ldy26098
/
BMAD-METHOD
mirror of https://github.com/bmad-code-org/BMAD-METHOD.git
1
0
Fork 0
Code Packages Projects Releases Wiki Activity
BMAD-METHOD/README.md

7.6 KiB
Raw Blame History

BMAD METHOD - Build, Manage, Adapt & Deliver

A comprehensive Agent-based software development methodology that orchestrates specialized AI personas through the complete software lifecycle. The BMAD Method transforms how teams approach product development by providing memory-enhanced, quality-enforced workflows that adapt and improve over time.

What is BMAD?

BMAD is more than a workflow—it's an intelligent development companion that:

  • 🎭 Orchestrates specialized AI personas for every development role
  • 🧠 Learns from experience through integrated memory systems
  • ✅ Enforces quality standards with zero-tolerance for anti-patterns
  • 🔄 Adapts to your patterns becoming more effective over time
  • 🤝 Enables collaboration through multi-persona consultations

Key Components

  • 🎭 Specialized Personas - Expert agents for PM, Architect, Dev, QA, and more
  • 📋 Smart Task System - Context-aware task execution with quality gates
  • ✅ Quality Enforcement - Automated standards compliance and validation
  • 📝 Templates - Standardized document templates for consistent deliverables
  • 🧠 Memory Integration - Persistent learning and context management via OpenMemory MCP
  • ⚡ Performance Optimization - Smart caching and resource management

Orchestrator Variations

The BMAD Method includes two orchestrator implementations, each optimized for different contexts:

IDE Orchestrator (Primary)

Files: bmad-agent/ide-bmad-orchestrator.md & bmad-agent/ide-bmad-orchestrator.cfg.md

Purpose: Optimized for IDE integration with comprehensive memory enhancement and quality enforcement

Key Features:

  • Memory-enhanced context continuity
  • Proactive intelligence and pattern recognition
  • Multi-persona consultation mode
  • Integrated quality enforcement framework
  • Performance optimization for IDE environments

Best For: Active development in IDE environments where memory persistence and quality enforcement are critical

Web Orchestrator (Alternative)

Files: bmad-agent/web-bmad-orchestrator-agent.md & bmad-agent/web-bmad-orchestrator-agent.cfg.md

Purpose: Streamlined for web-based or lightweight environments

Key Features:

  • Simplified persona management
  • Basic task orchestration
  • Minimal resource footprint
  • Web-friendly command structure

Best For: Web interfaces, demos, or resource-constrained environments

Choosing an Orchestrator

  • Use IDE Orchestrator for full-featured development with memory and quality enforcement
  • Use Web Orchestrator for lightweight deployments or web-based interfaces
  • Both orchestrators share the same persona and task definitions for consistency

Key Features

🧠 Memory-Enhanced Development

  • Persistent Learning: Remembers decisions, patterns, and outcomes across sessions
  • Proactive Intelligence: Warns about potential issues based on past experiences
  • Context-Rich Handoffs: Smooth transitions between personas with full historical context
  • Pattern Recognition: Identifies and suggests successful approaches from past projects

✅ Quality Enforcement Framework

  • Zero-Tolerance Anti-Patterns: Automated detection and prevention of poor practices
  • Ultra-Deep Thinking Mode (UDTM): Systematic multi-angle analysis for critical decisions
  • Quality Gates: Mandatory checkpoints before phase transitions
  • Brotherhood Reviews: Honest, specific peer feedback requirements
  • Evidence-Based Decisions: All choices backed by data and validation

🎭 Specialized Personas

Each persona is an expert in their domain with specific skills, tasks, and quality standards:

  • PM (Product Manager): Market research, requirements, prioritization
  • Architect: System design, technical decisions, patterns
  • Dev: Implementation with quality compliance
  • QA/Quality Enforcer: Standards enforcement, validation
  • SM (Scrum Master): Story creation, sprint management
  • Analyst: Research, brainstorming, documentation
  • PO (Product Owner): Validation, acceptance, delivery

🔄 Intelligent Workflows

  • Adaptive Recommendations: Suggests next steps based on context
  • Multi-Persona Consultations: Coordinate multiple experts for complex decisions
  • Workflow Templates: Pre-defined paths for common scenarios
  • Progress Tracking: Real-time visibility into project status

Getting Started

Quick Start (IDE)

  1. Copy the BMAD agent folder to your project
  2. Open bmad-agent/ide-bmad-orchestrator.md in your AI assistant
  3. The orchestrator will initialize and guide you through available commands
  4. Start with /start to begin a new session

Quick Start (Web)

  1. Copy the BMAD agent folder to your web project
  2. Load bmad-agent/web-bmad-orchestrator-agent.md in your interface
  3. Use web-friendly commands to interact with personas
  4. Begin with /help to see available options

Core Commands

  • /start - Initialize a new session
  • /status - Check current state and active persona
  • /[persona] - Switch to a specific persona (e.g., /pm, /dev)
  • /consult - Start multi-persona consultation
  • /memory-status - View memory integration status
  • /help - Get context-aware assistance

Example Workflow

# Starting a new feature
/start
/pm analyze "Payment processing feature"
> PM analyzes market, creates requirements with UDTM

/architect design
> Architect creates technical design with quality gates

/consult pm, architect, dev
> Multi-persona consultation validates approach

/sm create-stories
> SM creates quality-validated user stories

/dev implement STORY-001
> Dev implements with anti-pattern detection

/quality validate
> Quality enforcer runs comprehensive validation

Project Structure

bmad-agent/
├── personas/           # Persona definitions with quality standards
├── tasks/             # Executable task definitions
├── quality-tasks/     # Quality-specific validation tasks
├── templates/         # Document templates
├── checklists/        # Validation checklists
├── memory/            # Memory integration guides
├── workflows/         # Standard workflow definitions
├── config/            # Performance and system configuration
└── orchestrators/     # IDE and Web orchestrator files

Memory System Integration

BMAD integrates with OpenMemory MCP for persistent intelligence:

  • Automated Learning: Captures decisions, patterns, and outcomes
  • Search & Retrieval: Finds relevant past experiences
  • Pattern Recognition: Identifies successful approaches
  • Continuous Improvement: Gets smarter with each use

Setup: Follow the OpenMemory MCP Setup Guide to enable advanced memory features.

Quality Metrics

The framework tracks comprehensive quality metrics:

  • Code coverage requirements (>90%)
  • Technical debt ratios (<5%)
  • Anti-pattern detection rates
  • UDTM compliance scores
  • Brotherhood review effectiveness
  • Evidence-based decision percentages

Contributing

We welcome contributions! Please see our Contributing Guide for details on:

  • Code standards and quality requirements
  • Persona development guidelines
  • Task creation best practices
  • Memory integration patterns

Documentation

License

MIT License

Thank you and enjoy building amazing software with BMAD!

- BMad