Design a Twilio Agent Connect Integration in Codex
Map the connection between an existing AI agent and Twilio conversation services using a stubbed support request and a local fallback test.
中文Setup for Codex (Twilio Developer Kit)
Use the full Twilio Developer Kit in Codex. This exercise requires only a fake agent function and in-memory connection adapter. Live integration uses the appropriate Twilio SDK and your chosen agent runtime.
- Add Twilio Developer Kit from /plugins so the related Orchestrator, Memory, and Knowledge guides are available.
- Invoke $twilio-agent-connect in a scratch project and request the local stub described below. Keep credentials and real conversation data out of the fixture.
Example task
Give a local adapter two fixed requests, demo-user asking for store hours and demo-user repeating the request when the fake agent returns a timeout. Ask for a mock read-only answer on success and a handoff record for the failure.Expected result
The successful stub returns the configured demo store hours. The timeout produces one synthetic handoff record. The diagram identifies the agent logic and connection adapter as separate components.
Upstream documentation
Original SKILL.md on GitHub · MIT license · Source reviewed
Frequently asked question
Does Twilio Agent Connect replace an existing agent runtime?
The SDK connects your selected runtime to Twilio conversation services. Your runtime still owns the model behavior and the application logic.