agents.planning.plan_execute_v3.engines¶

Engines for Plan-and-Execute V3 Agent.

This module contains the specialized engines used by the Plan-and-Execute V3 agent for planning, validation, execution, and monitoring.

Classes¶

`ExecutorEngine`	Engine for executing individual plan steps.
`MonitorEngine`	Engine for monitoring plan execution.
`PlannerEngine`	Engine for generating execution plans.
`ReplannerEngine`	Engine for replanning when execution fails.
`ValidatorEngine`	Engine for validating and refining plans.

Module Contents¶

class agents.planning.plan_execute_v3.engines.ExecutorEngine(llm_config, tools)¶

Engine for executing individual plan steps.

Initialize executor engine.

Parameters:

llm_config (haive.core.engine.aug_llm.AugLLMConfig)
tools (list[langchain_core.tools.BaseTool])

async execute_step(step, context, previous_results)¶

Execute a single plan step.

Parameters:

step (agents.planning.plan_execute_v3.models.Step) – The step to execute
context (dict[str, Any]) – Shared execution context
previous_results (dict[str, Any]) – Results from previous steps

Returns:

Tuple of (result, error_message)

Return type:

tuple[Any, str | None]

class agents.planning.plan_execute_v3.engines.MonitorEngine(llm_config)¶

Engine for monitoring plan execution.

Initialize monitor engine.

Parameters:: llm_config (haive.core.engine.aug_llm.AugLLMConfig)

async analyze_execution(plan)¶

Analyze plan execution progress.

Parameters:: plan (agents.planning.plan_execute_v3.models.Plan) – The plan being executed
Returns:: Analysis results with metrics and suggestions
Return type:: dict[str, Any]

class agents.planning.plan_execute_v3.engines.PlannerEngine(llm_config, tools)¶

Engine for generating execution plans.

Initialize planner engine.

Parameters:

llm_config (haive.core.engine.aug_llm.AugLLMConfig) – LLM configuration
tools (list[langchain_core.tools.BaseTool]) – Available tools for planning

async generate_plan(goal)¶

Generate an execution plan for the given goal.

Parameters:: goal (str) – The goal to achieve
Returns:: Generated execution plan
Return type:: agents.planning.plan_execute_v3.models.Plan

class agents.planning.plan_execute_v3.engines.ReplannerEngine(llm_config, tools)¶

Engine for replanning when execution fails.

Initialize replanner engine.

Parameters:

llm_config (haive.core.engine.aug_llm.AugLLMConfig)
tools (list[langchain_core.tools.BaseTool])

async create_revised_plan(original_plan, issues)¶

Create a revised plan based on execution issues.

Parameters:

original_plan (agents.planning.plan_execute_v3.models.Plan) – The plan that encountered issues
issues (list[str]) – List of issues encountered

Returns:

Revised execution plan

Return type:

agents.planning.plan_execute_v3.models.Plan

class agents.planning.plan_execute_v3.engines.ValidatorEngine(llm_config, tools)¶

Engine for validating and refining plans.

Initialize validator engine.

Parameters:

llm_config (haive.core.engine.aug_llm.AugLLMConfig)
tools (list[langchain_core.tools.BaseTool])

async validate_plan(plan)¶

Validate a plan and suggest refinements.

Parameters:: plan (agents.planning.plan_execute_v3.models.Plan) – The plan to validate
Returns:: Validation result with issues and suggestions
Return type:: agents.planning.plan_execute_v3.models.PlanValidationResult