mcp-server-circleci

概述

CircleCI MCP 服务器在 CircleCI 的 CI/CD 基础设施与 Cursor、Windsurf、Claude 和 VS Code 等 AI 助手之间提供了全面的桥梁。它提供了 18 个专用工具，用于分析构建、查找不稳定的测试、重新运行工作流、管理工件和优化资源使用。该服务器支持多种部署方法，包括 NPX、Docker 和自管理的远程服务器，并为每个流行的 IDE 和客户端提供了清晰的配置说明。

试试问 AI

装完之后，这里有 5 个你可以让 AI 做的事：

• analyze_diff

  Analyze git diffs against cursor rules for violations

  试试:Check my current git diff for any issues试试:Analyze the diff against code quality rules
• config_helper

  Validate and get guidance for your CircleCI configuration

  试试:Check my CircleCI config for issues试试:Help me optimize my .circleci/config.yml
• create_prompt_template

  Generate structured prompt templates for AI applications

  试试:Create a prompt template for user feedback analysis试试:Generate a template for code review
• download_usage_api_data

  Download usage data from the CircleCI Usage API

  试试:Get our CircleCI usage statistics试试:Download pipeline usage data
• find_flaky_tests

  Identify flaky tests by analyzing test execution history

  试试:Find flaky tests in our project试试:Which tests are inconsistent?
• find_underused_resource_classes

  Find jobs with underused compute resources

  试试:Find jobs that are wasting resources试试:Check for underutilized resource classes
• get_build_failure_logs

  Retrieve detailed failure logs from CircleCI builds

  试试:Get logs for the failed build试试:Show me the error details from the last failed pipeline
• get_job_test_results

  Retrieve test metadata and results for CircleCI jobs

  试试:Get test results for the last build试试:Show me the test coverage metrics
• get_latest_pipeline_status

  Get the status of the latest pipeline for a branch

  试试:Check the status of the main pipeline试试:Is my feature branch pipeline complete?
• list_artifacts

  List artifacts produced by a CircleCI job

  试试:Show me the build artifacts试试:List the deployment packages
• rerun_workflow

  Rerun a workflow from start or from the failed job

  试试:Rerun the failed workflow试试:Restart the pipeline from the failing step
• run_pipeline

  Trigger a pipeline to run

  试试:Run the pipeline for main branch试试:Trigger a build for my changes

可对比工具

安装

{
  "mcpServers": {
    "circleci-mcp-server": {
      "command": "npx",
      "args": ["-y", "@circleci/mcp-server-circleci@latest"],
      "env": {
        "CIRCLECI_TOKEN": "your-circleci-token",
        "CIRCLECI_BASE_URL": "https://circleci.com",
        "MAX_MCP_OUTPUT_LENGTH": "50000"
      }
    }
  }
}

{
  "servers": {
    "circleci-mcp-server": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@circleci/mcp-server-circleci@latest"],
      "env": {
        "CIRCLECI_TOKEN": "${input:circleci-token}",
        "CIRCLECI_BASE_URL": "${input:circleci-base-url}"
      }
    }
  }
}

FAQ

我的 CircleCI API 令牌需要什么权限？

该令牌需要 'Status' 和 'Settings' 范围才能访问管道信息、构建细节和配置验证功能。

我可以将此服务器用于自托管 CircleCI 吗？

可以，您需要将 CIRCLECI_BASE_URL 环境变量设置为您的自托管实例 URL，而不是默认的 'https://circleci.com'。

mcp-server-circleci 对比