Class AgentExecutionResult

Namespace: LMKit.Agents

Assembly: LM-Kit.NET.dll

Represents the outcome of an agent task execution.

public sealed class AgentExecutionResult

Inheritance: object

AgentExecutionResult

Inherited Members: object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.ReferenceEquals(object, object)

object.ToString()

Examples

Processing an execution result:

using LMKit.Model;
using LMKit.Agents;

using var model = new LM("path/to/model.gguf");
var agent = new Agent(model);
agent.Identity = new AgentIdentity("Assistant", "Help with questions.");

var result = agent.Run("What is the capital of France?");

if (result.IsSuccess)
{
    Console.WriteLine($"Response: {result.Content}");
    Console.WriteLine($"Completed in {result.Duration.TotalMilliseconds:F0}ms");
    Console.WriteLine($"Inference calls: {result.InferenceCount}");
}
else if (result.IsFailed)
{
    Console.WriteLine($"Error: {result.Error.Message}");
}
else if (result.IsCancelled)
{
    Console.WriteLine("Task was cancelled.");
}

Inspecting tool calls:

using LMKit.Model;
using LMKit.Agents;

using var model = new LM("path/to/model.gguf");
var agent = Agent.CreateBuilder(model)
    .WithPersona("Research Assistant")
    .WithPlanning(PlanningStrategy.ReAct)
    .WithTools(t => t.Register(new SearchTool()))
    .Build();

var result = await agent.RunAsync("Find recent news about AI.");

Console.WriteLine($"Response: {result.Content}");
Console.WriteLine($"Tool calls made: {result.ToolCalls.Count}");

foreach (var toolCall in result.ToolCalls)
{
    Console.WriteLine($"  Tool: {toolCall.ToolName}");
    Console.WriteLine($"  Status: {toolCall.ResultType}");
}

Remarks

The execution result contains everything you need to understand what happened during agent execution: the final response, any tool calls made, reasoning traces, timing information, and error details if something went wrong.

Result Components

Content: The final text response from the agent.
ToolCalls: A record of all tool invocations during execution.
ReasoningTrace: Internal reasoning steps if a planning strategy produced them.
Status: The terminal state indicating success, failure, or cancellation.
Duration: Total execution time.
InferenceCount: Number of LLM inference calls made.

Checking Results

Use the convenience properties IsSuccess, IsFailed, and IsCancelled to quickly check the outcome.

Properties

Content: Gets the final text content produced by the agent.

Duration: Gets the total execution duration.

Error: Gets the exception that caused execution to fail, if applicable.

InferenceCount: Gets the number of LLM inference calls made during execution.

IsCancelled: Gets a value indicating whether the execution was cancelled.

IsFailed: Gets a value indicating whether the execution failed.

IsSuccess: Gets a value indicating whether the execution completed successfully.

ReasoningTrace: Gets the reasoning trace produced during execution, if any.

Status: Gets the execution status indicating how the task completed.

ToolCalls: Gets the list of tool calls made during execution, in order of invocation.

Methods

Cancelled(string, IReadOnlyList<ToolCallResult>, TimeSpan, int): Creates a cancelled execution result.

Failure(Exception, string, IReadOnlyList<ToolCallResult>, TimeSpan, int): Creates a failed execution result.

Success(string, IReadOnlyList<ToolCallResult>, string, TimeSpan, int): Creates a successful execution result.

Table of Contents

Class AgentExecutionResult

Examples

Remarks

Properties

Methods