Yuva AI

A powerful multi-agent AI framework for software development AND life/personal assistance with multi-LLM support, plugin system, workflow composer, and analytics.

What is this?

A universal AI assistant framework that turns your AI into a comprehensive multi-agent system:

• 34+ Specialized Agents — 11 for development + 23 for life/personal assistance
• Multi-LLM Support — Works with Claude, GPT, Gemini, and GitHub Copilot
• Plugin System — Create, install, and manage custom agents
• Workflow Composer — Define multi-agent pipelines with YAML
• Analytics Dashboard — Track agent usage and productivity (opt-in, local only)
• Session Persistence — Never lose progress, even across conversations
• Quality Gates — Built-in checklists, standards, and security reviews
• Safety Protocols — Crisis escalation, mandatory disclaimers for sensitive topics
• Zero Dependencies — Pure Node.js, installs in seconds

Quick Start

# Install and initialize
npx yuva-ai init

# Or use the short alias
npx yuva init

That's it. Open your project in Claude Code, Cursor, or VS Code and start working.

CLI Commands

# Setup
yuva init                  # Initialize in current directory
yuva init --force          # Overwrite existing files
yuva init --dry-run        # Preview what would be created
yuva upgrade               # Update agents and system files
yuva doctor                # Diagnose setup issues

# Agents
yuva list                  # List all installed agents
yuva list dev              # List development agents only
yuva list life             # List life/personal agents only
yuva add create <name>     # Create a custom agent
yuva add remove <name>     # Remove an agent

# Workflows
yuva workflow list         # List all workflows
yuva workflow create <n>   # Create a new workflow
yuva workflow show <n>     # Show workflow details
yuva workflow delete <n>   # Delete a workflow

# Configuration
yuva config                # Show current config
yuva config set <k> <v>    # Set a config value
yuva config get <key>      # Get a config value
yuva config reset          # Reset to defaults

# Multi-LLM
yuva llm list              # List supported LLMs
yuva llm use <name>        # Switch LLM platform
yuva llm detect            # Detect current LLM
yuva llm generate          # Generate configs for all LLMs

# Analytics
yuva status                # Show project status
yuva telemetry on/off      # Enable/disable analytics
yuva telemetry stats       # View usage statistics
yuva analytics             # View analytics dashboard

# General
yuva help                  # Show full help
yuva --version             # Show version

How It Works

┌─────────────────────────────────────────────────────────────┐
│                    CLAUDE.md (Controller)                    │
│  - Reads session state & user memory                        │
│  - Detects task type (Development vs Life)                  │
│  - Routes to appropriate agent                              │
│  - Enforces quality gates & safety protocols                │
└─────────────────────────────────────────────────────────────┘
                              │
              ┌───────────────┴───────────────┐
              ▼                               ▼
┌──────────────────────────┐    ┌──────────────────────────┐
│   DEVELOPMENT AGENTS     │    │      LIFE AGENTS         │
├──────────────────────────┤    ├──────────────────────────┤
│ Requirements │ Planner   │    │ Interview │ Finance     │
│ Execution    │ Tester    │    │ Doctor    │ Lawyer      │
│ Reviewer     │ Security  │    │ Career    │ Business    │
│ Debugger     │ Refactor  │    │ Teaching  │ Research    │
│ Continuity   │ Risk      │    │ Wellness  │ Stress      │
│ State Manager│           │    │ Travel    │ Writer      │
│              │           │    │ Fitness   │ Productivity│
│              │           │    │ + 10 more               │
└──────────────────────────┘    └──────────────────────────┘
              │                               │
              ▼                               ▼
┌──────────────────────────┐    ┌──────────────────────────┐
│  Standards & Checklists  │    │  Safety & Disclaimers    │
│  Quality Gates           │    │  Crisis Escalation       │
│  Workflow Pipelines      │    │  Professional Referrals  │
└──────────────────────────┘    └──────────────────────────┘

Available Agents

Development Agents (11)

Life Agents (23)

Multi-LLM Support

Works with 18+ AI platforms — commercial, open-source, and terminal-based:

Commercial

Open Source / Local

Terminal / CLI

# List all supported LLMs
yuva llm list

# List by category
yuva llm list open-source

# Switch LLM platform
yuva llm use ollama

# Generate configs for all platforms
yuva llm generate

# See suggested models for a platform
yuva llm models ollama

Plugin System

Create Custom Agents

# Create a new agent
yuva add create my-agent

# This creates:
# .aiautomations/prompts/my-agentagent.md (prompt template)
# .aiautomations/agents/my-agent.json (agent config)

# Remove an agent
yuva add remove my-agent

Agent Configuration

Each agent can have a config file in .aiautomations/agents/:

{
  "name": "my-agent",
  "version": "1.0.0",
  "type": "custom",
  "category": "custom",
  "triggers": ["keyword1", "keyword2"],
  "description": "What this agent does",
  "file": "my-agentagent.md",
  "dependencies": []
}

Workflow Composer

Define multi-agent pipelines with YAML:

# .aiautomations/workflows/blog-post.yml
name: blog-post
description: Research and write a blog post

steps:
  - agent: research
    task: "Research the topic thoroughly"
    output: research_notes

  - agent: writer
    task: "Write article based on research"
    input: research_notes
    output: draft

  - agent: reviewer
    task: "Review for quality and accuracy"
    input: draft
    output: final

Built-in Workflows

# Manage workflows
yuva workflow list
yuva workflow create my-pipeline
yuva workflow show my-pipeline

Analytics Dashboard

Track your usage patterns locally (opt-in):

# Enable analytics
yuva telemetry on

# View statistics
yuva telemetry stats

# View full dashboard with charts
yuva analytics

# All data stays local - nothing sent externally

Project Structure

your-project/
├── CLAUDE.md                    # Main controller
├── .aiautomations/
│   ├── config.json              # Project configuration
│   ├── prompts/                 # 34+ agent prompts
│   ├── protocols/               # System protocols
│   ├── standards/               # Code & design standards
│   ├── checklists/              # Quality checklists
│   ├── templates/               # Planning templates
│   ├── workflows/               # Multi-agent pipelines
│   └── agents/                  # Custom agent configs
├── .session/                    # Session persistence
│   ├── state.md
│   ├── log.md
│   └── next.md
└── .memory/                     # User memory
    ├── user.md
    └── context.md

Quality Gates

For Development

• Lint passes (no errors)
• Type check passes
• Tests pass (all green)
• Build succeeds
• Security reviewed
• Checklists completed
• Session updated

For Life Agents

• Question fully addressed
• Appropriate disclaimers included
• Professional referral when needed
• Crisis resources for emergencies

Safety Features

Crisis Escalation

Agents automatically provide crisis resources:

• Suicidal thoughts → 988, Crisis Text Line
• Domestic violence → National DV Hotline
• Sexual assault → RAINN
• Severe symptoms → "Seek immediate help"

Development

# Install dev dependencies
npm install

# Run tests
npm test

# Run tests with coverage
npm run test:coverage

# Lint code
npm run lint

# Fix lint issues
npm run lint:fix

# Check project health
npm run doctor

Works With

Commercial: Claude Code, Cursor, Windsurf, GitHub Copilot, Gemini CLI, Amazon Q, Cody Open Source: Ollama, LM Studio, Jan.ai, Continue.dev, Open Interpreter, LLM CLI, Tabby Terminal: OpenCode, Codex CLI, Kilo Code, Aider Models: Llama 3, Mistral, CodeLlama, DeepSeek Coder, Qwen, Phi-3, Gemma, StarCoder, GPT-4, Claude, Gemini

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Adding Agents

1. Create prompt in .aiautomations/prompts/
2. Create config in .aiautomations/agents/
3. Add to routing table in CLAUDE.md
4. Add tests
5. Submit PR

License

MIT License - see LICENSE file for details.

Links

• NPM Package
• GitHub Repository
• Landing Page
• Report Issues