Ask a question to get started
Enter to send•Shift+Enter new line
before_model( self, state: AgentState[Any], runtime: Runtime[Any] )
Inject one-shot policy nudges before the model acts.
