home / mcp / alibabacloud devops mcp server
Provides AI-assisted DevOps actions by connecting to Yunxiao for reading work items, generating code, and submitting merge requests.
Configuration
View docs{
"mcpServers": {
"aliyun-alibabacloud-devops-mcp-server": {
"command": "npx",
"args": [
"-y",
"alibabacloud-devops-mcp-server"
],
"env": {
"YUNXIAO_ACCESS_TOKEN": "<YOUR_TOKEN>"
}
}
}
}AlibabaCloud Devops MCP Server lets AI assistants interact with Yunxiao to read work items, automatically generate code, and submit merge requests. It helps your enterprise development team streamline reviews, optimize task management, and focus on delivering products faster.
You connect an MCP client to the server to enable AI-assisted workflows. The server supports stdio mode for local, direct communication and can also be run in Docker with SSE for HTTP-based access. Your client can then request organization, code, project, pipeline, application delivery, packages, and test management actions via the MCP endpoints exposed by the server.
Prerequisites you need to meet before running the server or integrating with an MCP client are: Node.js version 18 or newer and an active Yunxiao access token with read and write permissions for the APIs you plan to use.
Option A: Quick start with NPX (stdio mode) in your MCP client configuration:
{
"mcpServers": {
"yunxiao": {
"command": "npx",
"args": [
"-y",
"alibabacloud-devops-mcp-server"
],
"env": {
"YUNXIAO_ACCESS_TOKEN": "<YOUR_TOKEN>"
}
}
}
}Option B: Run via Docker in stdio mode using a prebuilt image or your own build.
{
"mcpServers": {
"yunxiao": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"YUNXIAO_ACCESS_TOKEN",
"build-steps-public-registry.cn-beijing.cr.aliyuncs.com/build-steps/alibabacloud-devops-mcp-server:v0.2.0"
],
"env": {
"YUNXIAO_ACCESS_TOKEN": "<YOUR_TOKEN>"
}
}
}
}If you prefer running SSE over HTTP, you can configure a service that exposes a URL like http://localhost:3000/sse and connect your MCP client to that endpoint. The tokens can be supplied via the query string or request header to support multiple users.
The server supports toolsets that let you enable only the tools you need for better performance. By default, all tool groups are enabled if you do not specify toolsets.
Explicit toolsets available include organization-management, code-management, project-management, pipeline-management, packages-management, application-delivery, and test-management. Use these to tailor the MCP capabilities exposed to the AI assistant.
If you run in SSE mode, you can pass your own token in the URL or via the x-yunxiao-token header to authenticate individual users.
View logs to diagnose issues when running in Docker SSE or stdio mode. If you stop the service, ensure you properly terminate the container or process to free resources.
Protect your Yunxiao access token. Do not expose it in client configurations or logs. Use environment isolation for tokens per deployment and rotate tokens periodically.
Retrieve the current user's organization information.
Fetch the list of organizations the current user has joined.
Obtain information about an organization role.
List departments within an organization.
Get details about a specific department.
List members in an organization.
Get details for a project.
Search for projects.
Get details of a work item.
Create a new work item.
List all work item types in an organization.
Create a new repository branch.
List repositories in a code base.
Create a merge request.
Query a merge request.
Query merge request list.
List pipeline runs.
Get pipeline details.
Get pipeline list.
Create a deployment change order.
View deployment order versions.