m365-agents-dotnet

作者: microsoft

Microsoft 365 Agents SDK for .NET. Build multichannel agents for Teams/M365/Copilot Studio with ASP.NET Core hosting, AgentApplication routing, and MSAL-based auth. Triggers: "Microsoft 365 Agents SDK", "Microsoft.Agents", "AddAgentApplicationOptions", "AgentApplication", "AddAgentAspNetAuthentication", "Copilot Studio client", "IAgentHttpAdapter".

npx skills add https://github.com/microsoft/skills --skill m365-agents-dotnet

Microsoft 365 Agents SDK (.NET)

Overview

Build enterprise agents for Microsoft 365, Teams, and Copilot Studio using the Microsoft.Agents SDK with ASP.NET Core hosting, agent routing, and MSAL-based authentication.

Before implementation

  • Use the microsoft-docs MCP to verify the latest APIs for AddAgent, AgentApplication, and authentication options.
  • Confirm package versions in NuGet for the Microsoft.Agents.* packages you plan to use.

Installation

dotnet add package Microsoft.Agents.Hosting.AspNetCore
dotnet add package Microsoft.Agents.Authentication.Msal
dotnet add package Microsoft.Agents.Storage
dotnet add package Microsoft.Agents.CopilotStudio.Client
dotnet add package Microsoft.Identity.Client.Extensions.Msal

Configuration (appsettings.json)

{
  "TokenValidation": {
    "Enabled": true,
    "Audiences": ["{{ClientId}}"],
    "TenantId": "{{TenantId}}"
  },
  "AgentApplication": {
    "StartTypingTimer": false,
    "RemoveRecipientMention": false,
    "NormalizeMentions": false
  },
  "Connections": {
    "ServiceConnection": {
      "Settings": {
        "AuthType": "ClientSecret",
        "ClientId": "{{ClientId}}",
        "ClientSecret": "{{ClientSecret}}",
        "AuthorityEndpoint": "https://login.microsoftonline.com/{{TenantId}}",
        "Scopes": ["https://api.botframework.com/.default"]
      }
    }
  },
  "ConnectionsMap": [
    {
      "ServiceUrl": "*",
      "Connection": "ServiceConnection"
    }
  ],
  "CopilotStudioClientSettings": {
    "DirectConnectUrl": "",
    "EnvironmentId": "",
    "SchemaName": "",
    "TenantId": "",
    "AppClientId": "",
    "AppClientSecret": ""
  }
}

Core Workflow: ASP.NET Core agent host

using Microsoft.Agents.Builder;
using Microsoft.Agents.Hosting.AspNetCore;
using Microsoft.Agents.Storage;
using Microsoft.AspNetCore.Builder;
using Microsoft.AspNetCore.Http;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;

var builder = WebApplication.CreateBuilder(args);

builder.Services.AddHttpClient();
builder.AddAgentApplicationOptions();
builder.AddAgent<MyAgent>();
builder.Services.AddSingleton<IStorage, MemoryStorage>();

builder.Services.AddControllers();
builder.Services.AddAgentAspNetAuthentication(builder.Configuration);

WebApplication app = builder.Build();

app.UseAuthentication();
app.UseAuthorization();

app.MapGet("/", () => "Microsoft Agents SDK Sample");

var incomingRoute = app.MapPost("/api/messages",
    async (HttpRequest request, HttpResponse response, IAgentHttpAdapter adapter, IAgent agent, CancellationToken ct) =>
    {
        await adapter.ProcessAsync(request, response, agent, ct);
    });

if (!app.Environment.IsDevelopment())
{
    incomingRoute.RequireAuthorization();
}
else
{
    app.Urls.Add("http://localhost:3978");
}

app.Run();

AgentApplication routing

using Microsoft.Agents.Builder;
using Microsoft.Agents.Builder.App;
using Microsoft.Agents.Builder.State;
using Microsoft.Agents.Core.Models;
using System;
using System.Threading;
using System.Threading.Tasks;

public sealed class MyAgent : AgentApplication
{
    public MyAgent(AgentApplicationOptions options) : base(options)
    {
        OnConversationUpdate(ConversationUpdateEvents.MembersAdded, WelcomeAsync);
        OnActivity(ActivityTypes.Message, OnMessageAsync, rank: RouteRank.Last);
        OnTurnError(OnTurnErrorAsync);
    }

    private static async Task WelcomeAsync(ITurnContext turnContext, ITurnState turnState, CancellationToken ct)
    {
        foreach (ChannelAccount member in turnContext.Activity.MembersAdded)
        {
            if (member.Id != turnContext.Activity.Recipient.Id)
            {
                await turnContext.SendActivityAsync(
                    MessageFactory.Text("Welcome to the agent."),
                    ct);
            }
        }
    }

    private static async Task OnMessageAsync(ITurnContext turnContext, ITurnState turnState, CancellationToken ct)
    {
        await turnContext.SendActivityAsync(
            MessageFactory.Text($"You said: {turnContext.Activity.Text}"),
            ct);
    }

    private static async Task OnTurnErrorAsync(
        ITurnContext turnContext,
        ITurnState turnState,
        Exception exception,
        CancellationToken ct)
    {
        await turnState.Conversation.DeleteStateAsync(turnContext, ct);

        var endOfConversation = Activity.CreateEndOfConversationActivity();
        endOfConversation.Code = EndOfConversationCodes.Error;
        endOfConversation.Text = exception.Message;
        await turnContext.SendActivityAsync(endOfConversation, ct);
    }
}

Copilot Studio direct-to-engine client

DelegatingHandler for token acquisition (interactive flow)

using System.Net.Http.Headers;
using Microsoft.Agents.CopilotStudio.Client;
using Microsoft.Identity.Client;

internal sealed class AddTokenHandler : DelegatingHandler
{
    private readonly SampleConnectionSettings _settings;

    public AddTokenHandler(SampleConnectionSettings settings) : base(new HttpClientHandler())
    {
        _settings = settings;
    }

    protected override async Task<HttpResponseMessage> SendAsync(
        HttpRequestMessage request,
        CancellationToken cancellationToken)
    {
        if (request.Headers.Authorization is null)
        {
            string[] scopes = [CopilotClient.ScopeFromSettings(_settings)];

            IPublicClientApplication app = PublicClientApplicationBuilder
                .Create(_settings.AppClientId)
                .WithAuthority(AadAuthorityAudience.AzureAdMyOrg)
                .WithTenantId(_settings.TenantId)
                .WithRedirectUri("http://localhost")
                .Build();

            AuthenticationResult authResponse;
            try
            {
                var account = (await app.GetAccountsAsync()).FirstOrDefault();
                authResponse = await app.AcquireTokenSilent(scopes, account).ExecuteAsync(cancellationToken);
            }
            catch (MsalUiRequiredException)
            {
                authResponse = await app.AcquireTokenInteractive(scopes).ExecuteAsync(cancellationToken);
            }

            request.Headers.Authorization = new AuthenticationHeaderValue("Bearer", authResponse.AccessToken);
        }

        return await base.SendAsync(request, cancellationToken);
    }
}

Console host with CopilotClient

using Microsoft.Agents.CopilotStudio.Client;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;

HostApplicationBuilder builder = Host.CreateApplicationBuilder(args);

var settings = new SampleConnectionSettings(
    builder.Configuration.GetSection("CopilotStudioClientSettings"));

builder.Services.AddHttpClient("mcs").ConfigurePrimaryHttpMessageHandler(() =>
{
    return new AddTokenHandler(settings);
});

builder.Services
    .AddSingleton(settings)
    .AddTransient<CopilotClient>(sp =>
    {
        var logger = sp.GetRequiredService<ILoggerFactory>().CreateLogger<CopilotClient>();
        return new CopilotClient(settings, sp.GetRequiredService<IHttpClientFactory>(), logger, "mcs");
    });

IHost host = builder.Build();
var client = host.Services.GetRequiredService<CopilotClient>();

await foreach (var activity in client.StartConversationAsync(emitStartConversationEvent: true))
{
    Console.WriteLine(activity.Type);
}

await foreach (var activity in client.AskQuestionAsync("Hello!", null))
{
    Console.WriteLine(activity.Type);
}

Best Practices

  1. Use AgentApplication subclasses to centralize routing and error handling.
  2. Use MemoryStorage only for development; use persisted storage in production.
  3. Enable TokenValidation in production and require authorization on /api/messages.
  4. Keep auth secrets in configuration providers (Key Vault, managed identity, env vars).
  5. Reuse HttpClient from IHttpClientFactory and cache MSAL tokens.
  6. Prefer async handlers and pass CancellationToken to SDK calls.

Reference Links

ResourceURL
Microsoft 365 Agents SDKhttps://learn.microsoft.com/en-us/microsoft-365/agents-sdk/
AddAgent APIhttps://learn.microsoft.com/en-us/dotnet/api/microsoft.agents.hosting.aspnetcore.servicecollectionextensions.addagent?view=m365-agents-sdk
AgentApplication APIhttps://learn.microsoft.com/en-us/dotnet/api/microsoft.agents.builder.app.agentapplication?view=m365-agents-sdk
Auth configuration optionshttps://learn.microsoft.com/en-us/microsoft-365/agents-sdk/microsoft-authentication-library-configuration-options
Copilot Studio integrationhttps://learn.microsoft.com/en-us/microsoft-365/agents-sdk/integrate-with-mcs
GitHub sampleshttps://github.com/microsoft/agents

來自 microsoft 的更多技能

oss-growth
microsoft
開源增長駭客角色
agent-framework-azure-ai-py
microsoft
使用Microsoft Agent Framework Python SDK(agent-framework-azure-ai)构建Azure AI Foundry代理。适用于使用AzureAIAgentsProvider创建持久化代理、使用托管工具(代码解释器、文件搜索、网络搜索)、集成MCP服务器、管理对话线程或实现流式响应。涵盖函数工具、结构化输出和多工具代理。
development
airunway-aks-setup
microsoft
在AKS上設定AI Runway——從裸叢集到執行模型。涵蓋叢集驗證、控制器安裝、GPU評估、供應商設定及首次部署。時機:「設定AI Runway」、「上線AKS叢集」、「安裝AI Runway」、「airunway設定」、「部署模型至AKS」、「在AKS上進行GPU推論」、「在AKS上設定KAITO」、「在AKS上執行LLM」、「在AKS上使用vLLM」、「在AKS上設定模型服務」、「AI Runway控制器」。
devops
appinsights-instrumentation
microsoft
使用Azure Application Insights檢測Web應用程式的指南。提供遙測模式、SDK設定與組態參考。適用時機:如何檢測應用程式、App Insights SDK、遙測模式、什麼是App Insights、Application Insights指南、檢測範例、APM最佳實踐。
devops
applicationinsights-web-ts
microsoft
使用Application Insights JavaScript SDK(@microsoft/applicationinsights-web)為瀏覽器/Web應用程式進行檢測。適用於真實使用者監控(RUM)——頁面檢視、點擊、AJAX/fetch依賴、例外、自訂事件,以及與後端OpenTelemetry追蹤關聯的瀏覽器端GenAI代理追蹤。涵蓋SDK載入器指令碼與npm設定、框架擴充(React、React Native、Angular)、點擊分析、遙測初始化器,以及從瀏覽器發出的代理/工具/模型span的OTel GenAI語意慣例。
devops
azure-ai-anomalydetector-java
microsoft
使用適用於 Java 的 Azure AI 異常偵測器 SDK 建置異常偵測應用程式。在實作單變量/多變量異常偵測、時間序列分析或 AI 驅動監控時使用。
development
azure-ai-language-conversations-py
microsoft
使用 azure-ai-language-conversations Python SDK 實作對話語言理解(CLU)。當使用 ConversationAnalysisClient 分析對話意圖與實體、建置 NLP 功能,或將語言理解整合至應用程式時使用。
development
azure-ai-ml-py
microsoft
Azure Machine Learning SDK v2 for Python。用於機器學習工作區、作業、模型、資料集、計算資源與管線。 觸發詞:「azure-ai-ml」、「MLClient」、「workspace」、「model registry」、「training jobs」、「datasets」。
development