agents-core/ai.koog.agents.core.feature.handler.llm

Package-level declarations

Types

data class LLMCallCompletedContext(val runId: String, val prompt: Prompt, val model: LLModel, val tools: List<ToolDescriptor>, val responses: List<Message.Response>, val moderationResponse: ModerationResult?) : LLMCallEventContext

Represents the context for handling an after LLM call event.

LLMCallCompletedHandler

fun interface LLMCallCompletedHandler

Represents a functional interface for handling operations or logic that should occur after a call to a large language model (LLM) is made. The implementation of this interface provides a mechanism to perform custom logic or processing based on the provided inputs, such as the prompt, tools, model, and generated responses.

LLMCallEventContext

interface LLMCallEventContext : AgentLifecycleEventContext

Represents the context for handling LLM-specific events within the framework.

LLMCallEventHandler

class LLMCallEventHandler

A handler responsible for managing the execution flow of a Large Language Model (LLM) call. It allows customization of logic to be executed before and after the LLM is called.

LLMCallStartingContext

data class LLMCallStartingContext(val runId: String, val prompt: Prompt, val model: LLModel, val tools: List<ToolDescriptor>) : LLMCallEventContext

Represents the context for handling a before LLM call event.

LLMCallStartingHandler

fun interface LLMCallStartingHandler

A functional interface implemented to handle logic that occurs before invoking a large language model (LLM). It allows preprocessing steps or validation based on the provided prompt, available tools, targeted LLM model, and a unique run identifier.