[{"_1":2,"_282":-5,"_283":-5},"loaderData",{"_3":4,"_22":23,"_24":25,"_229":230},"root",{"_5":6,"_7":8},"toasts",[],"visitor",{"_9":10,"_11":12},"viewportWidth",1280,"visitorSession",{"_13":14,"_16":17,"_18":-5,"_19":20,"_21":-5},"attributes",[15],"BOT","id",5778601,"membership","referenceId","019688c6-4b1b-7cb3-b825-e3308352aedf","userAccount","routes/_public/_layout",{},"routes/_public/mcp/servers/~namespace/~slug/_pages/_layout",{"_26":27,"_68":-5,"_69":28,"_29":70,"_146":147,"_228":116},"availableLocales",[28,40,47,54,61],{"_29":30,"_33":37,"_38":39},"mcpServer",{"_31":32,"_33":34,"_35":36},"description","A Model Context Protocol server that enables Microsoft Copilot Studio to fetch jokes from various sources including Chuck Norris jokes, Dad jokes, and Yo Mama jokes.","name","Jokes MCP Server","readme","# Microsoft Copilot Studio ❤️ MCP\n\nWelcome to the **Microsoft Copilot Studio ❤️ MCP** lab. In this lab, you will learn how to deploy an MCP Server, and how to add it to Microsoft Copilot Studio.\n\n## What is MCP?\n\n[Model Context Protocol (MCP)](https://modelcontextprotocol.io/introduction) is an open protocol that standardizes how applications provide context to LLMs, defined by [Anthropic](https://www.anthropic.com/). MCP provides a standardized way to connect AI models to different data sources and tools. MCP allows makers to seamlessly integrate existing knowledge servers and APIs directly into Copilot Studio.\n\nCurrently, Copilot Studio only supports Tools. To learn more about current capabilities, see [aka.ms/mcsmcp](https://aka.ms/mcsmcp). There are some known issues & planned improvements. These are listed [here](#known-issues-and-planned-improvements).\n\n## MCP vs Connectors\n\nWhen do you use MCP? And when do you use connectors? Will MCP replace connectors?\n\nMCP servers are made available to Copilot Studio using connector infrastructure, so these questions are not really applicable. The fact that MCP servers use the connector infrastructure means they can employ enterprise security and governance controls such as [Virtual Network](https://learn.microsoft.com/power-platform/admin/vnet-support-overview) integration, [Data Loss Prevention](https://learn.microsoft.com/power-platform/admin/wp-data-loss-prevention) controls, [multiple authentication methods](https://learn.microsoft.com/connectors/custom-connectors/#2-secure-your-api)—all of which are available in this release—while supporting real-time data access for AI-powered agents.\n\nSo, MCP and connectors are really **better together**.\n\n## Prerequisites\n\n- Azure Subscription (with payment method added)\n- Have a GitHub account and be logged in\n- Copilot Studio trial or developer account\n- Power Platform environment provisioned\n\n## Lab\n\nTo be able to deploy this MCP Server and use it in Microsoft Copilot Studio, you need to go through the following actions:\n\n- [Create a new GitHub repository based on the template](#create-a-new-github-repository-based-on-the-template)\n- [Deploy the Azure Web App](#deploy-the-azure-web-app)\n- [Create the Power Platform Connector](#create-the-power-platform-connector)\n- [Add the MCP Server as an action in Microsoft Copilot Studio](#add-the-mcp-server-as-an-action-in-microsoft-copilot-studio)\n\n### Create a new GitHub repository based on the template\n\n1. Select `Use this template`\n1. Select `Create a new repository\n\n ![](./assets/usetemplate.png)\n\n1. Select the right `Owner` \n1. Give it a `Repository name`\n1. Optionally you can give it a `Description`\n1. Select `Private`\n1. Select `Create repository`\n\n This will take a little while. After it's done, you will be directed to the newly created repository.\n\n> [!WARNING] \n> After completing the steps in this lab, you will have an MCP Server running on Azure that is publicly available. Ideally, you don't want that. Make sure to delete the Azure Web App after finishing the lab.\n\n### Deploy the Azure Web App\n\n1. Go to `portal.azure.com`\n\n ![Azure Portal](/assets/azureportal.png)\n\n1. Hover over the `Resource groups` tab and select `Create`\n\n ![Resource Group](/assets/selectResourceGroup.png)\n\n1. Click the `Subscription` dropdown and select your subscription from the list\n\n ![Select Subscription](/assets/selectSubscription.png)\n\n1. Click in the `Resource group` name text box and type a resource group name such as \"jokesgrp\"\n\n ![Resource Group Name](/assets/resourcegrpname.png)\n\n1. Select the `Region` dropdown choose the region closest to you\n\n ![Resource Group Region](/assets/resourcegrpregion.png)\n\n1. Click `Review + create`\n\n ![Review and create](/assets/resourcegrpreview.png)\n\n1. Click `Create`\n\n ![Create](/assets/resourcegrpcreate.png)\n\n1. Select `Go to resource group`\n\n ![Go to resource](/assets/gotoresource.png)\n\n1. Select `Create`\n\n ![Create](/assets/inresourcegrpcreate.png)\n\n1. Click into the `search box`\n\n ![Search](/assets/searchmarketplace.png)\n\n1. Type `web app`\n\n1. Select the `Create button` in the Web App search result\n\n ![Search Results](/assets/createwebapp1.png)\n\n1. Select `Web App`\n\n ![Create Web App](/assets/createwebapp2.png)\n\n1. Select the `Name` text input and put in a name for the web app, such as \"wegotjokes\"\n\n ![Web App Name](/assets/webappname.png)\n\n1. Select the `Runtime stack` dropdown and choose `Node 22 LTS` \n\n ![Web App Runtime](/assets/selectruntime.png)\n\n1. Select the `Region` dropdown and choose the region closest to you\n\n ![Region](/assets/webappselectregion.png)\n\n1. Click `Review + create`\n\n ![Review](/assets/webappreview.png)\n\n1. Click `Create`\n\n ![Create](/assets/webappcreate.png)\n\n1. Click `Refresh` to update your deployment progress\n\n ![Refresh](/assets/webapprefresh.png)\n\n1. Once deployment is complete, select `Go to resource`\n\n ![Go to resource](/assets/webappgotoresource.png)\n\n1. Expand the `Deployment` tab\n\n ![Deployment tab](/assets/webappdeployment.png)\n\n1. Click `Deployment Center`\n\n ![Deployment center](/assets/webappdeployctr.png)\n\n1. Click the `Source` dropdown and select `GitHub` \n\n ![Choose source](/assets/webappsource.png)\n\n1. Select `Authorize`\n\n ![Authorize](/assets/deployauthorize.png)\n\n1. Select `Continue`\n\n ![Continue](/assets/githubauthcontinue.png)\n\n1. Select `Authorize AzureAppService`\n\n ![Authorize](/assets/githubauthappservice.png)\n\n1. Click the `Organization` dropdown and choose your organization\n\n ![Org Selection](/assets/githuborg.png)\n\n1. Click on the `Repository` dropdown and choose the GitHub repository you created [earlier](#create-a-new-github-repository-based-on-the-template) from the list\n\n ![Repo Selection](/assets/githubrepo.png)\n\n1. Click the `Branch` dropdown and choose your branch\n\n ![Select Branch](/assets/githubbranch.png)\n\n1. Keep everything else default and select `Save`\n\n ![Save](/assets/githubsave.png)\n\n1. Click `Refresh` to update the deployment status\n\n ![Refresh](/assets/githubrefresh.png)\n\n1. Verify that the deployment is complete\n\n ![Deploy done](/assets/deploydone.png)\n\n### Create the Power Platform Connector\n\n1. Go to [https://make.powerapps.com/](https://make.powerapps.com/)\n1. Select `More` in the left menu\n1. Select the `Discover all` button\n1. Under Data, select the `pin icon` next to `Custom connectors` to pin it to the left menu\n1. Select `Custom connectors` in the left menu\n1. Select `New custom connector`\n1. Select `Create from blank`\n\n ![](./assets/newconnector.png)\n\n1. Give the connector a name (for instance `JokesMCP`)\n1. Select the blue `Continue` button\n1. Select the `Swagger editor` toggle\n \n ![](./assets/swaggereditor.png)\n\n1. Copy the yaml code from [here](./assets/connector.yml) and replace the code in the Swagger editor\n1. In the Swagger editor, replace `dummyurl.azurewebsites.net` with the URL of the web app you created [earlier](#deploy-the-azure-web-app). Make sure to remove `https://` and everything after `azurewebsites.net`\n1. Select `Create connector` to create the connector\n\n### Add the MCP Server as an action in Microsoft Copilot Studio\n\nNow you have an MCP Server running in Azure, and a connector available in the Power Platform. This step is about creating an agent in Microsoft Copilot Studio and adding the MCP Server to the agent.\n\n1. Go to [https://copilotstudio.microsoft.com](https://copilotstudio.microsoft.com)\n1. Select the environment picker at the top right corner\n1. Select the right environment\n1. Select `Create` in the left navigation\n1. Select the blue `New agent` button\n\n ![New agent](./assets/newagent.png)\n\n1. Select `Skip to configure` on the top right\n\n ![Skip to configure](./assets/skiptoconfigure.png)\n\n1. Change the name to `Jokester`\n1. Add the following `Description`\n\n ```text\n A humor-focused agent that delivers concise, engaging jokes only upon user request, adapting its style to match the user's tone and preferences. It remains in character, avoids repetition, and filters out offensive content to ensure a consistently appropriate and witty experience.\n ```\n\n1. Add the following `Instructions`\n\n ```text\n You are a joke-telling assistant. Your sole purpose is to deliver appropriate, clever, and engaging jokes upon request. Follow these rules:\n \n * Respond only when the user asks for a joke or something related (e.g., \"Tell me something funny\").\n * Match the tone and humor preference of the user based on their input—clean, dark, dry, pun-based, dad jokes, etc.\n * Never break character or provide information unrelated to humor.\n * Keep jokes concise and clearly formatted.\n * Avoid offensive, discriminatory, or NSFW content.\n * When unsure about humor preference, default to a clever and universally appropriate joke.\n * Do not repeat jokes within the same session.\n * Avoid explaining the joke unless explicitly asked.\n * Be responsive, witty, and quick.\n ```\n\n1. Select `Create` on the top right\n\n ![Create agent](./assets/createagent.png)\n\n1. Enable Generative AI `Orchestration`\n\n ![Turn on orchestration](./assets/turnonorchestration.png)\n\n1. Disable general knowledge in the `Knowledge` section\n\n ![Turn off general knowledge](./assets/turnoffgeneralknowledge.png)\n\n1. Select `Actions` in the top menu\n \n ![Actions](./assets/actions.png)\n\n1. Select `Add an action`\n\n ![Add an action](./assets/addanaction.png)\n\n1. Search for the name (in this case, `jokes`) of the connector you created [earlier](#create-the-power-platform-connector) (see number 1 in the screenshot below)\n1. Select the `Jokes MCP server` (see number 2 in the screenshot below)\n\n ![Search for and select the action](./assets/addaction.png)\n\n1. Wait for the connection to be created and select `Next` when it's done\n\n ![Action and connection](./assets/actionconnection.png)\n\n1. Change the `Description for the agent to know when to use this action` to the following text:\n\n ```text\n Trigger this action when a user asks for a joke. It can provide Chuck Norris jokes, Dad jokes and Yo Mama jokes.\n ```\n \n Leave the rest as default, like for instance end user authentication, where you will learn more about in a minute.\n\n1. Select `Add action` to add the action to the agent\n\n ![Add action](./assets/addactionend.png)\n\n1. Select the `refresh icon` in the `Test your agent` pane\n\n ![Refresh testing pane](./assets/refreshtestingpane.png)\n\n1. In the `Test your agent` pane send the following message:\n\n ```text\n Can I get a Chuck Norris joke?\n ```\n \n This will show you message that additional permissions are required to run this action. This is because of the user authentication in the action wizard.\n\n1. Select `Connect`\n\n ![Additional permissions](./assets/additionalpermissions.png)\n \n This will open a new window where you can manage your connections for this agent.\n\n1. Select `Connect` next to the `JokesMCP`\n\n ![Connect to JokesMCP](./assets/connect.png)\n\n1. Wait until the connection is created and select `Submit`\n\n ![Pick a connection](./assets/submitconnection.png)\n\n1. The connection should now be connected, so the status should be set to `Connected`\n\n ![Status connected](./assets/connected.png) \n\n1. Close the manage your connections tab in your browser\n\n Now you should be back in the Jokester agent screen.\n\n1. Select the `refresh icon` in the `Test your agent` pane\n\n ![Refresh testing pane](./assets/refreshtestingpane.png)\n\n1. In the `Test your agent` pane send the following message:\n\n ```text\n Can I get a Chuck Norris joke?\n ```\n\n This will now show a Chuck Norris joke - instead of the additional permissions.\n\n ![Chuck Norris joke](./assets/chucknorrisjoke.png)\n\n1. In the `Test your agent` pane send the following message:\n\n ```text\n Can I get a Dad joke?\n ```\n\n This will now show a Dad joke.\n\n ![Dad joke](./assets/dadjoke.png)\n\n1. In the `Test your agent` pane send the following message:\n\n ```text\n Can I get a Yo Mama joke?\n ```\n\n This will now show a Yo Mama joke.\n\n ![Yo Mama joke](./assets/yomamajoke.png)\n\nAnd that was the Jokes MCP Server working in Microsoft Copilot Studio. This is also the end of the lab! Hopefully you liked the lab. Please take the time to fill in our [feedback form](https://aka.ms/mcsmcp/lab/feedback).\n\n## Jokes MCP Server details\n\nThis is a [MCP](https://modelcontextprotocol.io/introduction) server built on the [TypeScript SDK](https://github.com/modelcontextprotocol/typescript-sdk).\n\nWith this MCP Server, you will able to fetch jokes from the following websites:\n- [chucknorris.io](https://api.chucknorris.io/)\n- [icanhazdadjoke.com](https://icanhazdadjoke.com/)\n- [yomama-jokes.com](https://www.yomama-jokes.com)\n\nIf you want to run the server locally, make sure to run `npm install` in the root of the repository.\n\nAfter that you can run `npm run build` to build the server and `npm start` to start the server.\n\n## Tools\n\nThe following tools are included:\n\n### get-chuck-joke\n\nThis tool retrieves a random Chuck Norris Joke from [chucknorris.io](https://api.chucknorris.io/).\n\n### get-chuck-categories\n\nThis tool retrieves the available categories from [chucknorris.io](https://api.chucknorris.io/).\n\n### get-dad-joke\n\nThis tool retrieves a random Dad Joke from [icanhazdadjoke.com](https://icanhazdadjoke.com/).\n\n### get-yo-mama-joke\n\nThis tool retrieves a random Yo Mama Joke from [yomama-jokes.com](https://www.yomama-jokes.com).\n\n## Known issues and planned improvements\n\nThere are some known issues and planned improvements for MCP in Microsoft Copilot Studio. They are listed in [this Microsoft Learn article](https://aka.ms/mcsmcpdocs#known-issues--planned-improvements).\n\n## Contributing\n\nThis project welcomes contributions and suggestions. Most contributions require you to agree to a\nContributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us\nthe rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.\n\nWhen you submit a pull request, a CLA bot will automatically determine whether you need to provide\na CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions\nprovided by the bot. You will only need to do this once across all repos using our CLA.\n\nThis project has adopted the [Microsoft Open Source Code of Conduct](https://opensource.microsoft.com/codeofconduct/).\nFor more information see the [Code of Conduct FAQ](https://opensource.microsoft.com/codeofconduct/faq/) or\ncontact [opencode@microsoft.com](mailto:opencode@microsoft.com) with any additional questions or comments.\n\n## Trademarks\n\nThis project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft \ntrademarks or logos is subject to and must follow \n[Microsoft's Trademark & Brand Guidelines](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general).\nUse of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship.\nAny use of third-party trademarks or logos are subject to those third-party's policies.\n","English","tag","en-US",{"_29":41,"_33":45,"_38":46},{"_31":42,"_33":43,"_35":44},"Un servidor de protocolo de contexto de modelo que permite a Microsoft Copilot Studio obtener chistes de varias fuentes, incluidos chistes de Chuck Norris, chistes de papá y chistes de Yo Mama.","Servidor MCP de chistes","# Microsoft Copilot Studio ❤️ MCP\n\nBienvenido al laboratorio **de Microsoft Copilot Studio ❤️ MCP** . En este laboratorio, aprenderá a implementar un servidor MCP y a agregarlo a Microsoft Copilot Studio.\n\n## ¿Qué es MCP?\n\n[El Protocolo de Contexto de Modelo (MCP)](https://modelcontextprotocol.io/introduction) es un protocolo abierto que estandariza la forma en que las aplicaciones proporcionan contexto a los LLM, definido por [Anthropic](https://www.anthropic.com/) . MCP proporciona una forma estandarizada de conectar modelos de IA a diferentes fuentes de datos y herramientas. MCP permite a los creadores integrar sin problemas los servidores de conocimiento y las API existentes directamente en Copilot Studio.\n\nActualmente, Copilot Studio solo es compatible con Herramientas. Para obtener más información sobre las funciones actuales, consulte [aka.ms/mcsmcp](https://aka.ms/mcsmcp) . Existen algunos problemas conocidos y mejoras planificadas. Se enumeran [aquí](#known-issues-and-planned-improvements) .\n\n## MCP vs. Conectores\n\n¿Cuándo se usa MCP? ¿Y cuándo se usan conectores? ¿Reemplazará el MCP a los conectores?\n\nLos servidores MCP se ponen a disposición de Copilot Studio mediante la infraestructura del conector, por lo que estas preguntas no son realmente aplicables. El hecho de que los servidores MCP utilicen la infraestructura del conector significa que pueden implementar controles de seguridad y gobernanza empresariales, como la integración [de redes virtuales](https://learn.microsoft.com/power-platform/admin/vnet-support-overview) , controles [de prevención de pérdida de datos](https://learn.microsoft.com/power-platform/admin/wp-data-loss-prevention) y [múltiples métodos de autenticación](https://learn.microsoft.com/connectors/custom-connectors/#2-secure-your-api) (todos disponibles en esta versión), a la vez que admiten el acceso a datos en tiempo real para agentes con IA.\n\nEntonces, el MCP y los conectores funcionan realmente **mejor juntos** .\n\n## Prerrequisitos\n\n* Suscripción a Azure (con método de pago agregado)\n* Tener una cuenta de GitHub y estar conectado\n* Cuenta de prueba o de desarrollador de Copilot Studio\n* Entorno de Power Platform aprovisionado\n\n## Laboratorio\n\nPara poder implementar este servidor MCP y usarlo en Microsoft Copilot Studio, debe realizar las siguientes acciones:\n\n* [Cree un nuevo repositorio de GitHub basado en la plantilla](#create-a-new-github-repository-based-on-the-template)\n* [Implementar la aplicación web de Azure](#deploy-the-azure-web-app)\n* [Crear el conector de Power Platform](#create-the-power-platform-connector)\n* [Agregue el servidor MCP como una acción en Microsoft Copilot Studio](#add-the-mcp-server-as-an-action-in-microsoft-copilot-studio)\n\n### Cree un nuevo repositorio de GitHub basado en la plantilla\n\n1. Seleccionar `Use this template`\n\n2. Seleccione \\`Crear un nuevo repositorio\n\n ![](./assets/usetemplate.png)\n\n3. Seleccione el `Owner` adecuado\n\n4. Dale un `Repository name`\n\n5. Opcionalmente puedes darle una `Description`\n\n6. Seleccionar `Private`\n\n7. Seleccione `Create repository`\n\n Esto tardará un poco. Una vez finalizado, se le redirigirá al repositorio recién creado.\n\n> \\[!ADVERTENCIA]\\\n> Después de completar los pasos de este laboratorio, tendrá un servidor MCP ejecutándose en Azure y disponible públicamente. Idealmente, no querrá que esto suceda. Asegúrese de eliminar la aplicación web de Azure después de finalizar el laboratorio.\n\n### Implementar la aplicación web de Azure\n\n1. Vaya a `portal.azure.com`\n\n ![Portal de Azure](/assets/azureportal.png)\n\n2. Coloque el cursor sobre la pestaña `Resource groups` y seleccione `Create`\n\n ![Grupo de recursos](/assets/selectResourceGroup.png)\n\n3. Haga clic en el menú desplegable `Subscription` y seleccione su suscripción de la lista\n\n ![Seleccionar suscripción](/assets/selectSubscription.png)\n\n4. Haga clic en el cuadro de texto Nombre `Resource group` y escriba un nombre de grupo de recursos como \"jokesgrp\".\n\n ![Nombre del grupo de recursos](/assets/resourcegrpname.png)\n\n5. Seleccione el menú desplegable `Region` y elija la región más cercana a usted\n\n ![Región del grupo de recursos](/assets/resourcegrpregion.png)\n\n6. Haga clic en `Review + create`\n\n ![Revisar y crear](/assets/resourcegrpreview.png)\n\n7. Haga clic en `Create`\n\n ![Crear](/assets/resourcegrpcreate.png)\n\n8. Seleccione `Go to resource group`\n\n ![Ir al recurso](/assets/gotoresource.png)\n\n9. Seleccionar `Create`\n\n ![Crear](/assets/inresourcegrpcreate.png)\n\n10. Haga clic en el `search box`\n\n![Buscar](/assets/searchmarketplace.png)\n\n11. Tipo de `web app`\n\n12. Seleccione el `Create button` en el resultado de búsqueda de la aplicación web\n\n![Resultados de la búsqueda](/assets/createwebapp1.png)\n\n13. Seleccionar `Web App`\n\n![Crear aplicación web](/assets/createwebapp2.png)\n\n14. Seleccione la entrada de texto `Name` y coloque un nombre para la aplicación web, como \"wegotjokes\".\n\n![Nombre de la aplicación web](/assets/webappname.png)\n\n15. Seleccione el menú desplegable `Runtime stack` y elija `Node 22 LTS`\n\n![Tiempo de ejecución de la aplicación web](/assets/selectruntime.png)\n\n16. Seleccione el menú desplegable `Region` y elija la región más cercana a usted\n\n![Región](/assets/webappselectregion.png)\n\n17. Haga clic en `Review + create`\n\n![Revisar](/assets/webappreview.png)\n\n18. Haga clic en `Create`\n\n![Crear](/assets/webappcreate.png)\n\n19. Haga clic en `Refresh` para actualizar el progreso de su implementación\n\n![Refrescar](/assets/webapprefresh.png)\n\n20. Una vez completada la implementación, seleccione `Go to resource`\n\n![Ir al recurso](/assets/webappgotoresource.png)\n\n21. Expandir la pestaña `Deployment`\n\n![Pestaña de implementación](/assets/webappdeployment.png)\n\n22. Haga clic en `Deployment Center`\n\n![Centro de despliegue](/assets/webappdeployctr.png)\n\n23. Haga clic en el menú desplegable `Source` y seleccione `GitHub`\n\n![Elija la fuente](/assets/webappsource.png)\n\n24. Seleccionar `Authorize`\n\n![Autorizar](/assets/deployauthorize.png)\n\n25. Seleccione `Continue`\n\n![Continuar](/assets/githubauthcontinue.png)\n\n26. Seleccione `Authorize AzureAppService`\n\n![Autorizar](/assets/githubauthappservice.png)\n\n27. Haga clic en el menú desplegable `Organization` y elija su organización\n\n![Selección de organización](/assets/githuborg.png)\n\n28. Haga clic en el menú desplegable `Repository` y elija el repositorio de GitHub que creó [anteriormente](#create-a-new-github-repository-based-on-the-template) de la lista.\n\n![Selección de repos](/assets/githubrepo.png)\n\n29. Haga clic en el menú desplegable `Branch` y elija su sucursal\n\n![Seleccionar sucursal](/assets/githubbranch.png)\n\n30. Mantenga todo lo demás predeterminado y seleccione `Save`\n\n![Ahorrar](/assets/githubsave.png)\n\n31. Haga clic `Refresh` para actualizar el estado de implementación\n\n![Refrescar](/assets/githubrefresh.png)\n\n32. Verificar que la implementación esté completa\n\n![Implementación realizada](/assets/deploydone.png)\n\n### Crear el conector de Power Platform\n\n1. Vaya a \n\n2. Seleccione `More` en el menú de la izquierda\n\n3. Seleccione el botón `Discover all`\n\n4. En Datos, seleccione el `pin icon` junto a `Custom connectors` para anclarlo al menú de la izquierda\n\n5. Seleccione `Custom connectors` en el menú de la izquierda\n\n6. Seleccionar `New custom connector`\n\n7. Seleccionar `Create from blank`\n\n ![](./assets/newconnector.png)\n\n8. Dale un nombre al conector (por ejemplo `JokesMCP` )\n\n9. Seleccione el botón azul `Continue`\n\n10. Seleccione el botón `Swagger editor`\n\n![](./assets/swaggereditor.png)\n\n11. Copie el código yaml desde [aquí](./assets/connector.yml) y reemplace el código en el editor Swagger\n\n12. En el editor Swagger, reemplaza `dummyurl.azurewebsites.net` con la URL de la aplicación web que creaste [anteriormente](#deploy-the-azure-web-app) . Asegúrate de eliminar `https://` y todo lo que esté después de `azurewebsites.net`\n\n13. Seleccione `Create connector` para crear el conector\n\n### Agregue el servidor MCP como una acción en Microsoft Copilot Studio\n\nAhora tiene un servidor MCP ejecutándose en Azure y un conector disponible en Power Platform. Este paso consiste en crear un agente en Microsoft Copilot Studio y agregarle el servidor MCP.\n\n1. Vaya a \n\n2. Seleccione el selector de entorno en la esquina superior derecha\n\n3. Seleccione el entorno adecuado\n\n4. Seleccione `Create` en la navegación izquierda.\n\n5. Seleccione el botón azul `New agent`\n\n ![Nuevo agente](./assets/newagent.png)\n\n6. Seleccione `Skip to configure` en la parte superior derecha\n\n ![Saltar a la configuración](./assets/skiptoconfigure.png)\n\n7. Cambiar el nombre a `Jokester`\n\n8. Añade la siguiente `Description`\n\n ```text\n A humor-focused agent that delivers concise, engaging jokes only upon user request, adapting its style to match the user's tone and preferences. It remains in character, avoids repetition, and filters out offensive content to ensure a consistently appropriate and witty experience.\n ```\n\n9. Añade las siguientes `Instructions`\n\n ```text\n You are a joke-telling assistant. Your sole purpose is to deliver appropriate, clever, and engaging jokes upon request. Follow these rules:\n\n * Respond only when the user asks for a joke or something related (e.g., \"Tell me something funny\").\n * Match the tone and humor preference of the user based on their input—clean, dark, dry, pun-based, dad jokes, etc.\n * Never break character or provide information unrelated to humor.\n * Keep jokes concise and clearly formatted.\n * Avoid offensive, discriminatory, or NSFW content.\n * When unsure about humor preference, default to a clever and universally appropriate joke.\n * Do not repeat jokes within the same session.\n * Avoid explaining the joke unless explicitly asked.\n * Be responsive, witty, and quick.\n ```\n\n10. Seleccione `Create` en la parte superior derecha\n\n![Crear agente](./assets/createagent.png)\n\n11. Habilitar `Orchestration` de IA generativa\n\n![Activar la orquestación](./assets/turnonorchestration.png)\n\n12. Desactivar conocimientos generales en la sección `Knowledge`\n\n![Desactivar conocimientos generales](./assets/turnoffgeneralknowledge.png)\n\n13. Seleccione `Actions` en el menú superior\n\n![Comportamiento](./assets/actions.png)\n\n14. Seleccionar `Add an action`\n\n![Agregar una acción](./assets/addanaction.png)\n\n15. Busca el nombre (en este caso, `jokes` ) del conector que creaste [anteriormente](#create-the-power-platform-connector) (ver el número 1 en la captura de pantalla a continuación)\n\n16. Seleccione el `Jokes MCP server` (ver el número 2 en la captura de pantalla a continuación)\n\n![Busca y selecciona la acción](./assets/addaction.png)\n\n17. Espere a que se cree la conexión y seleccione `Next` cuando haya terminado.\n\n![Acción y conexión](./assets/actionconnection.png)\n\n18. Cambie la `Description for the agent to know when to use this action` al siguiente texto:\n\n```text\nTrigger this action when a user asks for a joke. It can provide Chuck Norris jokes, Dad jokes and Yo Mama jokes.\n```\n\nDeje el resto como predeterminado, como por ejemplo la autenticación del usuario final, sobre lo cual aprenderá más en un minuto.\n\n19. Seleccione `Add action` para agregar la acción al agente\n\n![Añadir acción](./assets/addactionend.png)\n\n20. Seleccione el `refresh icon` en el panel `Test your agent`\n\n![Actualizar el panel de pruebas](./assets/refreshtestingpane.png)\n\n21. En el panel `Test your agent` envíe el siguiente mensaje:\n\n```text\nCan I get a Chuck Norris joke?\n```\n\nEsto le mostrará un mensaje que indica que se requieren permisos adicionales para ejecutar esta acción. Esto se debe a la autenticación del usuario en el asistente de acciones.\n\n22. Seleccione `Connect`\n\n![Permisos adicionales](./assets/additionalpermissions.png)\n\nEsto abrirá una nueva ventana donde podrás administrar tus conexiones para este agente.\n\n23. Seleccione `Connect` junto a `JokesMCP`\n\n![Conéctate a JokesMCP](./assets/connect.png)\n\n24. Espere hasta que se cree la conexión y seleccione `Submit`\n\n![Elige una conexión](./assets/submitconnection.png)\n\n25. La conexión ahora debería estar conectada, por lo que el estado debería establecerse en `Connected`\n\n![Estado conectado](./assets/connected.png)\n\n26. Cierra la pestaña Administrar tus conexiones en tu navegador\n\nAhora deberías volver a la pantalla del agente Jokester.\n\n27. Seleccione el `refresh icon` en el panel `Test your agent`\n\n![Actualizar el panel de pruebas](./assets/refreshtestingpane.png)\n\n28. En el panel `Test your agent` envíe el siguiente mensaje:\n\n```text\nCan I get a Chuck Norris joke?\n```\n\nAhora se mostrará un chiste de Chuck Norris, en lugar de los permisos adicionales.\n\n![Chiste de Chuck Norris](./assets/chucknorrisjoke.png)\n\n29. En el panel `Test your agent` envíe el siguiente mensaje:\n\n```text\nCan I get a Dad joke?\n```\n\nEsto ahora mostrará un chiste de papá.\n\n![Chiste de papá](./assets/dadjoke.png)\n\n30. En el panel `Test your agent` envíe el siguiente mensaje:\n\n```text\nCan I get a Yo Mama joke?\n```\n\nEsto ahora mostrará un chiste de Yo Mama.\n\n![Chiste de tu mamá](./assets/yomamajoke.png)\n\nY así fue el servidor Jokes MCP funcionando en Microsoft Copilot Studio. ¡Aquí termina el laboratorio! Esperamos que les haya gustado. Por favor, tómense un momento para completar nuestro [formulario de comentarios](https://aka.ms/mcsmcp/lab/feedback) .\n\n## Detalles del servidor MCP de Jokes\n\nEste es un servidor [MCP](https://modelcontextprotocol.io/introduction) creado en el [SDK de TypeScript](https://github.com/modelcontextprotocol/typescript-sdk) .\n\nCon este servidor MCP, podrás obtener chistes de los siguientes sitios web:\n\n* [chucknorris.io](https://api.chucknorris.io/)\n* [icanhazdadjoke.com](https://icanhazdadjoke.com/)\n* [yomama-jokes.com](https://www.yomama-jokes.com)\n\nSi desea ejecutar el servidor localmente, asegúrese de ejecutar `npm install` en la raíz del repositorio.\n\nDespués de eso, puedes ejecutar `npm run build` para construir el servidor y `npm start` para iniciar el servidor.\n\n## Herramientas\n\nSe incluyen las siguientes herramientas:\n\n### chiste de Chuck\n\nEsta herramienta recupera un chiste aleatorio de Chuck Norris de [chucknorris.io](https://api.chucknorris.io/) .\n\n### categorías de get-chuck\n\nEsta herramienta recupera las categorías disponibles de [chucknorris.io](https://api.chucknorris.io/) .\n\n### chiste de papá\n\nEsta herramienta recupera un chiste de papá aleatorio de [icanhazdadjoke.com](https://icanhazdadjoke.com/) .\n\n### chiste para conseguir a tu mamá\n\nEsta herramienta recupera un chiste de Yo Mama aleatorio de [yomama-jokes.com](https://www.yomama-jokes.com) .\n\n## Problemas conocidos y mejoras planificadas\n\nExisten algunos problemas conocidos y mejoras planificadas para MCP en Microsoft Copilot Studio. Se enumeran en [este artículo de Microsoft Learn](https://aka.ms/mcsmcpdocs#known-issues--planned-improvements) .\n\n## Contribuyendo\n\nEste proyecto agradece contribuciones y sugerencias. La mayoría de las contribuciones requieren la aceptación de un Acuerdo de Licencia de Colaborador (CLA), que declara que tiene derecho a otorgarnos, y efectivamente nos otorga, los derechos para usar su contribución. Para más información, visite .\n\nAl enviar una solicitud de extracción, un bot de CLA determinará automáticamente si es necesario proporcionar una CLA y la decorará adecuadamente (por ejemplo, comprobación de estado o comentario). Simplemente siga las instrucciones del bot. Solo tendrá que hacerlo una vez en todos los repositorios que utilicen nuestra CLA.\n\nEste proyecto ha adoptado el [Código de Conducta de Código Abierto de Microsoft](https://opensource.microsoft.com/codeofconduct/) . Para más información, consulte las [preguntas frecuentes sobre el Código de Conducta](https://opensource.microsoft.com/codeofconduct/faq/) o escriba [a opencode@microsoft.com](mailto:opencode@microsoft.com) si tiene alguna pregunta o comentario.\n\n## Marcas comerciales\n\nEste proyecto puede contener marcas comerciales o logotipos de proyectos, productos o servicios. El uso autorizado de las marcas comerciales o logotipos de Microsoft está sujeto a [las Directrices de Marcas Registradas y de Marca de Microsoft](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general) y debe cumplirlas. El uso de marcas comerciales o logotipos de Microsoft en versiones modificadas de este proyecto no debe causar confusión ni implicar patrocinio por parte de Microsoft. El uso de marcas comerciales o logotipos de terceros está sujeto a las políticas de dichos terceros.","Spanish","es-ES",{"_29":48,"_33":52,"_38":53},{"_31":49,"_33":50,"_35":51},"Microsoft Copilot Studio がチャック・ノリスジョーク、お父さんジョーク、ヨー・ママジョークなど、さまざまなソースからジョークを取得できるようにするモデルコンテキストプロトコルサーバー。","ジョークMCPサーバー","# Microsoft Copilot Studio ❤️ MCP\n\n**Microsoft Copilot Studio ❤️ MCP**ラボへようこそ。このラボでは、MCPサーバーの展開方法と、Microsoft Copilot Studioへの追加方法を学習します。\n\n## MCPとは何ですか?\n\n[モデルコンテキストプロトコル（MCP）](https://modelcontextprotocol.io/introduction)は、 [Anthropic](https://www.anthropic.com/)によって定義された、アプリケーションがLLMにコンテキストを提供する方法を標準化するオープンプロトコルです。MCPは、AIモデルをさまざまなデータソースやツールに接続するための標準化された方法を提供します。MCPにより、開発者は既存のナレッジサーバーとAPIをCopilot Studioに直接シームレスに統合できます。\n\n現在、Copilot Studio はツールのみをサポートしています。現在の機能の詳細については、 [aka.ms/mcsmcp](https://aka.ms/mcsmcp)をご覧ください。既知の問題と計画されている改善点がいくつかあります。[こちらに](#known-issues-and-planned-improvements)一覧を掲載しています。\n\n## MCPとコネクタ\n\nMCPはいつ使用しますか？コネクタはいつ使用しますか？MCPはコネクタに取って代わるのでしょうか？\n\nMCPサーバーはコネクタインフラストラクチャを介してCopilot Studioで利用可能になるため、これらの質問は実際には当てはまりません。MCPサーバーがコネクタインフラストラクチャを使用しているという事実は、[仮想ネットワーク](https://learn.microsoft.com/power-platform/admin/vnet-support-overview)統合、[データ損失防止](https://learn.microsoft.com/power-platform/admin/wp-data-loss-prevention)（DLP）、[複数の認証方法](https://learn.microsoft.com/connectors/custom-connectors/#2-secure-your-api)といったエンタープライズセキュリティとガバナンス制御（これらはすべて本リリースで利用可能）を採用しながら、AI搭載エージェントによるリアルタイムデータアクセスをサポートできることを意味します。\n\nしたがって、MCP とコネクタ**を組み合わせると非常に効果的**です。\n\n## 前提条件\n\n* Azure サブスクリプション (支払い方法を追加)\n* GitHubアカウントを持っていてログインしている\n* Copilot Studio のトライアルまたは開発者アカウント\n* Power Platform 環境がプロビジョニングされました\n\n## ラボ\n\nこの MCP サーバーを展開して Microsoft Copilot Studio で使用するには、次の操作を実行する必要があります。\n\n* [テンプレートに基づいて新しいGitHubリポジトリを作成する](#create-a-new-github-repository-based-on-the-template)\n* [Azure Web アプリをデプロイする](#deploy-the-azure-web-app)\n* [Power Platform コネクタを作成する](#create-the-power-platform-connector)\n* [Microsoft Copilot Studio で MCP サーバーをアクションとして追加します。](#add-the-mcp-server-as-an-action-in-microsoft-copilot-studio)\n\n### テンプレートに基づいて新しいGitHubリポジトリを作成する\n\n1. `Use this template`選択\n\n2. 「新しいリポジトリを作成」を選択します\n\n ![](./assets/usetemplate.png)\n\n3. 適切な`Owner`を選択する\n\n4. `Repository name`を付ける\n\n5. オプションで`Description`を記入できます\n\n6. `Private`を選択\n\n7. `Create repository`選択\n\n 少し時間がかかります。完了すると、新しく作成されたリポジトリに移動します。\n\n> \\[！警告]\\\n> このラボの手順を完了すると、Azure 上で実行され、公開されている MCP サーバーが作成されます。理想的には、これは不要なため、ラボ終了後は Azure Web アプリを必ず削除してください。\n\n### Azure Web アプリをデプロイする\n\n1. `portal.azure.com`にアクセスしてください\n\n ![Azureポータル](/assets/azureportal.png)\n\n2. `Resource groups`タブにマウスを移動し、 `Create`を選択します。\n\n ![リソースグループ](/assets/selectResourceGroup.png)\n\n3. `Subscription`ドロップダウンをクリックし、リストからサブスクリプションを選択します\n\n ![サブスクリプションを選択](/assets/selectSubscription.png)\n\n4. `Resource group`名テキストボックスをクリックし、「jokesgrp」などのリソースグループ名を入力します。\n\n ![リソースグループ名](/assets/resourcegrpname.png)\n\n5. `Region`ドロップダウンから最も近い地域を選択してください\n\n ![リソースグループリージョン](/assets/resourcegrpregion.png)\n\n6. `Review + create`をクリック\n\n ![レビューと作成](/assets/resourcegrpreview.png)\n\n7. `Create`クリック\n\n ![作成する](/assets/resourcegrpcreate.png)\n\n8. `Go to resource group`選択\n\n ![リソースへ移動](/assets/gotoresource.png)\n\n9. `Create`選択\n\n ![作成する](/assets/inresourcegrpcreate.png)\n\n10. `search box`をクリック\n\n![検索](/assets/searchmarketplace.png)\n\n11. `web app`を入力\n\n12. Webアプリの検索結果で`Create button`を選択します\n\n![検索結果](/assets/createwebapp1.png)\n\n13. `Web App`を選択\n\n![ウェブアプリを作成する](/assets/createwebapp2.png)\n\n14. `Name`テキスト入力を選択し、「wegotjokes」などのWebアプリの名前を入力します。\n\n![ウェブアプリ名](/assets/webappname.png)\n\n15. `Runtime stack`ドロップダウンを選択し、 `Node 22 LTS`を選択します。\n\n![Webアプリランタイム](/assets/selectruntime.png)\n\n16. `Region`ドロップダウンを選択し、最も近い地域を選択してください\n\n![地域](/assets/webappselectregion.png)\n\n17. `Review + create`をクリック\n\n![レビュー](/assets/webappreview.png)\n\n18. `Create`クリック\n\n![作成する](/assets/webappcreate.png)\n\n19. 展開の進行状況を更新するには、 `Refresh`クリックします。\n\n![リフレッシュ](/assets/webapprefresh.png)\n\n20. 展開が完了したら、 `Go to resource`を選択します。\n\n![リソースへ移動](/assets/webappgotoresource.png)\n\n21. `Deployment`タブを展開します\n\n![展開タブ](/assets/webappdeployment.png)\n\n22. `Deployment Center`をクリック\n\n![展開センター](/assets/webappdeployctr.png)\n\n23. `Source`ドロップダウンをクリックして`GitHub`を選択します\n\n![ソースを選択](/assets/webappsource.png)\n\n24. `Authorize`を選択\n\n![承認](/assets/deployauthorize.png)\n\n25. `Continue`選択\n\n![続く](/assets/githubauthcontinue.png)\n\n26. `Authorize AzureAppService`選択\n\n![承認](/assets/githubauthappservice.png)\n\n27. `Organization`ドロップダウンをクリックして組織を選択します\n\n![組織の選択](/assets/githuborg.png)\n\n28. `Repository`ドロップダウンをクリックし、リストから[先ほど](#create-a-new-github-repository-based-on-the-template)作成したGitHubリポジトリを選択します。\n\n![レポ選択](/assets/githubrepo.png)\n\n29. `Branch`ドロップダウンをクリックしてブランチを選択します\n\n![ブランチを選択](/assets/githubbranch.png)\n\n30. その他はすべてデフォルトのままにして`Save`を選択します\n\n![保存](/assets/githubsave.png)\n\n31. 展開ステータスを更新するには、 `Refresh`クリックします。\n\n![リフレッシュ](/assets/githubrefresh.png)\n\n32. 展開が完了したことを確認する\n\n![デプロイ完了](/assets/deploydone.png)\n\n### Power Platform コネクタを作成する\n\n1. にアクセスしてください。\n\n2. 左側のメニューで`More`を選択します\n\n3. `Discover all`ボタンを選択します\n\n4. 「データ」の下で、 `Custom connectors`の横にある`pin icon`を選択して、左側のメニューにピン留めします。\n\n5. 左側のメニューで`Custom connectors`を選択します\n\n6. `New custom connector`を選択\n\n7. `Create from blank`を選択\n\n ![](./assets/newconnector.png)\n\n8. コネクタに名前を付けます（例： `JokesMCP` ）\n\n9. 青い「 `Continue`ボタンを選択します\n\n10. `Swagger editor`トグルを選択する\n\n![](./assets/swaggereditor.png)\n\n11. [ここ](./assets/connector.yml)からyamlコードをコピーし、Swaggerエディタのコードを置き換えます\n\n12. Swaggerエディターで、 `dummyurl.azurewebsites.net`[先ほど](#deploy-the-azure-web-app)作成したWebアプリのURLに置き換えます。https `https://`と`azurewebsites.net`以降の部分をすべて削除してください。\n\n13. `Create connector`を選択します。\n\n### Microsoft Copilot Studio で MCP サーバーをアクションとして追加します。\n\nこれで、Azure で MCP サーバーが稼働し、Power Platform でコネクタが利用可能になりました。この手順では、Microsoft Copilot Studio でエージェントを作成し、MCP サーバーをエージェントに追加します。\n\n1. にアクセスしてください\n\n2. 右上隅の環境ピッカーを選択します\n\n3. 適切な環境を選択する\n\n4. 左側のナビゲーションで`Create`を選択します\n\n5. 青い`New agent`ボタンを選択します\n\n ![新しいエージェント](./assets/newagent.png)\n\n6. 右上の`Skip to configure`を選択します\n\n ![設定へスキップ](./assets/skiptoconfigure.png)\n\n7. 名前を`Jokester`に変更する\n\n8. 次の`Description`を追加します\n\n ```text\n A humor-focused agent that delivers concise, engaging jokes only upon user request, adapting its style to match the user's tone and preferences. It remains in character, avoids repetition, and filters out offensive content to ensure a consistently appropriate and witty experience.\n ```\n\n9. 以下の`Instructions`を追加します\n\n ```text\n You are a joke-telling assistant. Your sole purpose is to deliver appropriate, clever, and engaging jokes upon request. Follow these rules:\n\n * Respond only when the user asks for a joke or something related (e.g., \"Tell me something funny\").\n * Match the tone and humor preference of the user based on their input—clean, dark, dry, pun-based, dad jokes, etc.\n * Never break character or provide information unrelated to humor.\n * Keep jokes concise and clearly formatted.\n * Avoid offensive, discriminatory, or NSFW content.\n * When unsure about humor preference, default to a clever and universally appropriate joke.\n * Do not repeat jokes within the same session.\n * Avoid explaining the joke unless explicitly asked.\n * Be responsive, witty, and quick.\n ```\n\n10. 右上の`Create`を選択\n\n![エージェントを作成](./assets/createagent.png)\n\n11. 生成AI `Orchestration`有効にする\n\n![オーケストレーションをオンにする](./assets/turnonorchestration.png)\n\n12. `Knowledge`セクションで一般知識を無効にする\n\n![一般知識をオフにする](./assets/turnoffgeneralknowledge.png)\n\n13. 上部メニューの`Actions`を選択します\n\n![アクション](./assets/actions.png)\n\n14. `Add an action`選択\n\n![アクションを追加する](./assets/addanaction.png)\n\n15. [先ほど](#create-the-power-platform-connector)作成したコネクタの名前（この場合は`jokes` ）を検索します（下のスクリーンショットの番号 1 を参照）。\n\n16. `Jokes MCP server`を選択します（下のスクリーンショットの番号2を参照）\n\n![アクションを検索して選択する](./assets/addaction.png)\n\n17. 接続が確立されるまで待ち、完了したら`Next`を選択します。\n\n![行動とつながり](./assets/actionconnection.png)\n\n18. `Description for the agent to know when to use this action`次のテキストに変更します。\n\n```text\nTrigger this action when a user asks for a joke. It can provide Chuck Norris jokes, Dad jokes and Yo Mama jokes.\n```\n\n残りはデフォルトのままにしておきます。たとえば、エンドユーザー認証については、後ほど詳しく説明します。\n\n19. `Add action`を選択して、エージェントにアクションを追加します\n\n![アクションを追加](./assets/addactionend.png)\n\n20. `Test your agent`ペインで`refresh icon`を選択します\n\n![テストパネルを更新](./assets/refreshtestingpane.png)\n\n21. `Test your agent`ペインで次のメッセージを送信します。\n\n```text\nCan I get a Chuck Norris joke?\n```\n\nこのアクションを実行するには追加の権限が必要であることを示すメッセージが表示されます。これは、アクションウィザードでのユーザー認証のためです。\n\n22. `Connect`を選択\n\n![追加の権限](./assets/additionalpermissions.png)\n\nこれにより、このエージェントの接続を管理できる新しいウィンドウが開きます。\n\n23. `JokesMCP`横にある`Connect`を選択します\n\n![JokesMCPに接続する](./assets/connect.png)\n\n24. 接続が作成されるまで待ってから、 `Submit`を選択します。\n\n![接続を選択](./assets/submitconnection.png)\n\n25. 接続は確立されたので、ステータスは`Connected`に設定されます。\n\n![ステータス接続](./assets/connected.png)\n\n26. ブラウザの接続管理タブを閉じます\n\nこれで、Jokester エージェント画面に戻るはずです。\n\n27. `Test your agent`ペインで`refresh icon`を選択します\n\n![テストパネルを更新](./assets/refreshtestingpane.png)\n\n28. `Test your agent`ペインで次のメッセージを送信します。\n\n```text\nCan I get a Chuck Norris joke?\n```\n\nこれにより、追加の権限の代わりに、チャック・ノリスのジョークが表示されます。\n\n![チャック・ノリスのジョーク](./assets/chucknorrisjoke.png)\n\n29. `Test your agent`ペインで次のメッセージを送信します。\n\n```text\nCan I get a Dad joke?\n```\n\nこれでお父さんジョークが表示されます。\n\n![お父さんのジョーク](./assets/dadjoke.png)\n\n30. `Test your agent`ペインで次のメッセージを送信します。\n\n```text\nCan I get a Yo Mama joke?\n```\n\nこれでYo Mamaのジョークが表示されます。\n\n![ヨーママジョーク](./assets/yomamajoke.png)\n\n以上で、Microsoft Copilot Studio で動作する Jokes MCP Server のご紹介です。ラボもこれで終了です！ラボを気に入っていただけたでしょうか？[フィードバックフォーム](https://aka.ms/mcsmcp/lab/feedback)へのご記入をお待ちしております。\n\n## ジョークMCPサーバーの詳細\n\nこれは[TypeScript SDK](https://github.com/modelcontextprotocol/typescript-sdk)上に構築された[MCP](https://modelcontextprotocol.io/introduction)サーバーです。\n\nこの MCP サーバーを使用すると、次の Web サイトからジョークを取得できるようになります。\n\n* [chucknorris.io](https://api.chucknorris.io/)\n* [icanhazdadjoke.com](https://icanhazdadjoke.com/)\n* [yomama-jokes.com](https://www.yomama-jokes.com)\n\nサーバーをローカルで実行する場合は、リポジトリのルートで`npm install`を実行してください。\n\nその後、 `npm run build`実行してサーバーをビルドし、 `npm start`実行してサーバーを起動できます。\n\n## ツール\n\n以下のツールが含まれています:\n\n### ゲットチャックジョーク\n\nこのツールは、 [chucknorris.io](https://api.chucknorris.io/)からランダムにチャック・ノリスのジョークを取得します。\n\n### チャックカテゴリーを取得する\n\nこのツールは、 [chucknorris.io](https://api.chucknorris.io/)から利用可能なカテゴリを取得します。\n\n### お父さんのジョーク\n\nこのツールは、 [icanhazdadjoke.com](https://icanhazdadjoke.com/)からランダムな Dad Joke を取得します。\n\n### ママを騙すジョーク\n\nこのツールは、 [yomama-jokes.com](https://www.yomama-jokes.com)からランダムな Yo Mama Joke を取得します。\n\n## 既知の問題と計画されている改善\n\nMicrosoft Copilot Studio の MCP には、既知の問題と改善予定がいくつかあります。これらの問題は[、こちらの Microsoft Learn の記事](https://aka.ms/mcsmcpdocs#known-issues--planned-improvements)に記載されています。\n\n## 貢献\n\nこのプロジェクトでは、貢献と提案を歓迎しています。ほとんどの貢献には、貢献者ライセンス契約（CLA）への同意が必要です。CLAは、貢献者がその貢献を利用する権利を有し、実際にその権利をMicrosoftに付与することを宣言するものです。詳細については、をご覧ください。\n\nプルリクエストを送信すると、CLAボットがCLAが必要かどうかを自動的に判断し、適切な装飾（ステータス確認、コメントなど）をPRに施します。ボットの指示に従うだけで、CLAを使用しているすべてのリポジトリでこの操作を一度だけ実行すれば完了です。\n\nこのプロジェクトは[、Microsoft オープンソース行動規範](https://opensource.microsoft.com/codeofconduct/)を採用しています。詳細については[、行動規範に関する FAQ を](https://opensource.microsoft.com/codeofconduct/faq/)ご覧ください。ご質問やご意見がございましたら、までお問い合わせください。\n\n## 商標\n\nこのプロジェクトには、プロジェクト、製品、またはサービスの商標またはロゴが含まれている場合があります。Microsoftの商標またはロゴの使用は、 [Microsoftの商標およびブランドガイドラインの](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general)対象となり、これに従う必要があります。このプロジェクトの改変版におけるMicrosoftの商標またはロゴの使用は、混乱を招いたり、Microsoftのスポンサーシップを暗示したりしてはなりません。第三者の商標またはロゴの使用は、当該第三者のポリシーに従うものとします。","Japanese","ja-JP",{"_29":55,"_33":59,"_38":60},{"_31":56,"_33":57,"_35":58},"척 노리스 농담, 아빠 농담, 요마마 농담 등 다양한 출처에서 농담을 가져올 수 있는 Microsoft Copilot Studio를 구현하는 모델 컨텍스트 프로토콜 서버입니다.","농담 MCP 서버","# Microsoft Copilot Studio ❤️ MCP\n\n**Microsoft Copilot Studio ❤️ MCP** 랩에 오신 것을 환영합니다. 이 랩에서는 MCP 서버를 배포하고 Microsoft Copilot Studio에 추가하는 방법을 알아봅니다.\n\n## MCP란 무엇인가요?\n\n[모델 컨텍스트 프로토콜(MCP)](https://modelcontextprotocol.io/introduction) 은 [Anthropic](https://www.anthropic.com/) 에서 정의한 애플리케이션이 LLM에 컨텍스트를 제공하는 방식을 표준화하는 개방형 프로토콜입니다. MCP는 AI 모델을 다양한 데이터 소스 및 도구에 연결하는 표준화된 방식을 제공합니다. MCP를 통해 제작자는 기존 지식 서버와 API를 Copilot Studio에 직접 원활하게 통합할 수 있습니다.\n\n현재 Copilot Studio는 도구만 지원합니다. 현재 기능에 대한 자세한 내용은 [aka.ms/mcsmcp를](https://aka.ms/mcsmcp) 참조하세요. 몇 가지 알려진 문제와 계획된 개선 사항이 있으며, 자세한 내용은 [여기를](#known-issues-and-planned-improvements) 참조하세요.\n\n## MCP 대 커넥터\n\nMCP는 언제 사용하시나요? 커넥터는 언제 사용하시나요? MCP가 커넥터를 대체할까요?\n\nMCP 서버는 커넥터 인프라를 통해 Copilot Studio에서 사용할 수 있으므로 이러한 질문은 실제로 적용되지 않습니다. MCP 서버가 커넥터 인프라를 사용한다는 것은 [가상 네트워크](https://learn.microsoft.com/power-platform/admin/vnet-support-overview) 통합, [데이터 유출 방지(DLP)](https://learn.microsoft.com/power-platform/admin/wp-data-loss-prevention) 제어, [다중 인증 방법](https://learn.microsoft.com/connectors/custom-connectors/#2-secure-your-api) 등 엔터프라이즈 보안 및 거버넌스 제어 기능을 사용할 수 있다는 것을 의미하며, 이러한 모든 기능은 이번 릴리스에서 제공되며, AI 기반 에이전트의 실시간 데이터 액세스를 지원합니다.\n\n그러니까 MCP와 커넥터는 **함께 쓰면 정말 더 좋습니다** .\n\n## 필수 조건\n\n* Azure 구독(결제 방법 추가됨)\n* GitHub 계정이 있고 로그인되어 있어야 합니다.\n* Copilot Studio 평가판 또는 개발자 계정\n* Power Platform 환경이 프로비저닝됨\n\n## 랩\n\n이 MCP 서버를 배포하고 Microsoft Copilot Studio에서 사용하려면 다음 작업을 수행해야 합니다.\n\n* [템플릿을 기반으로 새로운 GitHub 저장소를 만듭니다.](#create-a-new-github-repository-based-on-the-template)\n* [Azure 웹 앱 배포](#deploy-the-azure-web-app)\n* [Power Platform 커넥터 만들기](#create-the-power-platform-connector)\n* [Microsoft Copilot Studio에서 MCP 서버를 작업으로 추가합니다.](#add-the-mcp-server-as-an-action-in-microsoft-copilot-studio)\n\n### 템플릿을 기반으로 새로운 GitHub 저장소를 만듭니다.\n\n1. `Use this template` 선택하세요\n\n2. \\`새 저장소 만들기\\`를 선택하세요\n\n ![](./assets/usetemplate.png)\n\n3. 올바른 `Owner` 선택하세요\n\n4. `Repository name` 지정하세요\n\n5. 선택적으로 `Description` 입력할 수 있습니다.\n\n6. `Private` 선택\n\n7. `Create repository` 선택하세요\n\n 이 작업에는 약간의 시간이 소요됩니다. 완료되면 새로 생성된 저장소로 이동하게 됩니다.\n\n> \\[!경고]\\\n> 이 랩의 단계를 완료하면 Azure에서 공개적으로 사용 가능한 MCP 서버가 생성됩니다. 이상적으로는 이러한 서버는 필요하지 않습니다. 랩을 완료한 후에는 Azure 웹앱을 삭제해야 합니다.\n\n### Azure 웹 앱 배포\n\n1. `portal.azure.com` 으로 이동\n\n ![Azure 포털](/assets/azureportal.png)\n\n2. `Resource groups` 탭 위에 마우스를 올려놓고 `Create` 선택하세요.\n\n ![리소스 그룹](/assets/selectResourceGroup.png)\n\n3. `Subscription` 드롭다운을 클릭하고 목록에서 구독을 선택하세요\n\n ![구독 선택](/assets/selectSubscription.png)\n\n4. `Resource group` 이름 텍스트 상자를 클릭하고 \"jokesgrp\"와 같은 리소스 그룹 이름을 입력합니다.\n\n ![리소스 그룹 이름](/assets/resourcegrpname.png)\n\n5. `Region` 드롭다운을 선택하여 가장 가까운 지역을 선택하세요\n\n ![리소스 그룹 지역](/assets/resourcegrpregion.png)\n\n6. `Review + create` 클릭하세요\n\n ![검토하고 생성하세요](/assets/resourcegrpreview.png)\n\n7. `Create` 클릭하세요\n\n ![만들다](/assets/resourcegrpcreate.png)\n\n8. `Go to resource group` 선택하세요\n\n ![리소스로 이동](/assets/gotoresource.png)\n\n9. `Create` 선택하세요\n\n ![만들다](/assets/inresourcegrpcreate.png)\n\n10. `search box` 을 클릭하세요\n\n![찾다](/assets/searchmarketplace.png)\n\n11. `web app` 입력\n\n12. 웹 앱 검색 결과에서 `Create button` 선택하세요\n\n![검색 결과](/assets/createwebapp1.png)\n\n13. `Web App` 선택\n\n![웹 앱 만들기](/assets/createwebapp2.png)\n\n14. `Name` 텍스트 입력을 선택하고 \"wegotjokes\"와 같이 웹 앱의 이름을 입력합니다.\n\n![웹 앱 이름](/assets/webappname.png)\n\n15. `Runtime stack` 드롭다운을 선택하고 `Node 22 LTS` 선택하세요.\n\n![웹 앱 런타임](/assets/selectruntime.png)\n\n16. `Region` 드롭다운을 선택하고 가장 가까운 지역을 선택하세요\n\n![지역](/assets/webappselectregion.png)\n\n17. `Review + create` 클릭하세요\n\n![검토](/assets/webappreview.png)\n\n18. `Create` 클릭하세요\n\n![만들다](/assets/webappcreate.png)\n\n19. 배포 진행 상황을 업데이트하려면 `Refresh` 클릭하세요.\n\n![새로 고치다](/assets/webapprefresh.png)\n\n20. 배포가 완료되면 `Go to resource` 선택하세요.\n\n![리소스로 이동](/assets/webappgotoresource.png)\n\n21. `Deployment` 탭을 확장합니다.\n\n![배포 탭](/assets/webappdeployment.png)\n\n22. `Deployment Center` 클릭하세요\n\n![배치 센터](/assets/webappdeployctr.png)\n\n23. `Source` 드롭다운을 클릭하고 `GitHub` 선택하세요\n\n![출처 선택](/assets/webappsource.png)\n\n24. `Authorize` 선택하세요\n\n![승인하다](/assets/deployauthorize.png)\n\n25. `Continue` 선택하세요\n\n![계속하다](/assets/githubauthcontinue.png)\n\n26. `Authorize AzureAppService` 선택하세요\n\n![승인하다](/assets/githubauthappservice.png)\n\n27. `Organization` 드롭다운을 클릭하고 조직을 선택하세요\n\n![조직 선택](/assets/githuborg.png)\n\n28. `Repository` 드롭다운을 클릭하고 목록에서 [앞서](#create-a-new-github-repository-based-on-the-template) 생성한 GitHub 저장소를 선택하세요.\n\n![레포 선택](/assets/githubrepo.png)\n\n29. `Branch` 드롭다운을 클릭하고 지점을 선택하세요\n\n![지점 선택](/assets/githubbranch.png)\n\n30. 나머지는 기본값으로 두고 `Save` 선택하세요.\n\n![구하다](/assets/githubsave.png)\n\n31. 배포 상태를 업데이트하려면 `Refresh` 클릭하세요.\n\n![새로 고치다](/assets/githubrefresh.png)\n\n32. 배포가 완료되었는지 확인하세요\n\n![배포 완료](/assets/deploydone.png)\n\n### Power Platform 커넥터 만들기\n\n1. 으로 이동하세요\n\n2. 왼쪽 메뉴에서 `More` 선택하세요\n\n3. `Discover all` 버튼을 선택하세요\n\n4. 데이터에서 `Custom connectors` 옆에 있는 `pin icon` 선택하여 왼쪽 메뉴에 고정합니다.\n\n5. 왼쪽 메뉴에서 `Custom connectors` 선택하세요\n\n6. `New custom connector` 선택\n\n7. `Create from blank` 선택하세요\n\n ![](./assets/newconnector.png)\n\n8. 커넥터에 이름을 지정합니다(예: `JokesMCP` )\n\n9. 파란색 `Continue` 버튼을 선택하세요\n\n10. `Swagger editor` 토글을 선택하세요\n\n![](./assets/swaggereditor.png)\n\n11. [여기](./assets/connector.yml) 에서 yaml 코드를 복사하고 Swagger 편집기에서 코드를 바꾸세요.\n\n12. Swagger 편집기에서 `dummyurl.azurewebsites.net`[앞서](#deploy-the-azure-web-app) 만든 웹앱의 URL로 바꾸세요. `https://` 와 `azurewebsites.net` 뒤의 모든 부분을 삭제하세요.\n\n13. 커넥터를 생성하려면 `Create connector` 선택하세요.\n\n### Microsoft Copilot Studio에서 MCP 서버를 작업으로 추가합니다.\n\n이제 Azure에서 MCP 서버를 실행하고 Power Platform에서 커넥터를 사용할 수 있습니다. 이 단계에서는 Microsoft Copilot Studio에서 에이전트를 만들고 에이전트에 MCP 서버를 추가하는 방법을 설명합니다.\n\n1. 으로 이동하세요\n\n2. 오른쪽 상단 모서리에 있는 환경 선택기를 선택하세요\n\n3. 올바른 환경을 선택하세요\n\n4. 왼쪽 탐색에서 `Create` 선택하세요\n\n5. 파란색 `New agent` 버튼을 선택하세요\n\n ![새로운 에이전트](./assets/newagent.png)\n\n6. 오른쪽 상단에서 `Skip to configure` 선택하세요\n\n ![구성으로 건너뛰기](./assets/skiptoconfigure.png)\n\n7. 이름을 `Jokester` 로 변경하세요\n\n8. 다음 `Description` 추가하세요\n\n 지엑스피1\n\n9. 다음 `Instructions` 추가하세요\n\n ```text\n You are a joke-telling assistant. Your sole purpose is to deliver appropriate, clever, and engaging jokes upon request. Follow these rules:\n\n * Respond only when the user asks for a joke or something related (e.g., \"Tell me something funny\").\n * Match the tone and humor preference of the user based on their input—clean, dark, dry, pun-based, dad jokes, etc.\n * Never break character or provide information unrelated to humor.\n * Keep jokes concise and clearly formatted.\n * Avoid offensive, discriminatory, or NSFW content.\n * When unsure about humor preference, default to a clever and universally appropriate joke.\n * Do not repeat jokes within the same session.\n * Avoid explaining the joke unless explicitly asked.\n * Be responsive, witty, and quick.\n ```\n\n10. 오른쪽 상단에서 `Create` 선택하세요\n\n![에이전트 생성](./assets/createagent.png)\n\n11. 생성적 AI `Orchestration` 활성화\n\n![오케스트레이션 켜기](./assets/turnonorchestration.png)\n\n12. `Knowledge` 섹션에서 일반 지식 비활성화\n\n![일반 지식 끄기](./assets/turnoffgeneralknowledge.png)\n\n13. 상단 메뉴에서 `Actions` 선택하세요\n\n![행위](./assets/actions.png)\n\n14. `Add an action` 선택하세요\n\n![동작 추가](./assets/addanaction.png)\n\n15. [이전에](#create-the-power-platform-connector) 생성한 커넥터의 이름(이 경우 `jokes` )을 검색합니다(아래 스크린샷의 번호 1 참조).\n\n16. `Jokes MCP server` 선택하세요(아래 스크린샷의 2번 참조)\n\n![작업을 검색하고 선택하세요](./assets/addaction.png)\n\n17. 연결이 생성될 때까지 기다린 후 완료되면 `Next` 선택하세요.\n\n![행동과 연결](./assets/actionconnection.png)\n\n18. `Description for the agent to know when to use this action` 다음 텍스트로 변경하세요.\n\n```text\nTrigger this action when a user asks for a joke. It can provide Chuck Norris jokes, Dad jokes and Yo Mama jokes.\n```\n\n나머지는 기본값으로 두세요. 예를 들어 최종 사용자 인증과 같은 항목은 잠시 후에 자세히 알아보겠습니다.\n\n19. 에이전트에 작업을 추가하려면 `Add action` 선택하세요.\n\n![작업 추가](./assets/addactionend.png)\n\n20. `Test your agent` 창에서 `refresh icon` 선택하세요.\n\n![테스트 창 새로 고침](./assets/refreshtestingpane.png)\n\n21. `Test your agent` 창에서 다음 메시지를 보내세요.\n\n```text\nCan I get a Chuck Norris joke?\n```\n\n이 작업을 실행하려면 추가 권한이 필요하다는 메시지가 표시됩니다. 이는 작업 마법사의 사용자 인증 때문입니다.\n\n22. `Connect` 선택하세요\n\n![추가 권한](./assets/additionalpermissions.png)\n\n이렇게 하면 해당 에이전트에 대한 연결을 관리할 수 있는 새 창이 열립니다.\n\n23. `JokesMCP` 옆에 있는 `Connect` 선택하세요\n\n![JokesMCP에 연결](./assets/connect.png)\n\n24. 연결이 생성될 때까지 기다린 후 `Submit` 선택하세요.\n\n![연결을 선택하세요](./assets/submitconnection.png)\n\n25. 이제 연결이 연결되었으므로 상태가 `Connected` 으로 설정되어야 합니다.\n\n![상태 연결됨](./assets/connected.png)\n\n26. 브라우저에서 연결 관리 탭을 닫습니다.\n\n이제 Jokester 에이전트 화면으로 돌아왔을 것입니다.\n\n27. `Test your agent` 창에서 `refresh icon` 선택하세요.\n\n![테스트 창 새로 고침](./assets/refreshtestingpane.png)\n\n28. `Test your agent` 창에서 다음 메시지를 보내세요.\n\n```text\nCan I get a Chuck Norris joke?\n```\n\n이제 추가 권한 대신 척 노리스 농담이 표시됩니다.\n\n![척 노리스 농담](./assets/chucknorrisjoke.png)\n\n29. `Test your agent` 창에서 다음 메시지를 보내세요.\n\n```text\nCan I get a Dad joke?\n```\n\n이제 아빠 농담이 표시됩니다.\n\n![아빠 농담](./assets/dadjoke.png)\n\n30. `Test your agent` 창에서 다음 메시지를 보내세요.\n\n```text\nCan I get a Yo Mama joke?\n```\n\n이제 요 마마 농담이 표시됩니다.\n\n![요 마마 농담](./assets/yomamajoke.png)\n\nMicrosoft Copilot Studio에서 작동하는 Jokes MCP 서버였습니다. 이제 실습도 끝났습니다! 실습이 마음에 드셨기를 바랍니다. 시간을 내어 [피드백 양식을](https://aka.ms/mcsmcp/lab/feedback) 작성해 주세요.\n\n## 농담 MCP 서버 세부 정보\n\n이는 [TypeScript SDK](https://github.com/modelcontextprotocol/typescript-sdk) 기반으로 구축된 [MCP](https://modelcontextprotocol.io/introduction) 서버입니다.\n\n이 MCP 서버를 사용하면 다음 웹사이트에서 농담을 가져올 수 있습니다.\n\n* [척노리스.io](https://api.chucknorris.io/)\n* [이칸하자다조크닷컴](https://icanhazdadjoke.com/)\n* [요마마-조크스닷컴](https://www.yomama-jokes.com)\n\n서버를 로컬에서 실행하려면 저장소 루트에서 `npm install` 실행해야 합니다.\n\n그런 다음 `npm run build` 실행하여 서버를 빌드하고 `npm start` 서버를 시작할 수 있습니다.\n\n## 도구\n\n다음 도구가 포함되어 있습니다.\n\n### 척 농담을 하다\n\n이 도구는 [chucknorris.io](https://api.chucknorris.io/) 에서 무작위 척 노리스 농담을 검색합니다.\n\n### 겟척카테고리\n\n이 도구는 [chucknorris.io](https://api.chucknorris.io/) 에서 사용 가능한 카테고리를 검색합니다.\n\n### 아빠 농담\n\n이 도구는 [icanhazdadjoke.com](https://icanhazdadjoke.com/) 에서 무작위 아빠 농담을 검색합니다.\n\n### 겟요마마 농담\n\n이 도구는 [yomama-jokes.com](https://www.yomama-jokes.com) 에서 무작위의 요 마마 농담을 검색합니다.\n\n## 알려진 문제 및 계획된 개선 사항\n\nMicrosoft Copilot Studio의 MCP에는 몇 가지 알려진 문제와 계획된 개선 사항이 있습니다. 자세한 내용은 [이 Microsoft Learn 문서](https://aka.ms/mcsmcpdocs#known-issues--planned-improvements) 에 나와 있습니다.\n\n## 기여하다\n\n이 프로젝트는 기여와 제안을 환영합니다. 대부분의 기여는 기여자 라이선스 계약(CLA)에 동의해야 합니다. 이 계약은 기여자가 기여한 내용을 사용할 권리를 보유하고 있으며, 실제로 이를 부여함을 명시합니다. 자세한 내용은 [https://cla.opensource.microsoft.com 을](https://cla.opensource.microsoft.com) 방문하세요.\n\n풀 리퀘스트를 제출하면 CLA 봇이 CLA 제출 필요 여부를 자동으로 판단하고 PR에 적절한 내용(예: 상태 확인, 댓글)을 추가합니다. 봇이 제공하는 안내를 따르세요. CLA를 사용하는 모든 저장소에서 이 작업은 한 번만 수행하면 됩니다.\n\n이 프로젝트는 [Microsoft 오픈 소스 행동 강령을](https://opensource.microsoft.com/codeofconduct/) 채택했습니다. 자세한 내용은 [행동 강령 FAQ를](https://opensource.microsoft.com/codeofconduct/faq/) 참조하시거나, 추가 질문이나 의견이 있으시면 으로 문의해 주세요.\n\n## 상표\n\n이 프로젝트에는 프로젝트, 제품 또는 서비스에 대한 상표 또는 로고가 포함될 수 있습니다. Microsoft 상표 또는 로고의 허가된 사용은 [Microsoft의 상표 및 브랜드 지침을](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general) 준수해야 합니다. 이 프로젝트의 수정된 버전에서 Microsoft 상표 또는 로고를 사용하는 것은 혼동을 야기하거나 Microsoft의 후원을 암시해서는 안 됩니다. 타사 상표 또는 로고를 사용하는 경우 해당 타사의 정책이 적용됩니다.","Korean","ko-KR",{"_29":62,"_33":66,"_38":67},{"_31":63,"_33":64,"_35":65},"模型上下文协议服务器使 Microsoft Copilot Studio 能够从各种来源获取笑话，包括查克·诺里斯笑话、老爸笑话和你妈妈笑话。","笑话 MCP 服务器","# Microsoft Copilot Studio ❤️ MCP\n\n欢迎来到**Microsoft Copilot Studio ❤️ MCP**实验室。在本实验室中，您将学习如何部署 MCP 服务器，以及如何将其添加到 Microsoft Copilot Studio。\n\n## 什么是 MCP？\n\n[模型上下文协议 (MCP)](https://modelcontextprotocol.io/introduction)是一种开放协议，它标准化了应用程序如何向 LLM 提供上下文，由[Anthropic](https://www.anthropic.com/)定义。MCP 提供了一种将 AI 模型连接到不同数据源和工具的标准化方法。MCP 允许创客将现有的知识服务器和 API 无缝集成到 Copilot Studio 中。\n\n目前，Copilot Studio 仅支持工具。要了解更多关于当前功能的信息，请访问[aka.ms/mcsmcp](https://aka.ms/mcsmcp) 。目前存在一些已知问题和计划改进。具体内容列[于此处](#known-issues-and-planned-improvements)。\n\n## MCP 与连接器\n\n什么时候使用 MCP？什么时候使用连接器？MCP 会取代连接器吗？\n\nMCP 服务器通过连接器基础架构提供给 Copilot Studio 使用，因此这些问题实际上并不适用。MCP 服务器使用连接器基础架构，这意味着它们可以采用企业安全和治理控制措施，例如[虚拟网络](https://learn.microsoft.com/power-platform/admin/vnet-support-overview)集成、[数据丢失防护](https://learn.microsoft.com/power-platform/admin/wp-data-loss-prevention)控制、[多种身份验证方法](https://learn.microsoft.com/connectors/custom-connectors/#2-secure-your-api)（所有这些功能在此版本中均已提供），同时支持 AI 客服人员的实时数据访问。\n\n因此，MCP 和连接器**结合在一起确实效果更好**。\n\n## 先决条件\n\n* Azure 订阅（已添加付款方式）\n* 拥有 GitHub 帐户并登录\n* Copilot Studio 试用版或开发者帐户\n* Power Platform 环境已预配\n\n## 实验室\n\n为了能够部署此 MCP 服务器并在 Microsoft Copilot Studio 中使用它，您需要执行以下操作：\n\n* [根据模板创建一个新的 GitHub 存储库](#create-a-new-github-repository-based-on-the-template)\n* [部署 Azure Web 应用](#deploy-the-azure-web-app)\n* [创建 Power Platform 连接器](#create-the-power-platform-connector)\n* [在 Microsoft Copilot Studio 中添加 MCP 服务器作为操作](#add-the-mcp-server-as-an-action-in-microsoft-copilot-studio)\n\n### 根据模板创建一个新的 GitHub 存储库\n\n1. 选择`Use this template`\n\n2. 选择“创建新存储库”\n\n ![](./assets/usetemplate.png)\n\n3. 选择合适的`Owner`\n\n4. 给它一个`Repository name`\n\n5. 您可以选择为其添加`Description`\n\n6. 选择`Private`\n\n7. 选择`Create repository`\n\n 这需要一点时间。完成后，您将被引导至新创建的存储库。\n\n> \\[！警告]\\\n> 完成本实验中的步骤后，您将拥有一个在 Azure 上运行的公开可用的 MCP 服务器。理想情况下，您不希望出现这种情况。请确保在完成本实验后删除 Azure Web 应用。\n\n### 部署 Azure Web 应用\n\n1. 前往`portal.azure.com`\n\n ![Azure 门户](/assets/azureportal.png)\n\n2. 将鼠标悬停在`Resource groups`选项卡上并选择`Create`\n\n ![资源组](/assets/selectResourceGroup.png)\n\n3. 单击`Subscription`下拉菜单并从列表中选择您的订阅\n\n ![选择订阅](/assets/selectSubscription.png)\n\n4. 单击`Resource group`名称文本框并输入资源组名称，例如“jokesgrp”\n\n ![资源组名称](/assets/resourcegrpname.png)\n\n5. 选择`Region`下拉菜单，选择距离您最近的区域\n\n ![资源组区域](/assets/resourcegrpregion.png)\n\n6. 单击`Review + create` Click Review + create\n\n ![审查并创建](/assets/resourcegrpreview.png)\n\n7. 单击`Create`\n\n ![创造](/assets/resourcegrpcreate.png)\n\n8. 选择`Go to resource group`\n\n ![前往资源](/assets/gotoresource.png)\n\n9. 选择`Create`\n\n ![创造](/assets/inresourcegrpcreate.png)\n\n10. 点击`search box`\n\n![搜索](/assets/searchmarketplace.png)\n\n11. 键入`web app`\n\n12. 在 Web 应用搜索结果中选择`Create button`\n\n![搜索结果](/assets/createwebapp1.png)\n\n13. 选择`Web App`\n\n![创建 Web 应用程序](/assets/createwebapp2.png)\n\n14. 选择`Name`文本输入并输入网络应用的名称，例如“wegotjokes”\n\n![Web 应用程序名称](/assets/webappname.png)\n\n15. 选择`Runtime stack`下拉菜单并选择`Node 22 LTS`\n\n![Web 应用运行时](/assets/selectruntime.png)\n\n16. 选择`Region`下拉菜单并选择距离您最近的区域\n\n![地区](/assets/webappselectregion.png)\n\n17. 单击`Review + create` Click Review + create\n\n![审查](/assets/webappreview.png)\n\n18. 单击`Create`\n\n![创造](/assets/webappcreate.png)\n\n19. 单击`Refresh`以更新部署进度\n\n![刷新](/assets/webapprefresh.png)\n\n20. 部署完成后，选择`Go to resource`\n\n![前往资源](/assets/webappgotoresource.png)\n\n21. 展开`Deployment`选项卡\n\n![部署选项卡](/assets/webappdeployment.png)\n\n22. 单击`Deployment Center`\n\n![部署中心](/assets/webappdeployctr.png)\n\n23. 单击`Source`下拉菜单并选择`GitHub`\n\n![选择来源](/assets/webappsource.png)\n\n24. 选择`Authorize`\n\n![授权](/assets/deployauthorize.png)\n\n25. 选择`Continue`\n\n![继续](/assets/githubauthcontinue.png)\n\n26. 选择`Authorize AzureAppService`\n\n![授权](/assets/githubauthappservice.png)\n\n27. 单击`Organization`下拉菜单并选择您的组织\n\n![组织选择](/assets/githuborg.png)\n\n28. 单击`Repository`下拉菜单，然后从列表中选择您[之前](#create-a-new-github-repository-based-on-the-template)创建的 GitHub 存储库\n\n![回购选择](/assets/githubrepo.png)\n\n29. 单击`Branch`下拉菜单并选择您的分支\n\n![选择分支](/assets/githubbranch.png)\n\n30. 保留其他所有内容的默认设置并选择`Save`\n\n![节省](/assets/githubsave.png)\n\n31. 单击`Refresh`以更新部署状态\n\n![刷新](/assets/githubrefresh.png)\n\n32. 验证部署是否完成\n\n![部署完成](/assets/deploydone.png)\n\n### 创建 Power Platform 连接器\n\n1. 转到\n\n2. 在左侧菜单中选择`More`\n\n3. 选择`Discover all`按钮\n\n4. 在“数据”下，选择`Custom connectors`旁边的`pin icon` ，将其固定到左侧菜单\n\n5. 在左侧菜单中选择`Custom connectors`\n\n6. 选择`New custom connector`\n\n7. 选择`Create from blank`\n\n ![](./assets/newconnector.png)\n\n8. 给连接器命名（例如`JokesMCP` ）\n\n9. 选择蓝色的`Continue`按钮\n\n10. 选择`Swagger editor`切换\n\n![](./assets/swaggereditor.png)\n\n11. 从[此处](./assets/connector.yml)复制 yaml 代码并替换 Swagger 编辑器中的代码\n\n12. 在 Swagger 编辑器中，将`dummyurl.azurewebsites.net`替换为你[之前](#deploy-the-azure-web-app)创建的 Web 应用的 URL。确保删除`https://`以及`azurewebsites.net`后面的所有内容。\n\n13. 选择`Create connector`来创建连接器\n\n### 在 Microsoft Copilot Studio 中添加 MCP 服务器作为操作\n\n现在，您已在 Azure 中运行 MCP 服务器，并在 Power Platform 中提供了连接器。此步骤涉及在 Microsoft Copilot Studio 中创建代理，并将 MCP 服务器添加到该代理。\n\n1. 转到\n\n2. 选择右上角的环境选择器\n\n3. 选择合适的环境\n\n4. 在左侧导航中选择`Create`\n\n5. 选择蓝色的`New agent`按钮\n\n ![新代理](./assets/newagent.png)\n\n6. 选择右上角的`Skip to configure`\n\n ![跳过配置](./assets/skiptoconfigure.png)\n\n7. 将名称更改为`Jokester`\n\n8. 添加以下`Description`\n\n ```text\n A humor-focused agent that delivers concise, engaging jokes only upon user request, adapting its style to match the user's tone and preferences. It remains in character, avoids repetition, and filters out offensive content to ensure a consistently appropriate and witty experience.\n ```\n\n9. 添加以下`Instructions`\n\n ```text\n You are a joke-telling assistant. Your sole purpose is to deliver appropriate, clever, and engaging jokes upon request. Follow these rules:\n\n * Respond only when the user asks for a joke or something related (e.g., \"Tell me something funny\").\n * Match the tone and humor preference of the user based on their input—clean, dark, dry, pun-based, dad jokes, etc.\n * Never break character or provide information unrelated to humor.\n * Keep jokes concise and clearly formatted.\n * Avoid offensive, discriminatory, or NSFW content.\n * When unsure about humor preference, default to a clever and universally appropriate joke.\n * Do not repeat jokes within the same session.\n * Avoid explaining the joke unless explicitly asked.\n * Be responsive, witty, and quick.\n ```\n\n10. 选择右上角的`Create`\n\n![创建代理](./assets/createagent.png)\n\n11. 启用生成式 AI `Orchestration`\n\n![启用业务流程管理](./assets/turnonorchestration.png)\n\n12. 在`Knowledge`部分中禁用常识\n\n![关闭常识](./assets/turnoffgeneralknowledge.png)\n\n13. 在顶部菜单中选择`Actions`\n\n![行动](./assets/actions.png)\n\n14. 选择`Add an action`\n\n![添加操作](./assets/addanaction.png)\n\n15. 搜索您[之前](#create-the-power-platform-connector)创建的连接器的名称（在本例中为`jokes` ）（参见下面屏幕截图中的数字 1）\n\n16. 选择`Jokes MCP server` （参见下面屏幕截图中的数字 2）\n\n![搜索并选择操作](./assets/addaction.png)\n\n17. 等待连接创建，完成后选择`Next`\n\n![动作与连接](./assets/actionconnection.png)\n\n18. `Description for the agent to know when to use this action`更改为以下文本：\n\n```text\nTrigger this action when a user asks for a joke. It can provide Chuck Norris jokes, Dad jokes and Yo Mama jokes.\n```\n\n将其余部分保留为默认值，例如最终用户身份验证，您将在一分钟内了解更多信息。\n\n19. 选择`Add action`将操作添加到代理\n\n![添加操作](./assets/addactionend.png)\n\n20. 在`Test your agent`窗格中选择`refresh icon`\n\n![刷新测试窗格](./assets/refreshtestingpane.png)\n\n21. 在`Test your agent`窗格中发送以下消息：\n\n```text\nCan I get a Chuck Norris joke?\n```\n\n系统会显示一条消息，提示您运行此操作需要额外权限。这是因为操作向导中存在用户身份验证。\n\n22. 选择`Connect`\n\n![附加权限](./assets/additionalpermissions.png)\n\n这将打开一个新窗口，您可以在其中管理此代理的连接。\n\n23. 选择`JokesMCP`旁边的`Connect`\n\n![连接到 JokesMCP](./assets/connect.png)\n\n24. 等待连接创建并选择`Submit`\n\n![选择一个连接](./assets/submitconnection.png)\n\n25. 现在应该已连接，因此状态应设置为`Connected`\n\n![状态已连接](./assets/connected.png)\n\n26. 关闭浏览器中的“管理连接”选项卡\n\n现在您应该回到 Jokester 代理屏幕。\n\n27. 在`Test your agent`窗格中选择`refresh icon`\n\n![刷新测试窗格](./assets/refreshtestingpane.png)\n\n28. 在`Test your agent`窗格中发送以下消息：\n\n```text\nCan I get a Chuck Norris joke?\n```\n\n现在将显示查克·诺里斯 (Chuck Norris) 的笑话 - 而不是额外的权限。\n\n![查克·诺里斯笑话](./assets/chucknorrisjoke.png)\n\n29. 在`Test your agent`窗格中发送以下消息：\n\n```text\nCan I get a Dad joke?\n```\n\n现在将显示一个老爸笑话。\n\n![老爸笑话](./assets/dadjoke.png)\n\n30. 在`Test your agent`窗格中发送以下消息：\n\n```text\nCan I get a Yo Mama joke?\n```\n\n现在将显示一个 Yo Mama 笑话。\n\n![你妈妈的笑话](./assets/yomamajoke.png)\n\n以上就是 Jokes MCP 服务器在 Microsoft Copilot Studio 中的运行情况。本次实验也到此结束！希望您喜欢本次实验。请抽出时间填写我们的[反馈表](https://aka.ms/mcsmcp/lab/feedback)。\n\n## 笑话 MCP 服务器详细信息\n\n这是一个基于[TypeScript SDK](https://github.com/modelcontextprotocol/typescript-sdk)构建的[MCP](https://modelcontextprotocol.io/introduction)服务器。\n\n通过此 MCP 服务器，您将能够从以下网站获取笑话：\n\n* [chucknorris.io](https://api.chucknorris.io/)\n* [icanhazdadjoke.com](https://icanhazdadjoke.com/)\n* [yomama-jokes.com](https://www.yomama-jokes.com)\n\n如果您想在本地运行服务器，请确保在存储库的根目录中运行`npm install` 。\n\n之后，您可以运行`npm run build`来构建服务器，并`npm start`来启动服务器。\n\n## 工具\n\n包含以下工具：\n\n### 得到查克笑话\n\n此工具从[chucknorris.io](https://api.chucknorris.io/)检索随机的 Chuck Norris 笑话。\n\n### 获取查克类别\n\n此工具从[chucknorris.io](https://api.chucknorris.io/)检索可用类别。\n\n### 爸爸笑话\n\n此工具从[icanhazdadjoke.com](https://icanhazdadjoke.com/)检索随机老爸笑话。\n\n### 让你妈妈笑话\n\n此工具从[yomama-jokes.com](https://www.yomama-jokes.com)检索随机的 Yo Mama Joke。\n\n## 已知问题和计划的改进\n\nMicrosoft Copilot Studio 中的 MCP 存在一些已知问题和计划改进。这些改进已在[Microsoft Learn 文章](https://aka.ms/mcsmcpdocs#known-issues--planned-improvements)中列出。\n\n## 贡献\n\n该项目欢迎您的贡献和建议。大多数贡献都要求您同意一份贡献者许可协议 (CLA)，该协议声明您有权并确实授予我们使用您贡献的权利。详情请访问。\n\n当您提交拉取请求时，CLA 机器人会自动判断您是否需要提供 CLA，并适当地修饰 PR（例如，状态检查、评论）。只需按照机器人提供的说明操作即可。您只需在所有使用我们 CLA 的仓库中执行一次此操作。\n\n本项目已采用[Microsoft 开源行为准则](https://opensource.microsoft.com/codeofconduct/)。更多信息请参阅[行为准则常见问题解答](https://opensource.microsoft.com/codeofconduct/faq/)，或联系提出其他问题或意见。\n\n## 商标\n\n本项目可能包含项目、产品或服务的商标或徽标。Microsoft 商标或徽标的授权使用须遵守[Microsoft 的商标和品牌指南](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general)。在本项目的修改版本中使用 Microsoft 商标或徽标不得引起混淆或暗示 Microsoft 的赞助。任何第三方商标或徽标的使用均须遵守该第三方的政策。","Chinese","zh-CN","githubUser","locale",{"_71":72,"_73":74,"_13":81,"_84":32,"_85":32,"_86":87,"_90":91,"_113":-5,"_16":114,"_115":116,"_117":118,"_33":34,"_129":104,"_130":-5,"_131":-5,"_132":-5,"_133":-5,"_134":135,"_140":-5,"_125":100,"_141":-5,"_142":-5,"_105":143,"_144":145},"addedAt",["D",1746352920541],"argumentsJsonSchema",{"_75":76,"_77":78,"_79":80},"properties",{},"required",[],"type","object",[82,83],"language:typescript","license_grade:a","descriptionMarkdown","descriptionPlainText","githubProject",{"_35":36,"_88":89},"url","https://github.com/VoltScp/MCPMS","githubRepository",{"_92":93,"_94":95,"_96":97,"_98":99,"_33":100,"_101":102,"_107":108,"_109":110,"_112":108},"defaultBranch","main","fullName","VoltScp/MCPMS","githubTreeId",114537,"language","TypeScript","MCPMS","owner",{"_103":104,"_105":106},"login","VoltScp","uid","q7exVT2NWu","recentStargazers",0,"spdxLicense",{"_33":111},"MIT License","stargazers","glamaJson",6280,"inspectable",false,"integrations",[119],{"_120":121,"_31":127,"_105":128},"brand",{"_122":123,"_33":124,"_125":126},"logoSvgPath","M12 .297c-6.63 0-12 5.373-12 12 0 5.303 3.438 9.8 8.205 11.385.6.113.82-.258.82-.577 0-.285-.01-1.04-.015-2.04-3.338.724-4.042-1.61-4.042-1.61C4.422 18.07 3.633 17.7 3.633 17.7c-1.087-.744.084-.729.084-.729 1.205.084 1.838 1.236 1.838 1.236 1.07 1.835 2.809 1.305 3.495.998.108-.776.417-1.305.76-1.605-2.665-.3-5.466-1.332-5.466-5.93 0-1.31.465-2.38 1.235-3.22-.135-.303-.54-1.523.105-3.176 0 0 1.005-.322 3.3 1.23.96-.267 1.98-.399 3-.405 1.02.006 2.04.138 3 .405 2.28-1.552 3.285-1.23 3.285-1.23.645 1.653.24 2.873.12 3.176.765.84 1.23 1.91 1.23 3.22 0 4.61-2.805 5.625-5.475 5.92.42.36.81 1.096.81 2.22 0 1.606-.015 2.896-.015 3.286 0 .315.21.69.825.57C20.565 22.092 24 17.592 24 12.297c0-6.627-5.373-12-12-12","GitHub","slug","github","Allows deploying the MCP server through GitHub integration with Azure Web App's deployment center","sdst5kwbzs","namespace","npmPackage","prompts","redditPostForeignId","resources","scores",{"_136":137,"_138":-5,"_139":-5},"license",100,"quality","security","simpleIcon","supportedPlatforms","tools","lp7af3xkle","updatedAt",["D",1746352938253],"meta",[148,151,153,160,163,165,168,171,174,177,180,182,185,188,190,192,194,196,198],{"_149":150},"title","Jokes MCP Server | Glama",{"_152":32,"_33":31},"content",{"_154":155,"_156":157,"_158":159},"href","https://glama.ai/mcp/servers/@VoltScp/MCPMS","rel","canonical","tagName","link",{"_152":34,"_161":162},"property","og:title",{"_152":32,"_161":164},"og:description",{"_152":166,"_161":167},"https://glama.ai/logo.png","og:logo",{"_152":169,"_161":170},"https://glama.ai/mcp/servers/@VoltScp/MCPMS/og-image","og:image",{"_152":172,"_161":173},"630","og:image:height",{"_152":175,"_161":176},"1200","og:image:width",{"_152":178,"_161":179},"website","og:type",{"_152":155,"_161":181},"og:url",{"_152":183,"_161":184},"Glama – MCP Hosting Platform","og:site_name",{"_152":186,"_33":187},"summary_large_image","twitter:card",{"_152":183,"_33":189},"twitter:title",{"_152":32,"_33":191},"twitter:description",{"_152":169,"_33":193},"twitter:image",{"_152":172,"_33":195},"twitter:image:height",{"_152":175,"_33":197},"twitter:image:width",{"_199":200},"script:ld+json",{"_201":202,"_203":204},"@context","http://schema.org","@graph",[205],{"_201":206,"_207":208,"_209":210},"https://schema.org","@type","BreadcrumbList","itemListElement",[211,218,222,226],{"_207":212,"_213":214,"_33":215,"_216":217},"ListItem","item","https://glama.ai/","Glama","position",1,{"_207":212,"_213":219,"_33":220,"_216":221},"https://glama.ai/mcp","MCP",2,{"_207":212,"_213":223,"_33":224,"_216":225},"https://glama.ai/mcp/servers","Servers",3,{"_207":212,"_213":155,"_33":34,"_216":227},4,"visitorIsMaintainer","routes/_public/mcp/servers/~namespace/~slug/_pages/reviews/_route",{"_68":-5,"_29":231,"_146":246,"_281":-5},{"_71":232,"_73":233,"_13":236,"_84":32,"_85":32,"_86":237,"_90":238,"_113":-5,"_16":114,"_115":116,"_117":241,"_33":34,"_129":104,"_130":-5,"_131":-5,"_132":-5,"_133":-5,"_134":244,"_140":-5,"_125":100,"_141":-5,"_142":-5,"_105":143,"_144":245},["D",1746352920541],{"_75":234,"_77":235,"_79":80},{},[],[82,83],{"_35":36,"_88":89},{"_92":93,"_94":95,"_96":97,"_98":99,"_33":100,"_101":239,"_107":108,"_109":240,"_112":108},{"_103":104,"_105":106},{"_33":111},[242],{"_120":243,"_31":127,"_105":128},{"_122":123,"_33":124,"_125":126},{"_136":137,"_138":-5,"_139":-5},["D",1746352938253],[247,249,251,253,255,256,257,258,259,260,261,262,263,264,265,266,267,268,269],{"_149":248},"Reviews | Jokes MCP Server | Glama",{"_152":250,"_33":31},"Read community reviews and feedback for Jokes MCP Server. Discover user experiences, ratings, strengths, and potential limitations of this MCP implementation from actual users.",{"_154":252,"_156":157,"_158":159},"https://glama.ai/mcp/servers/@VoltScp/MCPMS/reviews",{"_152":254,"_161":162},"Reviews | Jokes MCP Server",{"_152":250,"_161":164},{"_152":166,"_161":167},{"_152":169,"_161":170},{"_152":172,"_161":173},{"_152":175,"_161":176},{"_152":178,"_161":179},{"_152":252,"_161":181},{"_152":183,"_161":184},{"_152":186,"_33":187},{"_152":183,"_33":189},{"_152":250,"_33":191},{"_152":169,"_33":193},{"_152":172,"_33":195},{"_152":175,"_33":197},{"_199":270},{"_201":202,"_203":271},[272],{"_201":206,"_207":208,"_209":273},[274,275,276,277,278],{"_207":212,"_213":214,"_33":215,"_216":217},{"_207":212,"_213":219,"_33":220,"_216":221},{"_207":212,"_213":223,"_33":224,"_216":225},{"_207":212,"_213":155,"_33":34,"_216":227},{"_207":212,"_213":252,"_33":279,"_216":280},"Reviews",5,"review","actionData","errors"]

Rate & Review