agents.planning.plan_execute_v3.state¶

State schema for Plan-and-Execute V3 Agent.

This module defines the state schema used by the Plan-and-Execute V3 agent, extending MessagesState with computed fields for plan tracking.

Classes¶

State schema for Plan-and-Execute V3 agent.

class agents.planning.plan_execute_v3.state.PlanExecuteV3State(/, **data)¶

Bases: haive.core.schema.prebuilt.messages_state.MessagesState

State schema for Plan-and-Execute V3 agent.

This state is shared across the planner, executor, evaluator, and replanner sub-agents to maintain full context throughout the execution.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

add_evaluation(evaluation)¶

Add an evaluation result.

Parameters:: evaluation (agents.planning.plan_execute_v3.models.PlanEvaluation)
Return type:: None

add_step_execution(execution)¶

Add a step execution result and update plan.

Parameters:: execution (agents.planning.plan_execute_v3.models.StepExecution)
Return type:: None

revise_plan(new_plan)¶

Replace current plan with a revised version.

Parameters:: new_plan (agents.planning.plan_execute_v3.models.ExecutionPlan)
Return type:: None

property current_step: str | None¶

Get the current step description for the executor.

property execution_summary: str¶

Get a summary of the entire execution.

property execution_time: float | None¶

Total execution time in seconds.

property key_findings: list[str]¶

Extract key findings from executions.

property objective: str¶

Extract the objective from the plan or messages.

property plan_status: str¶

Get formatted plan status for agents.

property previous_results: str¶

Get formatted previous step execution results.

property should_evaluate: bool¶

Determine if we should run evaluation.