Token导航 LogoToken导航

Agent Skill · 研究检索

bmad-orchestrator(bmad 协调器)

bmad-orchestrator 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

bmad-orchestrator 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证为 MIT,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 17,715。

总安装 6,449周安装 277GitHub Stars 404下载 2,260核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:bmad-orchestrator(bmad 协调器)
来源仓库:https://github.com/aj-geddes/claude-code-bmad-skills
仓库路径:skills/bmad-orchestrator
安装命令:
npx skills add https://github.com/aj-geddes/claude-code-bmad-skills --skill bmad-orchestrator
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/aj-geddes/claude-code-bmad-skills --skill bmad-orchestrator

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

bmad-orchestrator 是 BMAD 方法的核心协调器,用于管理开发工作流、跟踪状态并引导用户完成结构化开发阶段。适用于需要初始化 BMAD 项目、检查工作流进度或启动结构化开发方法的场景。通过 /workflow-init 或 /init 等指令触发,自动路由用户至相应开发阶段并提供建议。

安装前需确认权限范围和维护状态,注意可能涉及命令执行和文件读写操作。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/bmad-orchestrator

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/aj-geddes/claude-code-bmad-skills --skill bmad…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

bmad-orchestrator 怎么安装?

在终端执行 npx skills add https://github.com/aj-geddes/claude-code-bmad-skills --skill bmad-orchestra…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

bmad-orchestrator 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

bmad-orchestrator 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

bmad-orchestrator 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

bmad-orchestrator 用的人多吗?

公开数据是累计安装 6,449 次,近一周安装 277 次,下载 2,260 次,GitHub 404 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

bmad-orchestrator 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
bmad-orchestrator

bmad 协调器 · 本页

6,449404需要联网MIT通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

BMAD Orchestrator

Purpose: Core orchestrator for the BMAD Method (Breakthrough Method for Agile AI-Driven Development), managing workflows, tracking status, and routing users through structured development phases.

When to Use This Skill

Use this skill when:

  • User requests /workflow-init or /init - Initialize BMAD in a project
  • User requests /workflow-status or /status - Check progress and get recommendations
  • User mentions "BMAD setup" or "start BMAD workflow"
  • Project needs structured development methodology
  • Coordination between multiple development phases is required

Core Responsibilities

  1. Project Initialization - Set up BMAD directory structure and configuration
  2. Status Tracking - Monitor progress across 4 development phases
  3. Workflow Routing - Direct users to appropriate next steps based on project state
  4. Progress Management - Maintain workflow status and completion tracking

BMAD Method Overview

4 Development Phases

  1. Analysis (Optional) - Research, brainstorming, product brief
  2. Planning (Required) - PRD or Tech Spec based on project complexity
  3. Solutioning (Conditional) - Architecture design for medium+ projects
  4. Implementation (Required) - Sprint planning, stories, development

Project Levels

  • Level 0: Single atomic change (1 story) - Quick fixes, small tweaks
  • Level 1: Small feature (1-10 stories) - Single feature additions
  • Level 2: Medium feature set (5-15 stories) - Multiple related features
  • Level 3: Complex integration (12-40 stories) - System integrations
  • Level 4: Enterprise expansion (40+ stories) - Large-scale projects

Planning Requirements by Level:

  • Level 0-1: Tech Spec required, PRD optional/recommended
  • Level 2+: PRD required, Tech Spec optional
  • Level 2+: Architecture required

Available Commands

/workflow-init or /init

Initialize BMAD structure in the current project.

Steps:

  1. Create directory structure: bmad/ ├── config.yaml └── agent-overrides/ docs/ ├── bmm-workflow-status.yaml └── stories/.claude/commands/bmad/ (if not exists)
  2. Collect project information:

- Project name - Project type (web-app, mobile-app, api, game, library, other) - Project level (0-4)

  1. Create project config using config.template.yaml
  2. Create workflow status file with conditional requirements based on level:

- Use workflow-status template - Set PRD: required if level >= 2, else recommended - Set Tech-spec: required if level <= 1, else optional - Set Architecture: required if level >= 2, else optional

  1. Display initialization summary and recommend next workflow
  2. Offer to start recommended workflow

Example interaction:

User: /workflow-init

Orchestrator: I'll initialize BMAD for your project.
[Collects info, creates files...]

✓ BMAD Method initialized!

Project: MyApp
Type: web-app
Level: 2 (Medium feature set)

Configuration: bmad/config.yaml
Status tracking: docs/bmm-workflow-status.yaml

Recommended next step: Create Product Brief with /product-brief
This will help define your product vision and requirements.

Would you like to start with /product-brief?

/workflow-status or /status

Check project status and get recommendations for next steps.

Steps:

  1. Load project config (bmad/config.yaml)
  2. Load workflow status (docs/bmm-workflow-status.yaml)
  3. Determine current phase and next recommended workflow
  4. Display status with visual indicators
  5. Offer to execute recommended workflow

Status Indicators:

  • ✓ = Completed (shows file path)
  • ⚠ = Required but not started
  • → = Current phase
  • - = Optional/not required

Example interaction:

User: /status

Orchestrator: Checking your project status...

Project: MyApp (Web Application, Level 2)

✓ Phase 1: Analysis
  ✓ product-brief (docs/product-brief-myapp-2025-01-11.md)

→ Phase 2: Planning [CURRENT]
  ⚠ prd (required - NOT STARTED)
  - tech-spec (optional)

Phase 3: Solutioning
  - architecture (required)

Phase 4: Implementation
  - sprint-planning (required)

Recommended next step: Create PRD with /prd command

Would you like to run /prd to create your PRD?

If project not initialized:

  • Inform user BMAD not detected
  • Offer to run /workflow-init

Workflow Routing Logic

After determining project status, route users to specialized workflows:

  • Analysis workflows: /product-brief, /brainstorm, /research
  • Planning workflows: /prd, /tech-spec
  • UX workflows: /create-ux-design
  • Architecture workflows: /architecture
  • Sprint workflows: /sprint-planning, /create-story
  • Development workflows: /dev-story, /code-review

Recommendation logic:

  1. If no product-brief and project new → Recommend: /product-brief
  2. If product-brief complete, no PRD/tech-spec:

- Level 0-1 → Recommend: /tech-spec - Level 2+ → Recommend: /prd

  1. If PRD/tech-spec complete, no architecture, level 2+ → Recommend: /architecture
  2. If planning complete → Recommend: /sprint-planning
  3. If sprint active → Recommend: /create-story or /dev-story

See REFERENCE.md for detailed routing logic.

Configuration Files

Project Config (bmad/config.yaml)

project_name: "MyApp"
project_type: "web-app"  # web-app, mobile-app, api, game, library, other
project_level: 2         # 0-4
output_folder: "docs"
communication_language: "English"

Workflow Status (docs/bmm-workflow-status.yaml)

Tracks completion of each workflow with status values:

  • "optional" - Can be skipped
  • "recommended" - Strongly suggested
  • "required" - Must be completed
  • "{file-path}" - Completed (shows output file)
  • "skipped" - Explicitly skipped

See templates/config.template.yaml for full template.

Helper Scripts

Execute via Bash tool:

  • init-project.sh - Automated project initialization bash scripts/init-project.sh --name "MyApp" --type web-app --level 2
  • check-status.sh - Display current workflow status bash scripts/check-status.sh
  • validate-config.sh - Validate YAML configuration bash scripts/validate-config.sh bmad/config.yaml

See scripts documentation for details.

Error Handling

Config missing:

  • Suggest /workflow-init
  • Explain BMAD not initialized

Invalid YAML:

  • Show error location
  • Offer to fix or reinitialize

Template missing:

  • Use inline fallback
  • Log warning
  • Continue operation

Status file inconsistent:

  • Validate against project level
  • Offer to regenerate

Integration with Other Skills

This orchestrator coordinates with specialized BMAD skills:

  • business-analyst - Analysis phase workflows
  • product-manager - Planning phase workflows
  • system-architect - Architecture design
  • scrum-master - Sprint and story management
  • developer - Development workflows

When routing to these skills, pass context:

  • Current project config
  • Workflow status
  • Project level
  • Output folder location

Token Optimization

  • Use script automation for repetitive tasks
  • Reference REFERENCE.md for detailed logic
  • Load files only when needed
  • Keep status displays concise
  • Delegate detailed work to specialized skills

Subagent Strategy

This skill leverages parallel subagents to maximize context utilization (each agent has up to 1M tokens on Claude Sonnet 4.6 / Opus 4.6).

Workflow Status Check Workflow

Pattern: Fan-Out Research Agents: 3-4 parallel agents

AgentTaskOutput
Agent 1Check project config and validate structurebmad/outputs/config-status.md
Agent 2Analyze workflow status file and phase completionbmad/outputs/workflow-status.md
Agent 3Scan docs directory for completed artifactsbmad/outputs/artifacts-status.md
Agent 4Generate recommendations based on project levelbmad/outputs/recommendations.md

Coordination:

  1. Launch all agents with shared project context
  2. Each agent writes status findings to designated output
  3. Main context synthesizes results into unified status report
  4. Display visual status indicators and next steps

Project Initialization Workflow

Pattern: Parallel Section Generation Agents: 3 parallel agents

AgentTaskOutput
Agent 1Create directory structure and validate pathsbmad/outputs/directory-setup.md
Agent 2Generate project config from templatebmad/config.yaml
Agent 3Generate workflow status file with level-based requirementsdocs/bmm-workflow-status.yaml

Coordination:

  1. Gather project information from user (sequential)
  2. Launch parallel agents to create structures and configs
  3. Main context validates all outputs and displays summary

Example Subagent Prompt

Task: Analyze workflow status and determine current phase
Context: Read bmad/config.yaml and docs/bmm-workflow-status.yaml
Objective: Identify completed workflows, current phase, and required next steps
Output: Write analysis to bmad/outputs/workflow-status.md

Deliverables:
1. List of completed workflows with file paths
2. Current phase determination
3. Required vs optional next workflows
4. Blocking issues or missing dependencies

Constraints:
- Use project level to determine requirements
- Flag any inconsistencies in status file

Notes for Claude

  • This is the entry point for BMAD workflows
  • Always check if project is initialized before operations
  • Maintain phase-based progression (don't skip required phases)
  • Use TodoWrite for multi-step initialization
  • Keep responses focused and actionable
  • Hand off to specialized skills for detailed workflows
  • Update workflow status after completing workflows

Quick Reference

平台分布

Claude Code

28.43%
按下载量换算643

Antigravity

24.01%
按下载量换算543

OpenCode

18.73%
按下载量换算423

Cursor

15.22%
按下载量换算344

Gemini CLI

8.33%
按下载量换算188

windsurf

3.96%
按下载量换算89

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。

继续浏览同类 Skills