koog-ktor/ai.koog.ktor/KoogAgentsConfig/AgentConfig

AgentConfig

inner class AgentConfig(source)

Configuration class for managing agent-specific settings and tools.

The AgentConfig class allows customization of the agent's behavior, including:

Setting the prompt and language model to use
Managing tools available to the agent
Defining strategies for handling missing tools
Installing additional features

Constructors

AgentConfig

constructor()

Properties

maxAgentIterations

var maxAgentIterations: Int

Specifies the maximum number of iterations an agent is permitted to perform during its execution cycle.

missingToolsConversionStrategy

var missingToolsConversionStrategy: MissingToolsConversionStrategy

Defines the strategy for handling tool calls present in the prompt that do not have corresponding tool definitions registered in the current context. This is used to convert missing tool information into a format suitable for processing by the model.

Functions

install

fun <TConfig : FeatureConfig> install(feature: AIAgentGraphFeature<TConfig, *>, configure: TConfig.() -> Unit = {})

Adds an AI agent feature to the current configuration by applying the specific configuration logic.

mcp

fun KoogAgentsConfig.AgentConfig.mcp(configure: McpToolsConfig.() -> Unit)

Configures the MCP (Modular Configuration Protocol) tools for the agent with the provided configuration block.

prompt

fun prompt(name: String = "agent", llmParams: LLMParams = LLMParams(), clock: Clock = Clock.System, build: PromptBuilder.() -> Unit)

Configures and sets the prompt for the agent using the provided parameters and a prompt-building function.

registerTools

fun registerTools(build: ToolRegistry.Builder.() -> Unit)

Registers tools into the tool registry using the provided configuration lambda.