The task instance currently running the hook.
Readonly view of the agent's current chat context.
Agent identifier.
Agent instructions.
LLM or realtime model configured for the agent.
Minimum delay between consecutive speech.
Current session for the agent.
Speech-to-text model configured for the agent.
Copy of the agent tool context.
Text-to-speech model configured for the agent.
Agent-level turn handling configuration.
Whether TTS-aligned transcripts are enabled for the agent.
Voice activity detector configured for the agent.
Complete the task with either a result or an error.
Context passed to hooks created with
AgentTask.create().