Struct PlanExecuteAgent 

pub struct PlanExecuteAgent { /* private fields */ }

Plan-Execute Agent: plans first, executes step by step, and replans on failure.

Supports any LLM provider that implements BaseChatModel.

Implementations

impl PlanExecuteAgent

pub fn new<L>(llm: L, tools: Vec<Arc<dyn BaseTool>>) -> Self

where L: BaseChatModel + Send + Sync + 'static, L::Error: Into<ProviderError>,

Creates a new Plan-Execute Agent

Parameters

• llm - LLM client (any type implementing BaseChatModel)
• tools - available tools

Backward compatibility

Legacy code PlanExecuteAgent::new(openai_chat, tools) still works.

pub fn from_arc( llm: Arc<dyn BaseChatModel<Error = ProviderError> + Send + Sync>, tools: Vec<Arc<dyn BaseTool>>, ) -> Self

Creates an agent from an already-wrapped Arc<dyn BaseChatModel>

pub fn with_max_replans(self, n: usize) -> Self

Sets the maximum number of replans.

pub fn with_agent_factory( self, factory: Arc<dyn Fn() -> Arc<dyn BaseAgent> + Send + Sync>, ) -> Self

Custom factory for the per-step execution agent (P1-2).

The BaseAgent returned by the factory is used for each execute_step of the PlanExecute. Defaults to FunctionCallingAgent. Useful for injecting ReAct / Streaming / custom agents as the execution agent.

pub async fn run(&self, objective: &str) -> Result<String, PlanExecuteError>

Runs the full task: plan -> execute each step -> replan on failure -> summarize

Trait Implementations

impl Orchestrator for PlanExecuteAgent

type Input = String

Input type (usually a String objective/question).

type Output = String

Output type.

fn run_with_context<'life0, 'life1, 'async_trait>( &'life0 self, input: Self::Input, ctx: &'life1 RunContext, ) -> Pin<Box<dyn Future<Output = Result<Self::Output, AgentError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Execution entry point carrying the run context.