BYAN v2.3.2 - Build Your AI Network

🏛️ Intelligent AI Agent Ecosystem | Powered by Merise Agile + TDD + 64 Mantras

Create custom AI agents through intelligent interviews + Hermes Universal Dispatcher for intelligent routing across 35+ specialized agents.

🎯 New in v2.3.2: Hermes - Universal Dispatcher

One entry point to rule them all 🏛️

@hermes          # → Menu-driven navigation to 35+ agents
@hermes @dev     # → Direct invocation  
@hermes rec créer API backend  # → Smart routing
@hermes pipe feature complète  # → Multi-agent pipelines

Features:

• 🎯 Smart Routing: Describe your task → Hermes recommends best agents
• 🔗 Pipelines: Pre-configured multi-agent workflows (Feature Complete, Bug Fix, etc.)
• 📋 Agent Directory: Browse 35+ agents organized by module (core, bmm, bmb, cis, tea)
• 💰 Cost Optimizer: 87.5% LLM cost savings (optional integration)
• 🌐 Multi-Platform: GitHub Copilot CLI, Claude Code, Codex
• 🪶 Node 12+ Compatible: Works on legacy servers

→ Full Hermes Guide

🚀 Quick Start

Install & Run

# Using npx (recommended - no installation needed)
npx create-byan-agent

# Or install globally
npm install -g create-byan-agent

# Then run
create-byan-agent

Programmatic Usage

const ByanV2 = require('create-byan-agent');

const byan = new ByanV2({
  maxQuestions: 12,
  outputDir: './_byan-output'
});

await byan.startSession();

// Interactive interview (12 questions)
while (!byan.isComplete()) {
  const question = byan.getNextQuestion();
  const answer = await getUserInput(question.text);
  await byan.submitResponse(answer);
}

// Generate agent profile
const profile = await byan.generateProfile();
console.log('Agent created:', profile.filePath);

✨ What's New in v2.1.0

4 New BMAD Modules

📚 Glossary Builder - Auto-builds domain vocabularies

• Auto-triggers for complex domains (ecommerce, finance, healthcare)
• Validates definition clarity (min 70%)
• Suggests related concepts

🔍 Five Whys Analyzer - Root cause analysis

• Detects pain points automatically
• 5 sequential WHY questions
• Categorizes causes (technical/process/people/resource)

👂 Active Listener - Intelligent response processing

• Auto-reformulation every 3rd response
• Validates confirmations
• Session summaries with insights

✅ Mantras Validator - Quality validation

• 64 BMAD/IA mantras
• Compliance scoring (target: 80%+)
• Category-based validation

417 new tests | 100% passing | 95%+ coverage | < 10% overhead

🎯 Features

Core Capabilities

• Intelligent 4-Phase Interview: Context → Business → Agent Needs → Validation
• Automatic Profile Generation: Creates .md agent files with YAML frontmatter
• Built-in Validation: Validates against GitHub Copilot CLI requirements
• Template System: Flexible templates with placeholder resolution
• Quality Enforcement: Zero emoji policy, clean code principles
• State Machine Workflow: INTERVIEW → ANALYSIS → GENERATION → COMPLETED

Advanced BMAD Features (v2.1.0)

• Domain Glossaries: Auto-build glossaries with clarity validation
• Root Cause Analysis: 5 Whys technique for pain point analysis
• Active Listening: Intelligent reformulation and validation
• Mantra Validation: 64 mantras compliance checking

100% Backwards Compatible

All v2.0.0 code works unchanged in v2.1.0. BMAD features are opt-in.

📖 Usage Examples

Example 1: Basic Agent Creation

const ByanV2 = require('create-byan-agent');

async function createAgent() {
  const byan = new ByanV2();
  await byan.startSession();
  
  // Answer 12 interview questions
  await byan.submitResponse('code-review-assistant');
  await byan.submitResponse('Reviews code for bugs and best practices');
  // ... 10 more responses
  
  const profile = await byan.generateProfile();
  console.log('✅ Agent created');
}

Example 2: With BMAD Features

const ByanV2 = require('create-byan-agent');

async function createWithBMAD() {
  const byan = new ByanV2({
    bmad_features: {
      glossary_builder: { enabled: true },
      five_whys: { enabled: true },
      active_listener: { enabled: true },
      mantras_validator: { enabled: true }
    }
  });
  
  await byan.startSession();
  
  // Glossary auto-triggers for complex domains
  const glossary = byan.startGlossary('ecommerce');
  byan.addConcept('Order', 'Customer purchase request...');
  
  // Pain point detection
  const detection = byan.detectPainPoints('Slow checkout');
  if (detection.needsWhys) {
    const question = byan.askWhy();
    // ... 5 Whys analysis
  }
  
  // Generate and validate
  const profile = await byan.generateProfile();
  const validation = byan.validateAgent(profile.content);
  
  console.log(`Score: ${validation.score * 100}%`);
}

Example 3: Validate Existing Agent

const ByanV2 = require('create-byan-agent');
const fs = require('fs');

const byan = new ByanV2({
  bmad_features: {
    mantras_validator: { enabled: true }
  }
});

const agentContent = fs.readFileSync('my-agent.md', 'utf-8');
const validation = byan.validateAgent(agentContent);

console.log(`Compliance: ${validation.score * 100}%`);
console.log(`Compliant: ${validation.compliant.length}/64`);
console.log(`Non-compliant: ${validation.nonCompliant.length}/64`);

🔧 Configuration

Basic Configuration

const byan = new ByanV2({
  maxQuestions: 12,
  outputDir: './_byan-output/bmb-creations',
  sessionId: 'my-session'
});

BMAD Configuration

const byan = new ByanV2({
  bmad_features: {
    glossary_builder: {
      enabled: true,
      auto_trigger_domains: ['ecommerce', 'finance', 'healthcare'],
      min_concepts: 5,
      clarity_threshold: 0.7
    },
    five_whys: {
      enabled: true,
      max_depth: 5,
      auto_trigger: true
    },
    active_listener: {
      enabled: true,
      reformulation_frequency: 3
    },
    mantras_validator: {
      enabled: true,
      min_compliance_score: 0.8
    }
  }
});

📊 Quality Metrics

• Tests: 1,308/1,308 passing (100%)
  • Core v2.0: 891 tests
  • BMAD v2.1: 417 tests
• Coverage: 95%+
• Performance: < 10% overhead vs v2.0.0
• Principles: KISS, DRY, SOLID, TDD
• Standards: Zero emojis (Mantra IA-23)

📚 API Reference

Main Class: ByanV2

Core Methods

• constructor(config) - Initialize BYAN
• startSession() - Start interview session
• getNextQuestion() - Get next interview question
• submitResponse(answer) - Submit answer
• generateProfile() - Generate agent profile
• isComplete() - Check if interview complete

Glossary Builder (BMAD)

• startGlossary(domain, options) - Initialize glossary
• addConcept(term, definition) - Add concept
• getConcepts() - Get all concepts
• getGlossary() - Get full glossary

Five Whys Analyzer (BMAD)

• detectPainPoints(response) - Detect pain points
• askWhy() - Get WHY question
• processWhyAnswer(answer) - Process answer
• getRootCause() - Get root cause analysis

Active Listener (BMAD)

• listen(input, context) - Process input with listening
• reformulate(input, style) - Reformulate input
• validateResponse(input) - Validate confirmation
• summarizeSession() - Generate session summary

Mantras Validator (BMAD)

• validateAgent(definition) - Validate against mantras
• generateComplianceReport() - Generate report
• validateMantra(mantraId, content) - Validate specific mantra
• getMantras(category) - Get mantras by category

🔄 Migration from v2.0.0

Good news: Zero breaking changes! Your v2.0.0 code works unchanged.

// v2.0.0 code - still works in v2.1.0
const ByanV2 = require('create-byan-agent');
const byan = new ByanV2();
await byan.startSession();
// ... your existing workflow

To use new BMAD features, simply enable them in config. See MIGRATION.md for details.

📖 Documentation

• Full Documentation
• CHANGELOG v2.1.0
• Migration Guide
• Manual Testing Guide
• BMAD Quick Reference

🤝 Contributing

Contributions welcome! Please read our contributing guidelines and submit pull requests to our GitHub repository.

👥 Contributors

Core Team

• Yan-Acadenice - Creator & Lead Developer

Special Contributors

• Wazadriano - Hermes Universal Dispatcher (v2.3.2)
  • Designed and documented the Hermes agent architecture
  • Smart routing rules and multi-agent pipelines
  • Complete integration with BYAN ecosystem

💝 Credits

Créé avec passion par Yan-Acadenice
Pour la communauté Acadenice

📄 License

MIT © Yan-Acadenice

🆘 Support

• Issues: GitHub Issues
• Repository: github.com/Yan-Acadenice/BYAN

🎯 Use Cases

For Solo Developers

Create specialized agents for your workflow without complex setup.

For Teams

Build consistent agent profiles following team standards and mantras.

For Agencies

Rapid agent prototyping with built-in quality validation.

For Open Source

Generate well-documented agent profiles for public repositories.

🌟 Highlights

• ✅ Zero Setup: Works immediately with npx
• ✅ Intelligent: Smart interview adapts to your answers
• ✅ Quality: Built-in validation against 64 mantras
• ✅ Fast: Create agents in 10-15 minutes
• ✅ Flexible: Use as CLI or programmatically
• ✅ Tested: 1,308 tests, 100% passing
• ✅ Compatible: Works with v2.0.0 code unchanged

Ready to create your first AI agent?

npx create-byan-agent

🚀 Let's build something amazing!