is required to set up and connect AI agents for evaluation. This authentication manages the agent lifecycle, including endpoint validation and system prompt assignment.

You link test sets and metrics during the agent configuration process. The setup automatically attaches these default resources to the agent endpoint, ensuring the agent is immediately ready for evaluation runs.

Agent endpoint validation fails during setup when communication formats are incorrect. Ensure telephone numbers strictly follow E.164 formatting and digital endpoints use proper URL formatting standards.