U_Agentix Agent Builder VS Code Extension

Visual agent creation and strategy development for AI trading

Build, test, and deploy intelligent trading agents directly from VS Code with drag-and-drop interface, battle simulation, and real-time performance tracking.

Features

Agent Explorer

57 Pre-Built Agents - Browse and customize legendary investor personas
Category Organization - Agents grouped by strategy type (Value, Momentum, Quantitative, etc.)
Performance Metrics - View win rates, Sharpe ratios, and battle history
Quick Actions - Edit, duplicate, export, or delete agents with one click

Agent Creation

Visual Wizard - Step-by-step agent creation with smart defaults
4 Strategy Templates - Momentum, Value, Contrarian, Quantitative
Custom Configuration - Fine-tune risk management, execution, and indicators
Code Generation - Export to TypeScript, Python, JSON, or Pine Script

Battle Simulator

Local Testing - Run battles before deploying to production
Multiple Scenarios - Bull, bear, sideways, volatile, or historical markets
Real-Time Results - See live rankings, equity curves, and trade history
Performance Analytics - Sharpe ratio, max drawdown, win rate, profit factor
Export Results - Save battles as JSON, CSV, or PDF reports

Strategy Builder (Phase 2)

Drag-and-Drop UI - Visual strategy design with components
Real-Time Preview - See strategy logic as you build
Backtest Integration - Test strategies on historical data
Code Export - Generate production-ready agent code

MCP Integration

Data Ops MCP - Load/save agent configurations and battle data
VS Code MCP - Generate code components and templates
Analytics MCP - Track performance and get ML recommendations
Deployment MCP - Deploy agents to staging or production

Installation

Requirements

VS Code 1.60.0 or higher
Node.js 16.x or higher
Git (for version control features)

Quick Install

Download Extension

cd /Users/haymanhymanb.t./Desktop/U_Agentix/u-agentix-agent-builder-extension
npm install
npm run build
npm run package:vsix

Install in VS Code
- Open VS Code
- Go to Extensions (Cmd+Shift+X)
- Click "..." menu > "Install from VSIX"
- Select u-agentix-agent-builder-0.1.0.vsix
Verify Installation
- Look for "Agent Builder" icon in activity bar (left sidebar)
- Click to open Agent Explorer

Manual Setup

# Clone or navigate to extension directory
cd u-agentix-agent-builder-extension

# Install dependencies
npm install

# Build extension
npm run compile

# Run tests
npm test

# Package for distribution
npm run package:vsix

Quick Start

1. Create Your First Agent

Cmd+Shift+P > Agent Builder: Create New Agent

Follow the wizard:

Enter agent name (e.g., "My Momentum Trader")
Select strategy type (Momentum, Value, Contrarian, Quantitative)
Choose category (Custom recommended for first agent)
Add description

Your agent is created and appears in Agent Explorer!

2. Customize Agent Configuration

Option A: Visual Editor

Right-click agent in tree view
Select "Edit Agent Configuration"
Modify JSON configuration
Save changes

Option B: Strategy Builder (Coming Phase 2)

Right-click agent
Select "Open Strategy Builder"
Drag-and-drop components
Connect indicators to actions

3. Run a Battle Simulation

Cmd+Shift+P > Agent Builder: Run Battle Simulation

Steps:

Select 2+ agents to compete
Choose market scenario (Bull, Bear, Sideways, Volatile)
Set duration (1 Hour, 1 Day, 1 Week, 1 Month)
Watch battle run with progress bar
View results with rankings and statistics

4. Export Agent to Code

Right-click agent > Export Strategy

Choose format:

TypeScript Class - Full agent implementation
Python Script - Python trading bot
JSON Config - Agent configuration file
Pine Script - TradingView strategy

Code opens in new editor ready to save and use!

Agent Explorer

Tree View Structure

Agent Builder
├── Legendary Investors (8)
│   ├── Value Oracle AI
│   ├── Growth Scout AI
│   ├── Fundamental Sage AI
│   └── ...
├── Modern Mavericks (8)
│   ├── Innovation Hunter AI
│   ├── Capital Maverick AI
│   └── ...
├── Quantitative Wizards (5)
│   ├── Quant Master AI
│   ├── Casino Quant AI
│   └── ...
├── Specialists (9)
├── Educational Personas (10)
└── Custom
    └── Your custom agents

Agent Details Panel

Click any agent to view:

Overview: Strategy, category, description
Performance: Return, Sharpe, drawdown, win rate
Risk Management: Position sizing, stop loss, take profit
Execution: Order types, slippage, commission
Backstory: Agent personality and approach

Right-click any agent:

View Agent Details
Edit Agent Configuration
Duplicate Agent
Export Strategy to Code
Delete Agent

Battle Simulator

Running Battles

Select Participants
- Minimum 2 agents required
- Select from all 57 agents
- Can battle same strategy types
- Duplicate agents allowed
Choose Scenario
- Bull Market: Strong uptrend (2024-style)
- Bear Market: Downturn and decline
- Sideways: Range-bound trading
- Volatile: High volatility environment
- Historical: Use actual historical data
Set Duration
- 1 Hour: Quick test
- 1 Day: Standard battle
- 1 Week: Medium-term analysis
- 1 Month: Long-term performance
View Results
- Winner announcement
- Full rankings table
- Performance statistics
- Trade-by-trade history

Battle Results

Battle Results

Winner: Quant Master AI
Return: 12.5%
Sharpe Ratio: 1.85
Max Drawdown: 5.2%
Win Rate: 65.0%
Total Trades: 47

Rankings:
1. Quant Master AI - 12.5%
2. Momentum Mike AI - 10.2%
3. Value Oracle AI - 8.7%
4. Contrarian Prophet AI - 6.3%

Export Options

JSON: Full battle data with trades
CSV: Rankings and statistics
PDF: Professional report (Phase 2)

Code Generation

TypeScript Agent Class

import { TradingAgent, Signal, MarketData } from '@u-agentix/core';

export class MyMomentumAgent extends TradingAgent {
  name = 'My Momentum Trader';
  strategy = 'momentum';

  async analyzeMarket(data: MarketData): Promise<Signal> {
    const rsi = this.calculateRSI(data, 14);
    const macd = this.calculateMACD(data, 12, 26, 9);

    if (rsi > 70 && macd.signal === 'buy') {
      return {
        action: 'buy',
        quantity: this.calculatePositionSize(data),
        reason: 'Strong momentum detected'
      };
    }

    return { action: 'hold' };
  }

  // Full implementation included...
}

Python Trading Script

from trading_agent import TradingAgent, Signal, MarketData

class MyMomentumAgent(TradingAgent):
    def __init__(self):
        self.name = 'My Momentum Trader'
        self.strategy = 'momentum'

    async def analyze_market(self, data: MarketData) -> Signal:
        rsi = self.calculate_rsi(data, 14)
        macd = self.calculate_macd(data, 12, 26, 9)

        if rsi > 70 and macd['signal'] == 'buy':
            return Signal(
                action='buy',
                quantity=self.calculate_position_size(data),
                reason='Strong momentum detected'
            )

        return Signal(action='hold')

Pine Script (TradingView)

//@version=5
indicator("My Momentum Trader", overlay=true)

// Strategy Parameters
fastMA = ta.sma(close, 10)
slowMA = ta.sma(close, 30)
rsi = ta.rsi(close, 14)

buySignal = ta.crossover(fastMA, slowMA) and rsi > 70
sellSignal = ta.crossunder(fastMA, slowMA) or rsi < 30

plotshape(buySignal, style=shape.triangleup, location=location.belowbar, color=color.green)
plotshape(sellSignal, style=shape.triangledown, location=location.abovebar, color=color.red)

Agent Templates

Momentum Strategy

Entry Conditions:

RSI > 70 (strong momentum)
MACD bullish crossover
Volume > 1.5x average

Exit Conditions:

RSI < 30 (momentum weakening)
MACD bearish crossover

Risk Management:

Stop Loss: 2%
Take Profit: 5%
Max Position: 10% of capital

Value Strategy

Entry Conditions:

P/E < 15
P/B < 1.5
Dividend Yield > 3%
25% margin of safety

Exit Conditions:

Price >= Intrinsic Value
Fundamentals deteriorate

Risk Management:

Stop Loss: 15%
Take Profit: Fair value
Position sizing based on discount

Contrarian Strategy

Entry Conditions:

Fear & Greed Index < 20
RSI < 30 (oversold)
Price 10%+ below 50 MA
High volatility

Exit Conditions:

Fear & Greed Index > 80
RSI > 70 (overbought)
Price 10%+ above 50 MA

Risk Management:

Stop Loss: 10%
Conservative position sizing
Scale in on dips

Quantitative Strategy

Entry Conditions:

Multi-factor score > 2σ
Statistical edge > 2%
Positive momentum + mean reversion

Exit Conditions:

Score < -2σ
Edge disappears

Risk Management:

Kelly Criterion position sizing
Dynamic stop loss
Probabilistic exits

MCP Integration

Data Ops MCP

// Load agent personas
const agents = await mcp.loadAgentPersonas({
  includeStrategies: true,
  includePerformance: true
});

// Get battle history
const battles = await mcp.getBattleHistory({
  timeframe: 'last-30-days',
  limit: 10
});

// Run battle simulation
const results = await mcp.simulateBattle({
  agents: ['agent1', 'agent2'],
  scenario: 'bull-market',
  duration: '1-month'
});

VS Code MCP

// Generate agent code
await mcp.generateAgentCode({
  name: 'MyAgent',
  template: 'momentum',
  includeTests: true
});

// Refactor existing code
await mcp.refactorAgentCode({
  filePath: 'agents/MyAgent.ts',
  refactorType: 'add-typescript'
});

Analytics MCP

// Get performance metrics
const performance = await mcp.getAgentPerformance({
  agentId: 'quant_master_ai',
  timeframe: 'all-time'
});

// Get strategy recommendations
const recommendations = await mcp.getRecommendedStrategies({
  agentType: 'momentum',
  maxResults: 5
});

Configuration

Extension Settings

Open VS Code Settings > Extensions > Agent Builder:

{
  "agentBuilder.agentPersonasPath": "/path/to/agent-personas",
  "agentBuilder.mcpDataOpsEnabled": true,
  "agentBuilder.mcpVSCodeEnabled": true,
  "agentBuilder.mcpAnalyticsEnabled": true,
  "agentBuilder.autoSaveOnBuild": true,
  "agentBuilder.showPerformanceMetrics": true,
  "agentBuilder.defaultStrategy": "momentum",
  "agentBuilder.battleSimulationDuration": "1-day"
}

Agent Configuration

{
  "strategy": {
    "type": "momentum",
    "parameters": {
      "rsiPeriod": 14,
      "macdFast": 12,
      "macdSlow": 26
    },
    "entryConditions": [
      { "indicator": "RSI", "operator": "gt", "value": 70 },
      { "indicator": "MACD", "operator": "cross-above", "value": 0 }
    ],
    "exitConditions": [
      { "indicator": "RSI", "operator": "lt", "value": 30 }
    ]
  },
  "riskManagement": {
    "maxPositionSize": 0.1,
    "stopLoss": 0.02,
    "takeProfit": 0.05,
    "maxDrawdown": 0.2,
    "positionSizing": "fixed"
  },
  "execution": {
    "orderType": "market",
    "timeInForce": "day",
    "slippage": 0.001,
    "commission": 0.001
  }
}

Commands

Command Palette

Access via Cmd+Shift+P (Mac) or Ctrl+Shift+P (Windows/Linux):

Agent Builder: Create New Agent - Create agent with wizard
Agent Builder: Run Battle Simulation - Battle 2+ agents
Agent Builder: Open Strategy Builder - Visual strategy editor
Agent Builder: Export Strategy to Code - Generate code
Agent Builder: Import Strategy from File - Load JSON config
Agent Builder: Refresh Agent List - Reload all agents

Keyboard Shortcuts

No default shortcuts. Add your own in VS Code Keyboard Shortcuts:

{
  "key": "cmd+shift+a",
  "command": "agentBuilder.createAgent"
},
{
  "key": "cmd+shift+b",
  "command": "agentBuilder.runBattle"
}

Development

Build from Source

# Clone repository
git clone https://github.com/u-agentix/agent-builder-extension.git
cd agent-builder-extension

# Install dependencies
npm install

# Build TypeScript
npm run compile

# Watch mode for development
npm run watch

# Run tests
npm test

# Package extension
npm run package:vsix

Project Structure

u-agentix-agent-builder-extension/
├── src/
│   ├── extension.ts              # Main entry point
│   ├── commands/
│   │   ├── createAgent.ts        # Create agent wizard
│   │   ├── runBattle.ts          # Battle simulator
│   │   └── exportStrategy.ts     # Code generation
│   ├── providers/
│   │   ├── agentTreeProvider.ts  # Agent tree view
│   │   └── ...
│   ├── services/
│   │   ├── mcpIntegration.ts     # MCP client
│   │   ├── agentService.ts       # Agent CRUD
│   │   ├── battleEngine.ts       # Battle simulation
│   │   └── ...
│   ├── webviews/
│   │   ├── strategyBuilder.ts    # Visual builder
│   │   ├── battleSimulator.ts    # Battle UI
│   │   └── ...
│   └── types/
│       ├── agent.ts              # Agent types
│       ├── strategy.ts           # Strategy types
│       └── battle.ts             # Battle types
├── templates/
│   ├── momentum-agent.template.ts
│   ├── value-agent.template.ts
│   ├── contrarian-agent.template.ts
│   └── quant-agent.template.ts
├── test/
│   └── extension.test.ts
├── package.json
├── tsconfig.json
├── webpack.config.js
└── README.md

Running Tests

# All tests
npm test

# Watch mode
npm run test:watch

# Coverage
npm run test:coverage

Adding New Templates

Create template file in templates/:

// templates/my-strategy.template.ts
import { TradingAgent } from '@u-agentix/core';

export class {{AGENT_CLASS_NAME}} extends TradingAgent {
  // Your strategy logic
}

Add to export command in commands/exportStrategy.ts
Update README with strategy description

Troubleshooting

Extension not activating

Solution: Check VS Code version (must be 1.60.0+)

code --version

Agent tree not loading

Solution: Verify agent personas path in settings

{
  "agentBuilder.agentPersonasPath": "/absolute/path/to/agent-personas"
}

Battle simulation fails

Solution: Ensure at least 2 agents selected Solution: Check MCP Data Ops is running

Code generation errors

Solution: Verify VS Code MCP is enabled in settings Solution: Check template files exist in templates/ directory

MCP integration not working

Solution: Verify MCP servers are running:

# Check MCP status
curl http://localhost:3100/health  # Data Ops MCP
curl http://localhost:3101/health  # VS Code MCP

Roadmap

Phase 1 (Current - v0.1.0) ✅

[x] Agent Explorer with 57 agents
[x] Create agent from template
[x] Battle simulator (local)
[x] Code export (TypeScript, Python, JSON, Pine Script)
[x] Agent details panel
[x] MCP integration skeleton
[x] Basic testing

Phase 2 (v0.2.0) - Q1 2025

[ ] Visual strategy builder with drag-and-drop
[ ] Advanced battle scenarios
[ ] Real-time backtesting
[ ] Performance charts and visualizations
[ ] Strategy marketplace
[ ] Collaborative features
[ ] Full MCP integration

Phase 3 (v0.3.0) - Q2 2025

[ ] Machine learning strategy optimization
[ ] Live trading integration
[ ] Portfolio management
[ ] Advanced analytics dashboard
[ ] Community agent sharing
[ ] Mobile companion app

Contributing

This extension is part of the U_Agentix platform. For contributions:

Fork the repository
Create feature branch (git checkout -b feature/amazing-feature)
Commit changes (git commit -m 'Add amazing feature')
Push to branch (git push origin feature/amazing-feature)
Open Pull Request

Code Style:

TypeScript for all code
ESLint + Prettier for formatting
Jest for testing
Document all public APIs

Support

Documentation: See U_Agentix Platform Docs
MCP Integration: See MCP Setup Guide
Issues: GitHub Issues
Community: Discord

License

Acknowledgments

Version: 0.1.0 Last Updated: October 18, 2025 Status: Phase 1 Complete Next Release: v0.2.0 (Visual Strategy Builder)

Made with ❤️ by the U_Agentix Spec Team

U_Agentix Agent Builder

CAMPINTL

U_Agentix Agent Builder VS Code Extension

Features

Agent Explorer

Agent Creation

Battle Simulator

Strategy Builder (Phase 2)

MCP Integration

Installation

Requirements

Quick Install

Manual Setup

Quick Start

1. Create Your First Agent

2. Customize Agent Configuration

3. Run a Battle Simulation

4. Export Agent to Code

Agent Explorer

Tree View Structure

Agent Details Panel

Context Menu Actions

Battle Simulator

Running Battles

Battle Results

Export Options

Code Generation

TypeScript Agent Class

Python Trading Script

Pine Script (TradingView)

Agent Templates

Momentum Strategy

Value Strategy

Contrarian Strategy

Quantitative Strategy

MCP Integration

Data Ops MCP

VS Code MCP

Analytics MCP

Configuration

Extension Settings

Agent Configuration

Commands

Command Palette

Keyboard Shortcuts

Development

Build from Source

Project Structure

Running Tests

Adding New Templates

Troubleshooting

Extension not activating

Agent tree not loading

Battle simulation fails

Code generation errors

MCP integration not working

Roadmap

Phase 1 (Current - v0.1.0) ✅

Phase 2 (v0.2.0) - Q1 2025

Phase 3 (v0.3.0) - Q2 2025

Contributing

Support

License

Acknowledgments