MDAN

MDAN — An AI-driven development platform with specialized agents for FinTech, DevOps/Azure, and Database Optimization. Built on the Better Agents framework for production-ready agent development.

100% free and open source. No paywalls. No gated content. Empowering teams with intelligent AI agents.

Why MDAN?

Traditional AI tools do the thinking for you, producing average results. MDAN agents act as expert collaborators who guide you through structured processes to bring out your best thinking in partnership with the AI.

• Better Agents Framework — Built on industry standards for agent development
• Scenario Testing — Every agent tested end-to-end for guaranteed behavior
• Versioned Prompts — Collaborative prompt management with version control
• Specialized Packs — FinTech, DevOps/Azure, and Database Optimization agents
• Scale-Adaptive — Automatically adjusts planning depth based on project complexity
• Complete Lifecycle — From brainstorming to deployment

📦 Installation

Install from npm (Recommended)

npm install -g mdan-ai

Or use npx without installing:

npx mdan-ai install

Install from GitHub

git clone https://github.com/khalilbenaz/MDANV2.git
cd MDANV2
npm install
npm link

Quick Start

After installation:

mdan-ai install

Follow the installer prompts, then open your AI IDE (Claude Code, Cursor, etc.) in your project folder.

Non-Interactive Installation (for CI/CD):

mdan-ai install --directory /path/to/project --modules mmm --tools claude-code --yes

🎯 Key Features

Better Agents Integration

MDAN is built on the Better Agents framework, ensuring:

• ✅ Scenario Tests — End-to-end testing for every agent
• ✅ Versioned Prompts — YAML-based prompt management
• ✅ Evaluations — Jupyter notebooks for performance measurement
• ✅ Observability — Full tracing and monitoring
• ✅ MCP Integration — Model Context Protocol for tool access

Specialized Packs

MDAN includes three specialized packs out of the box:

🏦 FinTech Pack

• Financial Analyst — Financial modeling, market analysis, portfolio optimization
• Compliance Officer — Regulatory compliance, audit, risk assessment
• Risk Manager — Risk analysis, hedging strategies, stress testing

☁️ DevOps/Azure Pack

• DevOps Engineer — CI/CD, infrastructure, monitoring
• Azure Specialist — Azure services, cloud architecture, deployment
• CI/CD Architect — Pipeline design, automation, quality gates

🗄️ Database Optimization Pack

• DBA — Database administration, backup, recovery
• Performance Tuner — Query optimization, index tuning, performance analysis
• Migration Specialist — Database migration, schema changes, data migration

Core Modules

• MMM — Business Model & Management (PM, Architect, Developer, QA, etc.)
• MMB — Module Builder (Agent, Module, Workflow builders)
• CIS — Creative Innovation Strategy (Brainstorming, Design Thinking, Innovation)
• TEA — Test Engineering & Architecture (Test strategy, automation, CI/CD)

📚 Documentation

User Documentation

• INSTALL.md — Installation guide
• USAGE.md — Usage guide
• AGENTS_LIST.md — Complete list of all agents

Technical Documentation

• AGENTS.md — Agent development guidelines
• ARCHITECTURE.md — Architecture documentation
• CONTRIBUTING.md — Contribution guide

Configuration

• CLAUDE.md — Claude Code configuration
• USER_STORIES_AND_TASKS.md — Development tasks

🏗️ Project Structure

MDAN/
├── app/                      # Agent code
│   ├── core/                 # Core agents
│   ├── mmm/                  # Main module (Business Model & Management)
│   ├── mmb/                  # Module Builder
│   ├── cis/                  # Creative Innovation Strategy
│   ├── tea/                  # Test Engineering & Architecture
│   └── packs/                # Specialized packs
│       ├── fintech/
│       ├── devops-azure/
│       └── db-optimization/
├── tests/
│   ├── scenarios/            # End-to-end scenario tests
│   └── evaluations/          # Jupyter notebooks for evaluations
├── prompts/                  # Versioned prompts
├── prompts.json              # Prompt registry
├── .mcp.json                 # MCP server configuration
├── AGENTS.md                 # Agent development guidelines
└── CLAUDE.md                 # Claude Code compatibility

🧪 Testing

MDAN uses the Scenario testing framework for end-to-end agent testing.

# Run all tests
npm test

# Run scenario tests
npm run test:scenarios

# Run evaluations
npm run test:evaluations

🤝 Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

When contributing to MDAN:

1. Follow the agent structure defined in AGENTS.md
2. Include tests for all new agents
3. Update documentation
4. Follow code quality standards
5. Get code review before merging

📦 Modules

MDAN extends with official modules for specialized domains:

| Module | Purpose | |--------|---------| | MMM | Core framework with 34+ workflows | | MMB | Create custom MDAN agents and workflows | | TEA | Risk-based test strategy and automation | | CIS | Innovation, brainstorming, design thinking | | FinTech Pack | Financial analysis, compliance, risk management | | DevOps/Azure Pack | CI/CD, Azure services, infrastructure | | DB Optimization Pack | Database tuning, optimization, migration |

🌟 Based On

MDAN is built on and inspired by:

• Better Agents — Standards for building agents
• BMAD-METHOD — AI-driven agile development
• Scenario — Agent testing framework
• LangWatch — LLM observability and evaluation

📄 License

MIT License — see LICENSE for details.

Last Updated: 28 February 2026 Version: 2.6.0