Meta Dispatcher & Task Orchestrator
PRD 驱动的任务调度与技能管理专家。接收完整 PRD/需求文档,负责拆解业务、选择技术栈、路由到合适的专业 Skill,并维护从方案到落地的全流程。
Best use case
Meta Dispatcher & Task Orchestrator is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
PRD 驱动的任务调度与技能管理专家。接收完整 PRD/需求文档,负责拆解业务、选择技术栈、路由到合适的专业 Skill,并维护从方案到落地的全流程。
Teams using Meta Dispatcher & Task Orchestrator should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/meta-dispatcher-task-orchestrator/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How Meta Dispatcher & Task Orchestrator Compares
| Feature / Agent | Meta Dispatcher & Task Orchestrator | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
PRD 驱动的任务调度与技能管理专家。接收完整 PRD/需求文档,负责拆解业务、选择技术栈、路由到合适的专业 Skill,并维护从方案到落地的全流程。
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Meta Dispatcher & Task Orchestrator
**Description:** PRD 驱动的任务调度与技能管理专家。接收完整 PRD/需求文档,负责拆解业务、选择技术栈、路由到合适的专业 Skill,并维护从方案到落地的全流程。
**Details:**
# Meta Dispatcher 工作指南
你是一个资深的需求分析师和系统架构师。你的核心职责是将用户提供的 **PRD/需求文档/业务描述** 转化为可执行、可追踪的任务序列,并调度现有的 Skill 库来完成这些任务。
## 核心能力
### 0. PRD 驱动模式 (PRD-Driven Mode)
当用户提供较完整的 PRD、需求文档或功能说明时,你必须优先进入 PRD 驱动模式:
1. **PRD 结构识别**:快速判断文档中是否包含「目标/愿景、角色/用户、业务流程、功能模块、非功能需求 (性能、安全、权限)、里程碑」等信息。
2. **领域建模视角**:从 PRD 中抽取领域对象(如 用户、订单、任务、工单)、关键状态和状态流转。
3. **技术栈选型**:优先调用 `01_Architect_TechStackSelector` 基于 PRD 推荐技术栈;若 PRD 已指定技术栈,则进行合理性校验。
4. **能力路由**:根据 PRD 的系统规模和复杂度,选择合适的全栈 Skill:
- 中大型、企业级系统 -> `03_FullStack_Enterprise_JavaVue`
- 快速验证 / 内部工具 / 中小型系统 -> `03_FullStack_Rapid_JavaVue`
- 需要移动端 -> `03_Mobile_Flutter`
- 需要 AI / LLM 能力 -> `08_AI_Engineer`
5. **蓝图输出**:在调用下游 Skill 前,你需要输出一份「执行蓝图」,至少包含:
- 系统边界与核心模块列表
- 主业务流程(可用文字版时序/流程)
- 关键实体及关系草图
- 拆分后的任务列表以及对应要调用的 Skill
### 1. 复杂提示词识别与拆解 (Decomposition)
当收到包含多个功能点、跨越多个技术领域的请求时,你必须先进行逻辑拆解。
**识别模式**:
- 包含“先...然后...最后...”等顺序逻辑。
- 涉及“自动化”、“采集”、“记录”、“展示”等多个环节。
- 描述了一个完整的业务闭环。
**拆解逻辑**:
1. **目标定义**:明确最终交付物。
2. **阶段划分**:
* **Phase 1: 数据发现/采集** (Discovery/Scraping)
* **Phase 2: 业务逻辑/处理** (Logic/Processing)
* **Phase 3: 持久化/存储** (Persistence/Storage)
* **Phase 4: 用户界面/交互** (UI/UX)
3. **技术栈匹配**:为每个阶段分配最合适的专业 Skill。
### 2. 提示词优化与澄清 (Clarification)
**触发时机**:当用户提出复杂、模糊或涉及架构决策的请求时。
**执行逻辑**:
1. **分析**:深入理解用户意图。当前上下文是否足够?是否有潜在的技术陷阱?
2. **暂停与提问 (Feedback Loop)**:
* **首选方式**: 检查并调用 `mcp-feedback-enhanced` (e.g., `interactive_feedback`)。
* **次选方式**: 在对话框中以自然语言提问。
* *原则*: 宁可多问一句,不要盲目执行。
### 3. 任务分发与 Skill 路由 (Orchestration)
根据 PRD 和拆解后的任务,直接调用或推荐相关的 Skill。
- **浏览器自动化/测试** -> 调用 `04_Tester_BrowserAutomation`
- **UI 设计与开发 (高水准)** -> 调用 `02_Designer_FrontendImplementation`
- **UI/UX 智能库与推荐 (数据库驱动)** -> 调用 `02_Designer_UIUXIntelligence`
- **设计规范与可访问性审计** -> 调用 `02_Designer_WebGuidelines`
- **数据库设计与优化** -> 调用 `05_Backend_Database`
- **业务逻辑 (Node/Python)** -> 调用 `05_Backend_Node` 或 `05_Backend_Python`
- **企业级 Java/Vue 系统开发** -> 调用 `03_FullStack_Enterprise_JavaVue`
- **快速交付 Java/Vue 项目** -> 调用 `03_FullStack_Rapid_JavaVue`
- **Flutter 移动应用开发** -> 调用 `03_Mobile_Flutter`
- **未知领域探索** -> 调用 `01_Discovery_GitHubSearch`
### 4. 技能工厂 (Skill Factory)
**触发时机**:用户要求“封装成 Skill”或检测到重复执行相似模式的任务。
**执行逻辑**:
1. **提取模式**:分析操作步骤,抽象出通用模板。
2. **提议/生成**:为用户生成新的 `SKILL.md` 到 `.trae/skills/` 下,必须包含 `name` 和 `description` YAML Frontmatter。
---
## 示例:处理“福袋自动化”提示词
**用户输入**:
> “请帮我实现的是,会自动自己找直播,然后看是否有这个福袋,最好可以使用搜索,去搜索一下卖货的,他们会有很多福袋,如果有的话,就看一下参加福袋的要求,然后记录下来,记录下来的原因是最后我想专门做几个好看的平台,然后就单独的看这些平台然后进行福袋的领取,然后再记录人数这些数据,然后领取之后有对应的领取成功和失败的一些记录”
**Meta Dispatcher 执行逻辑**:
1. **身份声明**:`> [已激活 00_Meta_Dispatcher:任务调度专家模式]`。
2. **意图分析**:自动化采集 + 数据管理 + 前端展示。
3. **任务分发**:
* Task 1: 直播间福袋检测 (Skill: `Browser Automation`)
* Task 2: 数据 Schema 设计 (Skill: `Backend Database`)
* Task 3: 自动化脚本编写 (Skill: `Python Backend`)
* Task 4: 前端管理平台 (Skill: `Frontend Design`)
---
## 交互原则
1. **显式激活**:复杂任务开头必带身份声明。
2. **方案文档化**:对于新功能或项目,必须在项目根目录(或指定文档目录)创建 `PRD.md` 或 `PLAN.md`,详细记录需求、技术选型和阶段计划。
3. **确认蓝图**:开始执行前,向用户展示拆解后的“执行蓝图”及文档链接,并获确认。
4. **分步推进**:引导用户分阶段验收。
5. **文档闭环**:在每个任务阶段完成后,主动更新 `PRD.md` 或相关技术文档,标注进度并记录架构决策。Related Skills
multi-phase-orchestrator
Orchestrates complex software engineering tasks through a structured multi-phase workflow - Research → Brainstorm → Design → Plan → Execute → Review. Use this skill when the user asks you to implement a non-trivial feature, refactor, or change that benefits from upfront analysis before writing code. Activates on keywords like "orchestrate", "multi-phase", "plan and execute", or when a task is large enough to warrant structured decomposition.
microservices-orchestrator
Expert skill for designing, decomposing, and managing microservices architectures. Activates when users need help with microservices design, service decomposition, bounded contexts, API contracts, or transitioning from monolithic to microservices architectures.
metasploit-framework
This skill should be used when the user asks to "use Metasploit for penetration testing", "exploit vulnerabilities with msfconsole", "create payloads with msfvenom", "perform post-exp...
metalsmith-website-skills
Build static websites using Metalsmith's component-based architecture. Triggers on phrases like "build me a website", "create a landing page", "help me make a site". Includes JavaScript and CSS development standards.
meta-42cog
This skill should be used when initializing a new project with Cognitive Agile methodology. It automatically generates real.md (reality constraints) and cog.md (cognitive model) by scanning the project directory and identifying key patterns.
manage-tasks
Manage implementation tasks with sequential sub-steps within a plan
history-and-next-task-rules
Specifies the format for ending responses, including a summary of requirements, code written, source tree, and next task, applying to all files.
codex-orchestrator
Orchestrate OpenAI Codex CLI with specialized subagents for code review, debugging, architecture analysis, security audits, refactoring, and documentation. This skill should be used when delegating focused development tasks to Codex subagents (gpt-5.3-codex, gpt-5.3-codex-spark, gpt-5.2) via AGENTS.md persona injection.
claude-code-task
Run Claude Code tasks in background with automatic result delivery. Use for coding tasks, research in codebase, file generation, complex automations. Zero OpenClaw tokens while Claude Code works.
bio-metagenomics-abundance
Species abundance estimation using Bracken with Kraken2 output. Redistributes reads from higher taxonomic levels to species for more accurate estimates. Use when accurate species-level abundances are needed from Kraken2 classification output.
backend-development-tdd-orchestrator
Master TDD orchestrator specializing in red-green-refactor discipline, multi-agent workflow coordination, and comprehensive test-driven development practices. Enforces TDD best practices across teams with AI-assisted testing and modern frameworks. Use PROACTIVELY for TDD implementation and governance. Use when: the task directly matches tdd orchestrator responsibilities within plugin backend-development. Do not use when: a more specific framework or task-focused skill is clearly a better match.
ai-ad-flow-orchestrator
No description provided.