mcp-task-orchestrator

概述

MCP任务编排器是一个基于Python的综合服务器，通过智能任务分解和执行改变AI辅助开发方式。它采用专业AI角色（架构师、实现者、测试者、审查者等）确保项目各个方面都获得领域专业知识。采用Clean Architecture原则构建，它能自动检测项目结构并持久化保存所有决策和实现。该服务器包含工作区智能、工件管理和全面的模板系统，具有13个用于可重用任务模板的工具。

试试问 AI

装完之后，这里有 8 个你可以让 AI 做的事：

• orchestrator_initialize_sessionorchestrator_initialize_session(working_directory)

  Start a new orchestration session with optional working directory

  试试:Initialize a new orchestration session试试:Initialize a new orchestration session for my project at /home/user/project
• orchestrator_plan_taskorchestrator_plan_task(task_description)

  Create a structured breakdown of tasks from a request

  试试:Plan a Python script for processing CSV files试试:Break down the requirements for an e-commerce website
• orchestrator_execute_taskorchestrator_execute_task(task_id)

  Execute a specific task with specialist role context

  试试:Execute task 3 with Implementer role context试试:Run the API implementation task
• orchestrator_complete_taskorchestrator_complete_task(task_id, artifacts)

  Mark a task as complete with associated artifacts

  试试:Complete task 5 with the test artifacts试试:Finish the documentation task with the generated PDF
• orchestrator_synthesize_resultsorchestrator_synthesize_results(project_id)

  Combine all task results into a comprehensive solution

  试试:Synthesize the results for my project试试:Create a comprehensive report from all completed tasks
• orchestrator_get_statusorchestrator_get_status(project_id)

  Check the progress and status of the orchestration session

  试试:What's the current status of my project?试试:Show me the progress of all tasks
• orchestrator_maintenance_coordinatororchestrator_maintenance_coordinator(action_type)

  Perform automated cleanup and optimization of sessions

  试试:Run the maintenance coordinator to cleanup old sessions试试:Optimize the database and archive stale tasks

可对比工具

安装

# 克隆并运行通用安装程序
git clone https://github.com/EchoingVesper/mcp-task-orchestrator.git
cd mcp-task-orchestrator
python install.py

pip install mcp-task-orchestrator

{
  "mcpServers": {
    "task-orchestrator": {
      "command": "python",
      "args": ["-m", "mcp_task_orchestrator.server"]
    }
  }
}

FAQ

支持哪些MCP客户端？

该服务器完全支持Claude Desktop、Cursor IDE、Windsurf以及带有扩展的VS Code。对Continue.dev和Cline有部分支持，并且正在持续开发中。

持久化内存如何工作？

服务器使用SQLite存储所有任务决策、实现和工件。它提供超过24小时任务的自动恢复和归档功能，确保永远不会丢失上下文。

可以自定义专业AI角色吗？

可以。您可以编辑`.task_orchestrator/roles/project_roles.yaml`来创建具有专业知识和方法指南的项目特定专家。

mcp-task-orchestrator 对比