{
  "markdown": "# Customer Experience Agent Studio Managed MCP Extension\n\nThe Customer Experience Agent Studio managed MCP extension enables AI-assisted\ndevelopment workflows, drastically reducing the friction of building and\nmaintaining agent applications.\n\n## Why use the Customer Experience Agent Studio managed MCP server?\n\nGoogle and Google Cloud\n[managed MCP servers](https://docs.cloud.google.com/mcp/overview) can be used in\nyour AI applications with enterprise-ready governance, security, and access\ncontrol.\n\n## Before you begin\n\n1.  In the Google Cloud console, on the\n    [project selector page](https://console.cloud.google.com/projectselector2/home/dashboard),\n    select or create a Google Cloud project.\n\n    > **Note**: If you don't plan to keep the resources that you create in this\n    > procedure, create a project instead of selecting an existing project.\n    > After you finish these steps, you can delete the project, removing all\n    > resources associated with the project.\n\n2.  Get your administrator to grant you the\n    [MCP Tool User role](https://docs.cloud.google.com/iam/docs/roles-permissions/mcp#mcp.toolUser)\n    (`roles/mcp.toolUser`) on the Google Cloud project. If you created a new\n    project, then you already have the required permissions.\n\n3.  Ensure your administrator has enabled the\n    [Customer Experience Agent Studio API](https://console.cloud.google.com/marketplace/product/google/ces.googleapis.com)\n    on the Google Cloud project.\n\n## Configure authentication\n\nThis extension uses Google Application Default Credentials (ADC) to perform\nauthentication. To login with ADC, run the following command in your terminal:\n\n```bash\ngcloud auth application-default login\n```\n\nFor additional details, see the\n[ADC documentation](https://docs.cloud.google.com/docs/authentication/application-default-credentials#personal).\n\n## Available tools\n\nTo see a complete list of available tools and their schemas, see the\n[Customer Experience Agent Studio MCP reference](https://docs.cloud.google.com/customer-engagement-ai/conversational-agents/ps/reference/mcp).\n\n## Limitations\n\nThe following limitations apply:\n\n-   Token Context Limits: Large agent configurations (with many tools and\n    extensive instructions) may exceed the context window of some coding models\n    when retrieving full agent definitions. We recommend fetching specific\n    sub-components (for example, just one tool) rather than the entire app\n    definition at once. Additionally, users should monitor their context window\n    usage and restart their agent session periodically (every few requests) to\n    clear the buffer.\n-   Latency: \"Direct Mutation\" (API calls) are generally faster for small\n    changes. However, for massive architectural refactors (renaming variables\n    across 50 files), we recommend the \"Export -> Local Edit -> Import\" workflow\n    to ensure data integrity, which the MCP server also supports using\n    export_app and import_app tools.\n\n## Example use cases\n\nThe following are sample use cases for the Customer Experience Agent Studio MCP\nserver:\n\n-   Vibe Coding (Rapid Prototyping): Instead of manually clicking through the UI\n    to create an agent, you can simply tell your AI-assisted IDE \"Create a\n    retail support agent that uses the Shopify API and speaks in a friendly\n    tone.\" The coding agent uses the MCP server to construct the agent\n    architecture for you.\n-   Mass Refactoring & Clean-up: The MCP server excels at bulk operations that\n    are tedious in a UI. For example, you can command \"Rename the 'customer_id'\n    parameter across all 15 sub-agents\" or \"Find and delete all unused intents\".\n-   Interactive Eval-Driven Development: You can run a failing evaluation and\n    instruct the agent: \"Modify the instructions until this specific evaluation\n    passes\".\n-   Self-Healing & Optimization: An automated \"Helper Agent\" can monitor an\n    agent's performance (for example, failing a specific evaluation) and use the\n    MCP server to autonomously tweak instructions or fix tool definitions to\n    improve the score (\"Hill Climbing\").\n-   Contextual Awareness: The server allows coding assistants to \"read\" the\n    current state of a deployed agent, making it easier for you to understand\n    complex legacy configurations without digging through JSON files manually.\n\n## Optional security and safety configurations\n\nMCP introduces new security risks and considerations due to the wide variety of\nactions that you can take with MCP tools. To minimize and manage these risks,\nGoogle Cloud offers defaults and customizable policies to control the use of MCP\ntools in your Google Cloud organization or project.\n\nFor more information about MCP security and governance, see\n[AI security and safety](https://docs.cloud.google.com/mcp/ai-security-safety).\n\n## Quotas and limits\n\nThe Customer Experience Agent Studio MCP server doesn't have its own quotas.\nThere is no limit on the number of call that can be made to the MCP server. You\nare still subject to the quotas enforced by the APIs called by the MCP server\ntools.\n\n## Reference and resources\n\n*   Explore the\n    [Customer Experience Agent Studio remote MCP server reference documentation](https://docs.cloud.google.com/customer-engagement-ai/conversational-agents/ps/reference/mcp),\n    which includes a list of all available tools, and the full input and output\n    schema for each tool.\n*   See the\n    [Customer Experience Agent Studio overview](https://docs.cloud.google.com/customer-engagement-ai).\n*   Learn about\n    [MCP security and governance](https://docs.cloud.google.com/mcp/ai-security-safety).\n",
  "bytes": 5613,
  "sha": "f2622a49cbf76119010b051b4bf9813c5780c8e0f7409e75fb0d3f56dc96dabb",
  "repo_slug": "gemini-cli-extensions/customer-experience-agent-studio",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_gemini_cli_extensions_customer_experienc_5ab7c6f6/readme"
}