Skip to content
Development
Agent

BMAD-WORKFLOW

**BMAD (Business-Minded Agile Development)** - AI-driven agile development automation with role-based agents

From plugin
myclaude
2.7k23 skills23 agents13 commands2 hooks
Install
> /plugin marketplace add cexll/myclaude

How it fires

How this agent gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition โ†’
  • You can call itInvoke it directly when you want it.

Context preview

The summary Claude sees to decide when to auto-load this agent.

**BMAD (Business-Minded Agile Development)** - AI-driven agile development automation with role-based agents

Agent definition

BMAD-WORKFLOW.md

BMAD Workflow Complete Guide

> **BMAD (Business-Minded Agile Development)** - AI-driven agile development automation with role-based agents

๐ŸŽฏ What is BMAD?

BMAD is an enterprise-grade agile development methodology that transforms your development process into a fully automated workflow with 6 specialized AI agents and quality gates.

Core Principles

  • **Agent Planning**: Specialized agents collaborate to create detailed, consistent PRDs and architecture documents
  • **Context-Driven Development**: Transform detailed plans into ultra-detailed development stories
  • **Role Specialization**: Each agent focuses on specific domains, avoiding quality degradation from role switching

๐Ÿค– BMAD Agent System

Agent Roles

| Agent | Role | Quality Gate | Artifacts | |-------|------|--------------|-----------| | **bmad-po** (Sarah) | Product Owner - requirements gathering, user stories | PRD โ‰ฅ 90/100 | `01-product-requirements.md` | | **bmad-architect** (Winston) | System Architect - technical design, system architecture | Design โ‰ฅ 90/100 | `02-system-architecture.md` | | **bmad-sm** (Mike) | Scrum Master - task breakdown, sprint planning | User approval | `03-sprint-plan.md` | | **bmad-dev** (Alex) | Developer - code implementation, technical docs | Code completion | Implementation files | | **bmad-review** | Code Reviewer - independent review between Dev and QA | Pass/Risk/Fail | `04-dev-reviewed.md` | | **bmad-qa** (Emma) | QA Engineer - testing strategy, quality assurance | Test execution | `05-qa-report.md` |

๐Ÿš€ Quick Start

Command Overview

# Full BMAD workflow
/bmad-pilot "Build e-commerce checkout system with payment integration"

# Workflow: PO โ†’ Architect โ†’ SM โ†’ Dev โ†’ Review โ†’ QA

Command Options

# Skip testing phase
/bmad-pilot "Admin dashboard" --skip-tests

# Skip sprint planning (architecture โ†’ dev directly)
/bmad-pilot "API gateway implementation" --direct-dev

# Skip repository scan (not recommended)
/bmad-pilot "Add feature" --skip-scan

Individual Agent Usage

# Product requirements analysis only
/bmad-po "Enterprise CRM system requirements"

# Technical architecture design only
/bmad-architect "High-concurrency distributed system design"

# Orchestrator (can transform into any agent)
/bmad-orchestrator "Coordinate multi-agent complex project"

๐Ÿ“‹ Workflow Phases

Phase 0: Repository Scan (Automatic)

  • **Agent**: `bmad-orchestrator`
  • **Output**: `00-repository-context.md`
  • **Content**: Project type, tech stack, code organization, conventions, integration points

Phase 1: Product Requirements (PO)

  • **Agent**: `bmad-po` (Sarah - Product Owner)
  • **Quality Gate**: PRD score โ‰ฅ 90/100
  • **Output**: `01-product-requirements.md`
  • **Process**:

1. PO generates initial PRD 2. System calculates quality score (100-point scale) 3. If < 90: User provides feedback โ†’ PO revises โ†’ Recalculate 4. If โ‰ฅ 90: User confirms โ†’ Save artifact โ†’ Next phase

Phase 2: System Architecture (Architect)

  • **Agent**: `bmad-architect` (Winston - System Architect)
  • **Quality Gate**: Design score โ‰ฅ 90/100
  • **Output**: `02-system-architecture.md`
  • **Process**:

1. Architect reads PRD + repo context 2. Generates technical design document 3. System calculates design quality score 4. If < 90: User provides feedback โ†’ Architect revises 5. If โ‰ฅ 90: User confirms โ†’ Save artifact โ†’ Next phase

Phase 3: Sprint Planning (SM)

  • **Agent**: `bmad-sm` (Mike - Scrum Master)
  • **Quality Gate**: User approval
  • **Output**: `03-sprint-plan.md`
  • **Process**:

1. SM reads PRD + Architecture 2. Breaks down tasks with story points 3. User reviews sprint plan 4. User confirms โ†’ Save artifact โ†’ Next phase

  • **Skip**: Use `--direct-dev` to skip this phase

Phase 4: Development (Dev)

  • **Agent**: `bmad-dev` (Alex - Developer)
  • **Quality Gate**: Code completion
  • **Output**: Implementation files
  • **Process**:

1. Dev reads all previous artifacts 2. Implements features following sprint plan 3. Creates or modifies code files 4. Completes implementation โ†’ Next phase

Phase 5: Code Review (Review)

  • **Agent**: `bmad-review` (Independent Reviewer)
  • **Quality Gate**: Pass / Pass with Risk / Fail
  • **Output**: `04-dev-reviewed.md`
  • **Process**:

1. Review reads implementation + all specs 2. Performs comprehensive code review 3. Generates review report with status:

  • **Pass**: No issues, proceed to QA
  • **Pass with Risk**: Non-critical issues noted
  • **Fail**: Critical issues, return to Dev

4. Updates sprint plan with review findings

**Enhanced Review (Optional)**:

  • Use GPT-5 via Codex CLI for deeper analysis
  • Set via `BMAD_REVIEW_MODE=enhanced` environment variable

Phase 6: Quality Assurance (QA)

  • **Agent**: `bmad-qa` (Emma - QA Engineer)
  • **Quality Gate**: Test execution
  • **Output**: `05-qa-report.md`
  • **Process**:

1. QA reads implementation + review + all specs 2. Creates targeted test strategy 3. Executes tests 4. Generates QA report 5. Workflow complete

  • **Skip**: Use `--skip-tests` to skip this phase

๐Ÿ“Š Quality Scoring System

PRD Quality (100 points)

  • **Business Value** (30): Clear value proposition, user benefits
  • **Functional Requirements** (25): Complete, unambiguous requirements
  • **User Experience** (20): User flows, interaction patterns
  • **Technical Constraints** (15): Performance, security, scalability
  • **Scope & Priorities** (10): Clear boundaries, must-have vs nice-to-have

Architecture Quality (100 points)

  • **Design Quality** (30): Modularity, maintainability, clarity
  • **Technology Selection** (25): Appropriate tech stack, justification
  • **Scalability** (20): Growth handling, performance considerations
  • **Security** (15): Authentication, authorization, data protection
  • **Feasibility** (10): Realistic implementation, resource alignment

Review Status (3 levels)

  • **Pass**: No critical issues, code meets standards
  • **Pass with R
Read more
Ships withmyclaude

AI-powered development automation with multi-backend execution (Codex/Claude/Gemini/OpenCode)

Get the whole plugin