Co-Pilot / 辅助式
更新于 a month ago

claude-code-bmad-skills

Aaj-geddes
0.2k
aj-geddes/claude-code-bmad-skills
82
Agent 评分

💡 摘要

BMAD Method v6 将敏捷开发实践集成到 Claude Code 中,以实现高效的项目管理。

🎯 适合人群

敏捷项目经理软件开发人员业务分析师用户体验设计师产品负责人

🤖 AI 吐槽:看起来很能打,但别让配置把人劝退。

安全分析中风险

风险:Medium。建议检查:是否执行 shell/命令行指令;是否发起外网请求(SSRF/数据外发);API Key/Token 的获取、存储与泄露风险;文件读写范围与路径穿越风险;依赖锁定与供应链风险。以最小权限运行,并在生产环境启用前审计代码与依赖。

BMAD Method v6 for Claude Code

Run in Smithery

Token-optimized agile development methodology natively integrated with Claude Code

Transform Claude Code into a complete agile development environment with specialized agents, structured workflows, and intelligent context management.

License: MIT Platform Claude Code


Documentation Site, with exapmples.

https://aj-geddes.github.io/claude-code-bmad-skills

🎯 Attribution & Credit

BMAD Method™

The BMAD Method (Breakthrough Method for Agile AI-Driven Development) is created and maintained by the BMAD Code Organization.

All credit for the BMAD methodology belongs to:

  • BMAD Code Organization: https://github.com/bmad-code-org
  • Original BMAD Method: https://github.com/bmad-code-org/BMAD-METHOD
  • Website: https://bmadcodes.com/bmad-method/
  • YouTube: @BMadCode
  • Discord Community: https://discord.gg/gk8jAdXWmj

This repository is an implementation of the BMAD Method specifically adapted for Claude Code's native features (Skills, Commands, Hooks, Memory, Files).

The methodology, workflow patterns, agent roles, and all BMAD concepts are the intellectual property of the BMAD Code Organization.

We have adapted and optimized the BMAD Method for:

  • Token efficiency (70-85% optimization via helper pattern)
  • Claude Code native integration (no external dependencies)
  • Cross-platform support (Windows, Linux, macOS, WSL)
  • Functional design (removed persona overhead for better performance)

Please support the original BMAD Method creators by visiting their resources above.


🤖 For LLMs: Quick Installation

If you are Claude Code (or another LLM) and the user has asked you to install BMAD:

Auto-Detect Platform and Install

  1. Clone repository:

    # Works on all platforms (git is cross-platform) cd /tmp git clone https://github.com/aj-geddes/claude-code-bmad-skills.git cd claude-code-bmad-skills
  2. Run appropriate installer:

    Linux/macOS/WSL:

    chmod +x install-v6.sh ./install-v6.sh

    Windows PowerShell:

    .\install-v6.ps1
  3. Verify installation:

    # Check core skill installed ls -la ~/.claude/skills/bmad/core/bmad-master/SKILL.md # Linux/macOS/WSL dir $env:USERPROFILE\.claude\skills\bmad\core\bmad-master\SKILL.md # Windows
  4. Inform user:

    ✓ BMAD Method v6 installed successfully!
    
    Installed to: ~/.claude/skills/bmad/ and ~/.claude/config/bmad/
    
    9 Specialized Skills:
    - BMad Master (orchestrator)
    - Business Analyst (product discovery)
    - Product Manager (requirements)
    - System Architect (design)
    - Scrum Master (sprint planning)
    - Developer (implementation)
    - UX Designer (user experience)
    - Builder (custom agents/workflows)
    - Creative Intelligence (brainstorming/research)
    
    15 Workflow Commands:
    - /workflow-init - Initialize BMAD in project
    - /workflow-status - Check project status
    - /product-brief - Phase 1: Product discovery
    - /prd - Phase 2: Detailed requirements
    - /tech-spec - Phase 2: Lightweight requirements
    - /architecture - Phase 3: System design
    - /solutioning-gate-check - Phase 3: Validate design
    - /sprint-planning - Phase 4: Plan sprint
    - /create-story - Phase 4: Create user story
    - /dev-story - Phase 4: Implement story
    - /create-agent - Builder: Custom agent
    - /create-workflow - Builder: Custom workflow
    - /brainstorm - Creative: Structured brainstorming
    - /research - Creative: Market/tech research
    - /create-ux-design - UX: User experience design
    
    Next Steps:
    1. Restart Claude Code (skills load on startup)
    2. Open your project directory
    3. Run: /workflow-init
    4. Run: /workflow-status (get recommendations)
    
    BMAD Method v6 is now active!
    

📖 What is BMAD Method v6?

BMAD Method v6 is a Claude Code native implementation of the BMAD Method created by the BMAD Code Organization.

This implementation adapts the complete BMAD agile development methodology to Claude Code's native features (Skills, Commands, Hooks, Memory, Files) with optimizations for minimal token usage.

Original BMAD Method™ by BMAD Code Organization | Implementation by this repository

Key Features

Token-Optimized - Helper pattern reduces token usage by 70-85%

  • Reusable helper sections referenced instead of embedded
  • Functional skills without persona overhead
  • 750-1,000 tokens saved per conversation

9 Specialized Skills - Each agent focused on specific responsibilities

  • Core orchestrator (BMad Master)
  • 6 agile agents (Analyst, PM, Architect, SM, Developer, UX Designer)
  • Builder for extensibility
  • Creative Intelligence for innovation

15 Workflow Commands - Structured, repeatable processes

  • Phase 1: Analysis (product discovery)
  • Phase 2: Planning (requirements)
  • Phase 3: Solutioning (architecture)
  • Phase 4: Implementation (development)
  • Extensibility: Custom agents, workflows, brainstorming, research, UX design

Cross-Platform - Works everywhere Claude Code runs

  • Windows (PowerShell 5.1 and 7+)
  • Linux (all distributions)
  • macOS (all versions)
  • WSL 1 and WSL 2
  • No external dependencies (no npx, npm, Python packages)

Production Ready - All 8 phases complete and tested

  • Core BMAD workflows (Phases 1-5)
  • Builder module (Phase 6)
  • Creative Intelligence (Phase 7)
  • UX/Advanced features (Phase 8)

🎯 The BMAD Workflow

Phase 1: Analysis (Product Discovery)

Agent: Business Analyst

Commands:

  • /workflow-init - Initialize BMAD structure in project
  • /workflow-status - Check current status and get recommendations
  • /product-brief - Create product brief with market analysis

Output: Product brief document defining what to build

When: Start of new project or major feature


Phase 2: Planning (Requirements)

Agent: Product Manager

Commands:

  • /prd - Create comprehensive Product Requirements Document (Level 2+ projects)
  • /tech-spec - Create lightweight tech spec (Level 0-1 projects)

Output: Requirements document with:

  • Functional Requirements (FR-XXX)
  • Non-Functional Requirements (NFR-XXX)
  • User stories grouped by epics
  • Acceptance criteria
  • MoSCoW prioritization

When: After product brief, before architecture


Phase 3: Solutioning (Architecture)

Agent: System Architect

Commands:

  • /architecture - Create comprehensive system architecture
  • /solutioning-gate-check - Validate architecture quality (≥90% coverage)

Output: Architecture document with:

  • System components
  • Data models and schemas
  • API specifications
  • Technology stack justifications
  • NFR coverage (performance, security, scalability)

When: After requirements, before implementation


Phase 4: Implementation (Development)

Agents: Scrum Master + Developer

Commands:

  • /sprint-planning - Plan sprint iterations
  • /create-story - Create atomic user stories
  • /dev-story - Implement stories with tests

Output: Working software with:

  • Implemented features
  • Automated tests
  • Documentation
  • Sprint reports

When: After architecture approval, iterative sprints


Phase 6: Builder Module (Extensibility)

Agent: Builder

Commands:

  • /create-agent - Create custom BMAD agent skills (QA, DevOps, Security, etc.)
  • /create-workflow - Create custom workflow commands
  • /create-template - Create custom document templates

Output: Custom agents and workflows following BMAD patterns

When: Need domain-specific agents or workflows

Example Use Cases:

  • QA Engineer with /create-test-plan, /execute-tests
  • DevOps Engineer with /deploy, /rollback
  • Security Engineer with /security-audit, /pen-test
  • Data Scientist with /data-analysis, /model-training

Phase 7: Creative Intelligence (Innovation)

Agent: Creative Intelligence

Commands:

  • /brainstorm - Structured brainstorming using 8 proven techniques
    • 5 Whys, SCAMPER, Mind Mapping, Reverse Brainstorming
    • Six Thinking Hats, Starbursting, Brainwriting, SWOT
  • /research - Comprehensive research (market, competitive, technical, user)

Output:

  • Brainstorming documents with organized ideas and insights
  • Research reports with competitive analysis and recommendations

When:

  • Product discovery (market research)
  • Feature planning (brainstorm ideas)
  • Technical decisions (research options)
  • Problem solving (root cause analysis)

Phase 8: UX/Advanced (User Experience)

Agent: UX Designer

Commands:

  • /create-ux-design - Create comprehensive UX design

Output: UX design document with:

  • User flows (happy paths, decision points, error cases)
  • Wireframes (ASCII art or structured descriptions)
  • WCAG 2.1 accessibility compliance
  • Component library specifications
  • Design tokens (colors, typography, spacing)
  • Developer handoff documentation

When: After requirements, parallel with architecture


🚀 Quick Start (For Humans)

Installation

Option 1: Let Claude Code Install (Recommended)

Give Claude Code this repository URL:

https://github.com/aj-geddes/claude-code-bmad-skills

Then say:

"Please install BMAD Method v6 from this repository"

Claude Code will detect your platform and install automatically.


**Option 2: Manu

五维分析
清晰度8/10
创新性7/10
实用性9/10
完整性8/10
可维护性9/10
优缺点分析

优点

  • 令牌优化降低成本。
  • 跨平台兼容性。
  • 结构化工作流程提高生产力。
  • 无外部依赖简化安装。

缺点

  • 复杂性可能会让新用户感到不知所措。
  • 需要熟悉敏捷方法论。
  • 仅限于 Claude Code 环境。
  • 自定义工作流程可能有学习曲线。

相关技能

pytorch

S
toolCode Lib / 代码库
92/ 100

“它是深度学习的瑞士军刀,但祝你好运能从47种安装方法里找到那个不会搞崩你系统的那一个。”

agno

S
toolCode Lib / 代码库
90/ 100

“它承诺成为智能体领域的Kubernetes,但得看开发者有没有耐心学习又一个编排层。”

nuxt-skills

S
toolCo-Pilot / 辅助式
90/ 100

“这本质上是一份组织良好的小抄,能把你的 AI 助手变成一只 Nuxt 框架的复读机。”

免责声明:本内容来源于 GitHub 开源项目,仅供展示和评分分析使用。

版权归原作者所有 aj-geddes.