# Copilot Studio [Health: Active]

**Category:** 🗣️ Conversational AI  
**Repository:** https://github.com/microsoft/copilot-studio-mcp  
**GitHub Stars:** 3  
**npm Downloads (last month):** 507  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/copilot-studio

## Description
MCP bridge for Copilot Studio sign-in, environment selection, and Agentic Builder access.

## Tools
Capabilities this server exposes over MCP:

- **discover_resources** — Discover connector operations and schemas, existing connections, dependent parameter values, and attachable resources.
- **get_consumption** — Read historical agent or workflow credit consumption, or current environment credit capacity.
- **manage_agents** — Manage agent inventory, complete draft authoring, lifecycle, published-agent sharing, saved-draft previews, and agent evaluation suites and runs.
- **manage_workflows** — Manage workflow inventory, complete draft authoring, lifecycle, real tests, and run diagnostics.
- **get_copilot_studio_skills** — Start here for Copilot Studio guidance. Read the matching core skill before authoring, then read `schema.json`.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "copilot-studio": {
    "command": "npx",
    "args": ["-y","@microsoft/copilot-studio-mcp@latest"]
  }
}
```

## Documentation

## What Copilot Studio does

The Copilot Studio MCP server exposes Copilot Studio operations through the Model Context Protocol. It provides one entry point for working with agents and workflows across inventory, draft authoring, lifecycle actions, testing, previews, sharing, diagnostics, and evaluation runs.

Resource discovery covers connector operations and schemas, existing connections, dependent parameter values, and resources that can be attached to an agent or workflow. Consumption queries provide historical credit usage for agents or workflows and the current credit capacity of an environment.

The server is intended for the Conversational AI category and is useful when an MCP-compatible development tool or agent needs to work with Copilot Studio rather than calling its hosted service through a custom integration.

## How it works

A client starts the package with npx and communicates with the local process over standard input and output. After sign-in and environment selection, the package connects to the hosted Copilot Studio service. Available access remains subject to the signed-in identity, selected environment, permissions, and organizational policies.

The main management tools accept an action and parameters. Before constructing unfamiliar requests, use resource discovery with `action=describe_actions` to inspect an action's contract and side-effect information. The `get_copilot_studio_skills` tool provides Copilot Studio guidance; its instructions say to read the relevant core skill before authoring and then consult `schema.json`.

## Setup and configuration

The Copilot Studio MCP server requires Node.js 22.18 or newer; Node.js 24 LTS is recommended. You also need an MCP-compatible client and a Microsoft work account with access to a Copilot Studio environment.

Configure the client to launch:

```text
npx -y @microsoft/copilot-studio-mcp@latest
```

The client starts the process locally, so you do not provide a remote MCP URL or clone the source repository. Installation is also available through the GitHub MCP Registry and supported client galleries, including the VS Code MCP gallery.

If setup fails, stop without sharing credentials or configuration JSON. Check endpoint availability and network or proxy settings before retrying. DNS failures, refused or reset connections, and timeouts have separate safe cause codes.

## Tools and capabilities

- `discover_resources` finds connector operations, schemas, connections, dependent parameter values, and attachable resources.
- `get_consumption` reads historical agent or workflow credit consumption and current environment credit capacity.
- `manage_agents` handles agent inventory, draft authoring, lifecycle operations, published-agent sharing, saved-draft previews, and evaluation suites and runs.
- `manage_workflows` handles workflow inventory, draft authoring, lifecycle operations, real tests, and run diagnostics.
- `get_copilot_studio_skills` supplies guidance for Copilot Studio authoring and schema use.

Connection and sign-in utilities are separate from the listed service catalog tools.

## Limitations and notes

The server does not remove Copilot Studio access controls: operations depend on the selected environment, account permissions, and organizational policy. A retry does not correct missing environment routing.

Organizations requiring FedRAMP High should note that the GitHub Copilot harness agent turns off at runtime to prevent data egress from the compliance boundary. The project is licensed under MIT.

_Full upstream README: https://allmcps.com/mcp/copilot-studio/readme_