Namespace LMKit.Agents.Orchestration
Classes
- AfterAgentExecutionEventArgs
Event arguments for after agent execution in orchestration.
Provides access to the execution result for logging or post-processing.
- BeforeAgentExecutionEventArgs
Event arguments for before agent execution in orchestration.
Allows handlers to modify the input or skip the agent entirely.
- OrchestrationContext
Provides shared context and state during orchestrated execution.
The context is passed between agents in a workflow and can be used to share data, accumulate results, and coordinate execution.
- OrchestrationOptions
Configuration options for orchestration execution.
- OrchestrationResult
Represents the result of an orchestrated multi-agent execution.
- OrchestratorBase
Base class for orchestrator implementations.
Provides common functionality for timeout handling, context management, and result building.
- ParallelAgent
Represents an agent in a parallel orchestration.
- ParallelOrchestrator
Orchestrator that executes multiple agents concurrently and aggregates results.
All agents receive the same input and execute in parallel. Results are combined using a configurable aggregation strategy.
- PipelineOrchestrator
Orchestrator that executes agents sequentially in a pipeline.
Each agent receives the output of the previous agent as its input, forming a chain of processing steps.
- PipelineStage
Represents a stage in a pipeline orchestration.
- RouterOrchestrator
Orchestrator that routes the input to the most appropriate single agent.
Uses a routing function or agent to select which specialist agent should handle the request, then executes only that agent.
- SupervisorOrchestrator
Orchestrator where a supervisor agent dynamically coordinates worker agents.
The supervisor agent decides which worker agents to invoke and how to combine their outputs to accomplish the task. This enables flexible, adaptive workflows.
Interfaces
- IOrchestrator
Defines the contract for multi-agent orchestration strategies.
An orchestrator coordinates the execution of multiple agents to accomplish a task, managing how agents are selected, sequenced, and how their outputs are combined.