Skip to main content Module agent Copy item path Source pub use error::AgentResultError ;error memory prebuilt task turn_engine ActorAgent Marker type for actor-based agents. ActorAgentHandle Handle for an actor-based agent that contains both the agent and the
address of its actor. Use addr() to send messages directly or publish
Tasks to subscribed Topic<Task>. AgentBuilder Builder for creating BaseAgent instances from AgentDeriveT implementations AgentConfig BaseAgent Base agent type that wraps an AgentDeriveT implementation with additional runtime components Context Execution context shared across an agent run. DirectAgent Marker type for direct (non-actor) agents. DirectAgentHandle Handle for a direct agent containing the agent instance and an event stream
receiver. Use agent.run(...) for one-shot calls or agent.run_stream(...)
to receive streaming outputs. EventHelper Helper for managing event emissions ExecutorConfig Configuration for executors MemoryHelper Helper for managing agent memory operations ToolProcessor Handles all tool-related operations in a centralized manner AgentProtocol ContextError HookOutcome Outcome for hook execution: continue or abort the run. TurnResult Result of processing a single turn in the agent’s execution AgentDeriveT Core trait that defines agent metadata and behavior
This trait is implemented via the #[agent] macro AgentExecutor Base trait for agent execution strategies AgentHooks Lifecycle hooks that allow observing and customizing agent behavior.
Implementers can observe agent creation, per-run lifecycle (start/complete),
per-turn lifecycle (start/complete) for multi-turn executors, and tool
execution (pre-call gating, start, result, error). AgentOutputT Trait for agent output types that can generate structured output schemas