mcp-codepipeline

This MCP server provides a standardized interface to manage AWS CodePipeline resources through Windsurf and Cascade. It exposes a set of tools for listing pipelines, retrieving pipeline details and executions, triggering and stopping executions, managing approvals, tagging resources, and creating webhooks, along with access to pipeline metrics and logs. After starting the server, you can use Windsurf to translate natural language requests into MCP tool invocations, such as listing all pipelines, getting the state of a specific pipeline, triggering a build, or fetching performance metrics. The included tools cover core pipeline management (list, get state, list executions, trigger, stop), details and logs (get details, get execution logs, get metrics), and actions/integrations (approve/reject actions, retry stages, tag resources, create webhooks).

To use these capabilities, ensure Windsurf is configured with the MCP server under the mcp_config.json entry. Once configured, Cascade will route natural-language requests to the appropriate MCP tool calls. For example, asking for all pipelines will map to list_pipelines, requesting the state of a pipeline maps to get_pipeline_state, triggering a pipeline maps to trigger_pipeline, and requesting metrics or logs maps to get_pipeline_metrics and get_pipeline_execution_logs respectively.

Prerequisites:

• Node.js (v14 or later)
• npm (comes with Node.js)
• AWS account with CodePipeline access and credentials with permissions for CodePipeline, CloudWatch, and IAM (for tagging)
• Windsurf IDE with Cascade AI assistant

Installation steps:

1. Clone the repository

git clone https://github.com/cuongdev/mcp-codepipeline-server.git
cd mcp-codepipeline-server

2. Install dependencies

npm install

3. Create a .env file based on the template in .env.example (if provided) or set environment variables as needed. Example:

cp .env.example .env

Set AWS credentials and region in .env or via the environment:

AWS_REGION=us-east-1 AWS_ACCESS_KEY_ID=your_access_key_id AWS_SECRET_ACCESS_KEY=your_secret_access_key PORT=3000

4) Start the server

```bash
npm start

5. (Optional) Development with auto-restart

npm run dev

6. Configure Windsurf with MCP config (see how_to_use for example) and ensure Windsurf is restarted to pick up changes.

Tips and notes:

• Do not commit credentials to version control. Use environment variables or a secure secret manager for AWS credentials.
• Ensure the AWS region in your environment matches where your CodePipeline resources reside.
• If Windsurf fails to detect the MCP server, verify the mcp_config.json format and restart Windsurf.
• The provided mcp_config example uses npx to run the server binary from a dist/index.js entry point; adjust the path to match your build output if you package differently.
• For security, limit IAM permissions to only what is required for the pipelines and tagging operations you perform.
• Logs are printed to the console; run with DEBUG=* npm start for verbose output during troubleshooting.