BMAD-MCP

Business-Minded Agile Development workflow orchestrator as an MCP (Model Context Protocol) server.

Complete agile development workflow: PO → Architect → SM → Dev → Review → QA

Interactive Requirements Gathering - Asks clarifying questions to ensure complete requirements

Dynamic Engine Selection - Uses Claude by default, dual-engine when needed

Content Reference System - Efficient token usage via file references

Human-Readable Task Names - Organizes by task name, not UUID

---

🎯 What is BMAD-MCP?

BMAD-MCP is a lightweight workflow orchestrator that manages the complete agile development process. It:

Manages workflow state (which stage you're in, what's needed next)
Dispatches role prompts (provides detailed prompts for each role)
Saves artifacts (PRD, architecture, code, reports)
Does NOT call LLMs (that's Claude Code's job)

🏗️ Architecture

code

User → Claude Code → bmad-mcp tool
                       ↓
            Returns: {
              stage: "po",
              role_prompt: "",
              engines: ["claude", "codex"],
              context: {...}
            }
                       ↓
Claude Code executes:
  - Calls Claude (with role_prompt)
  - Calls Codex MCP (with role_prompt)
                       ↓
Claude Code submits results → bmad-mcp
                       ↓
bmad-mcp: merges, scores, saves, advances to next stage

📋 Workflow Stages

Stage	Role	Engines	Description
PO	Product Owner	Claude + Codex	Requirements analysis (merge both)
Architect	System Architect	Claude + Codex	Technical design (merge both)
SM	Scrum Master	Claude	Sprint planning
Dev	Developer	Codex	Code implementation
Review	Code Reviewer	Codex	Code review
QA	QA Engineer	Codex	Testing and quality assurance

🚀 Quick Start

Installation (3 Steps)

bash

# Step 1: Install globally from npm
npm install -g bmad-mcp

# Step 2: Add to Claude Code
claude mcp add-json --scope user bmad '{"type":"stdio","command":"bmad-mcp"}'

# Step 3: Verify installation
bmad-mcp
# Expected output: "BMAD MCP Server running on stdio"

That's it! Restart Claude Code and you're ready to use BMAD workflow.

Usage Example

Simply tell Claude Code to use BMAD:

code

User: Use bmad-task to create a user authentication system

Claude Code will:
1. Start BMAD workflow (PO stage)
2. Generate Product Requirements Document (with interactive Q&A)
3. Generate System Architecture (with interactive Q&A)
4. Create Sprint Plan
5. Implement code (using Codex)
6. Perform code review
7. Run quality assurance tests

All artifacts saved to: .claude/specs/user-authentication-system/

Configuration Location

MCP configuration is automatically added to ~/.claude/config.json:

json

{
  "mcpServers": {
    "bmad": {
      "type": "stdio",
      "command": "bmad-mcp"
    }
  }
}

---

🔧 Advanced Installation

Option 1: NPM Install (Recommended)

bash

npm install -g bmad-mcp
claude mcp add-json --scope user bmad '{"type":"stdio","command":"bmad-mcp"}'

Option 2: Build from Source

bash

git clone https://github.com/cexll/bmad-mcp-server
cd bmad-mcp-server
npm install
npm run build
npm link  # Makes bmad-mcp globally available

# Add to Claude Code
claude mcp add-json --scope user bmad '{"type":"stdio","command":"bmad-mcp"}'

Verify Installation

bash

# Check if binary is available
which bmad-mcp
# Output: /usr/local/bin/bmad-mcp (or similar)

# Test the server directly
bmad-mcp
# Expected output: "BMAD MCP Server running on stdio"
# Press Ctrl+C to exit

# Restart Claude Code to load the configuration

Uninstall

bash

# Remove from Claude Code
claude mcp remove bmad

# Uninstall npm package
npm uninstall -g bmad-mcp

📖 Usage

Basic Workflow

typescript

// 1. Start workflow
const startResult = await callTool("bmad-task", {
  action: "start",
  cwd: "/path/to/your/project",
  objective: "Implement user login system"
});

const { session_id, task_name, role_prompt, engines } = JSON.parse(startResult.content[0].text);

// 2. Execute with engines
if (engines.includes("claude")) {
  const claudeResult = await callClaude(role_prompt);
}
if (engines.includes("codex")) {
  const codexResult = await callCodexMCP(role_prompt);
}

// 3. Submit results
await callTool("bmad-task", {
  action: "submit",
  session_id: session_id,
  stage: "po",
  claude_result: claudeResult,
  codex_result: codexResult
});

// 4. Confirm and proceed (unified: saves + advances to next stage)
await callTool("bmad-task", {
  action: "confirm",
  session_id: session_id,
  confirmed: true
});

Actions

`start` - Start a new workflow

json

{
  "action": "start",
  "cwd": "/path/to/project",
  "objective": "Project description"
}

Returns:

json

{
  "session_id": "uuid",
  "task_name": "project-description",
  "stage": "po",
  "state": "generating",
  "stage_description": "Product Owner - Requirements Analysis",
  "requires_user_confirmation": true,
  "interaction_type": "awaiting_generation",
  "user_message": "📋 **BMAD 工作流已启动**...",
  "role_prompt": "",
  "engines": ["claude"],
  "context": {...},
  "pending_user_actions": ["review_and_confirm_generation"]
}

`submit` - Submit stage results

json

{
  "action": "submit",
  "session_id": "uuid",
  "stage": "po",
  "claude_result": "...",
  "codex_result": "..."
}

Returns (if score >= 90):

json

{
  "session_id": "uuid",
  "stage": "po",
  "state": "awaiting_confirmation",
  "score": 92,
  "requires_user_confirmation": true,
  "interaction_type": "user_decision",
  "user_message": "✅ **PRD生成完成**\n质量评分：92/100...",
  "final_draft_summary": "...",
  "final_draft_file": ".bmad-task/temp/uuid/po_final_result_xxx.md",
  "pending_user_actions": ["confirm", "reject_and_refine"]
}

Returns (if score ",

"engines": ["claude"],

"previous_artifact": ".claude/specs/task-name/01-product-requirements.md",

"pending_user_actions": ["review_and_confirm_generation"]

}

code

#### `answer` - Answer clarification questions

{

"action": "answer",

"session_id": "uuid",

"answers": {

"q1": "Target users are enterprise B2B customers",

"q2": "Expected 10k concurrent users with ",

"engines": ["claude"],

"pending_user_actions": ["regenerate_with_answers"]

}

code

#### `approve` - Approve current stage (SM stage only)

{

"action": "approve",

"session_id": "uuid",

"approved": true

}

code

**Returns** (when entering Dev stage):

{

"session_id": "uuid",

"stage": "dev",

"state": "generating",

"requires_user_confirmation": true,

"interaction_type": "awaiting_generation",

"user_message": "✅ Sprint Plan 已批准\n\n正在进入下一阶段：Developer - Implementation\n\nSprint Plan 包含 3 个 Sprint：\n1. Sprint 1: 基础架构\n2. Sprint 2: 核心功能\n3. Sprint 3: 优化和完善\n\n⚠️ 重要：请明确指示开发范围...",

"role_prompt": "",

"engines": ["codex"],

"pending_user_actions": ["specify_sprint_scope_then_generate"]

}

code

**Important - Dev Stage Behavior**:
- After approving Sprint Plan, workflow enters Dev stage but **does NOT auto-start development**
- User must explicitly specify development scope:
  - **"开始开发" / "start development"** → Implements ALL sprints (default)
  - **"开发 Sprint 1" / "implement sprint 1"** → Implements only Sprint 1
- This ensures users have full control over what gets implemented and when

#### `status` - Query workflow status

{

"action": "status",

"session_id": "uuid"

}

code

**Returns**:

{

"session_id": "uuid",

"current_stage": "dev",

"current_state": "generating",

"stages": {...},

"artifacts": [...]

}

code

## 📁 File Structure

### Your Project

your-project/

├── .bmad-task/

│ ├── session-abc-123.json # Workflow state (with content references)

│ ├── task-mapping.json # Maps session_id → task_name

│ └── temp/

│ └── abc-123/ # Temporary content files

│ ├── po_claude_result_xxx.md

│ ├── po_codex_result_xxx.md

│ └── po_final_result_xxx.md

├── .claude/

│ └── specs/

│ └── implement-user-login/ # Task name (human-readable slug)

│ ├── 01-product-requirements.md

│ ├── 02-system-architecture.md

│ ├── 03-sprint-plan.md

│ ├── 04-dev-reviewed.md

│ └── 05-qa-report.md

└── src/

code

### Session State File

{

"session_id": "abc-123",

"task_name": "implement-user-login",

"cwd": "/path/to/project",

"objective": "Implement user login",

"current_stage": "dev",

"current_state": "generating",

"stages": {

"po": {

"status": "completed",

"claude_result_ref": {

"summary": "First 300 chars...",

"file_path": ".bmad-task/temp/abc-123/po_claude_result_xxx.md",

"size": 12450,

"last_updated": "2025-01-15T10:30:00Z"

"final_result_ref": {...},

"score": 92,

"approved": true

...

"artifacts": [".claude/specs/implement-user-login/01-product-requirements.md", ...]

}

code

## 🎨 Engine Configuration

### PO & Architect Stages (Dynamic Engine Selection)

- **Default**: Only Claude (single engine)
- **Dual Engine**: Enabled when objective contains "codex" or "使用 codex"
- If dual engine enabled:
  - Calls both Claude and Codex
  - Each generates independent solution
  - BMAD-MCP merges results:
    - If both ≥ 90: choose higher score
    - If one ≥ 90: choose that one
    - If both  B[PO Stage: Generate]
    B --> C{Has Questions?}
    C -->|Yes| D[Clarifying: User Answers]
    D --> E[Refining: Regenerate]
    E --> F{Score >= 90?}
    C -->|No| F
    F -->|No| C
    F -->|Yes| G[Awaiting Confirmation]
    G -->|confirm| H[Saved + Architect Stage]
    H --> I{Has Questions?}
    I -->|Yes| J[Clarifying: User Answers]
    J --> K[Refining: Regenerate]
    K --> L{Score >= 90?}
    I -->|No| L
    L -->|No| I
    L -->|Yes| M[Awaiting Confirmation]
    M -->|confirm| N[Saved + SM Stage]
    N -->|approve| O[Dev Stage]
    O --> P[Review Stage]
    P --> Q[QA Stage]
    Q --> R[Complete]

🛠️ Development

Project Structure

code

bmad-mcp/
├── src/
│   ├── index.ts              # Main MCP server
│   └── master-prompt.ts      # All role prompts
├── dist/                     # Compiled output
├── package.json
├── tsconfig.json
└── README.md

Build

bash

npm run build

Development Mode

bash

npm run dev  # Watch mode

Test Locally

bash

npm run build
node dist/index.js

📚 Master Orchestrator Design

All role prompts are embedded in a single master-prompt.ts file:

Centralized management: All roles in one place
Workflow definition: Clear stage sequence
Engine configuration: Which engines for each stage
Quality gates: Score thresholds and approval points

🤝 Integration with Codex MCP

When calling Codex for Dev/Review/QA stages:

typescript

// Claude Code calls Codex MCP
await callTool("codex", {
  prompt: role_prompt,  // From bmad-task
  model: "gpt-5",  // IMPORTANT: Use "gpt-5", NOT "gpt-5-codex"
  sandbox: "danger-full-access",
  "approval-policy": "on-failure"
});

⚙️ Configuration

Quality Thresholds

Defined in master-prompt.ts:

typescript

quality_gates: {
  po: { min_score: 90, approval_required: true },
  architect: { min_score: 90, approval_required: true },
  sm: { approval_required: true },
  dev: {},
  review: {},
  qa: {}
}

Artifact Filenames

typescript

artifacts: {
  po: "01-product-requirements.md",
  architect: "02-system-architecture.md",
  sm: "03-sprint-plan.md",
  dev: "code-implementation",
  review: "04-dev-reviewed.md",
  qa: "05-qa-report.md"
}

🔍 Troubleshooting

Server not starting

bash

# Check installation
which bmad-mcp

# Test directly
bmad-mcp

Tool name error

Important: The tool name is bmad-task, not bmad
Use callTool("bmad-task", {...}) in your code
Claude Code configuration should use bmad-task as the tool name

Session not found

Ensure .bmad-task/ directory has write permissions
Check session_id is correct
Verify cwd path is absolute

Scores not detected

Ensure generated content includes: Quality Score: X/100 or "quality_score": 92 in JSON
Check score format matches pattern (0-100)
Score ≥ 90 required for PO/Architect stages to advance

Clarification workflow issues

If you see state: "clarifying", user must answer questions via answer action
Do NOT auto-generate answers - wait for real user input
Check requires_user_confirmation: true before proceeding

📝 License

MIT

🙋 Support

Documentation: This README
Issues: GitHub issues
Reference: https://github.com/cexll/myclaude

---

Transform your development with BMAD - One workflow, complete agile process, quality assured.

Bmad Mcp Server

Documentation

BMAD-MCP

🎯 What is BMAD-MCP?

🏗️ Architecture

📋 Workflow Stages

🚀 Quick Start

Installation (3 Steps)

Usage Example

Configuration Location

🔧 Advanced Installation

Option 1: NPM Install (Recommended)

Option 2: Build from Source

Verify Installation

Uninstall

📖 Usage

Basic Workflow

Actions

start - Start a new workflow

submit - Submit stage results

🛠️ Development

Project Structure

Build

Development Mode

Test Locally

📚 Master Orchestrator Design

🤝 Integration with Codex MCP

⚙️ Configuration

Quality Thresholds

Artifact Filenames

🔍 Troubleshooting

Server not starting

Tool name error

Session not found

Scores not detected

Clarification workflow issues

📝 License

🙋 Support

Similar MCP

Waha Mcp

Wizzy Mcp Tmdb

Rijksmuseum Mcp

Mcp Server Playwright

Trending MCP

Playwright Mcp

Serena

Mcp Playwright

Mcp Server Cloudflare

Waha Mcp

Wizzy Mcp Tmdb

Rijksmuseum Mcp

Mcp Server Playwright

Playwright Mcp

Serena

Mcp Playwright

Mcp Server Cloudflare

`start` - Start a new workflow

`submit` - Submit stage results