DAD Agent

AI-powered autonomous testing agent that intelligently explores and tests web applications using LangGraph decision-making and browser automation.

Prerequisites

• Node.js >= 18
• Docker (for Qdrant vector database)
• VS Code (optional, for extension)

Setup

1. Install Dependencies

npm install
cd runtime-discovery && npm install
cd ../vscode-extension && npm install

2. Start Qdrant Vector Database

Windows (PowerShell):

# Using Docker (recommended)
docker run -p 6333:6333 -v ${PWD}/qdrant_data:/qdrant/storage qdrant/qdrant

# Or download binary
Invoke-WebRequest -Uri "https://github.com/qdrant/qdrant/releases/latest/download/qdrant-x86_64-pc-windows-msvc.zip" -OutFile "qdrant.zip"
Expand-Archive -Path "qdrant.zip" -DestinationPath "."
.\qdrant.exe

Windows (CMD):

# Using Docker (recommended)
docker run -p 6333:6333 -v %cd%/qdrant_data:/qdrant/storage qdrant/qdrant

# Or download binary
curl -L https://github.com/qdrant/qdrant/releases/latest/download/qdrant-x86_64-pc-windows-msvc.zip -o qdrant.zip
tar -xf qdrant.zip
qdrant.exe

Linux/macOS:

# Using Docker (recommended)
docker run -p 6333:6333 -v $(pwd)/qdrant_data:/qdrant/storage qdrant/qdrant

# Or download binary
wget https://github.com/qdrant/qdrant/releases/latest/download/qdrant-x86_64-unknown-linux-gnu.tar.gz
tar -xzf qdrant-x86_64-unknown-linux-gnu.tar.gz
./qdrant

Initialize Collection:

PowerShell:

node createCollection.ts

CMD:

node createCollection.ts

Linux/macOS:

node createCollection.ts

3. Configure Environment (Optional)

# Copy and edit .env for Azure/OpenAI embeddings
cp .env.example .env
# Edit OPENAI_API_KEY, AZURE_* settings
# Default uses local embeddings (@xenova/transformers)

4. Install Playwright Browsers

cd runtime-discovery
npx playwright install

Usage

Run Autonomous Testing

cd runtime-discovery
npm run start https://your-app-url.com

Test Knowledge Base

node testKB.ts

Analyze Results

node error-intelligence/pipeline.js runs/<run-id>.json reports/analysis.json

Architecture

🧠 LangGraph Agent Pipeline

• Anomaly Detection - Identifies UI/network issues
• Diagnosis - Analyzes root causes
• Memory - Maintains testing context
• Decision Engine - AI-powered action selection
• Executor - Safe action execution with confidence thresholds
• Validator - Result verification
• Learner - Knowledge accumulation

🔍 Runtime Discovery

• Dynamic UI action discovery (buttons, links, forms, inputs, selects)
• Network monitoring and console error tracking
• Screenshot capture with stability detection
• Playwright-based browser automation

📊 Error Intelligence

• Multi-stage analysis pipeline
• Azure Vision safety checks
• Performance monitoring
• Evidence correlation (screenshots + network traces)

🧮 Knowledge Management

• Qdrant Vector Database (port 6333)
• Local Embeddings (@xenova/transformers, 384-dim)
• Azure Embeddings (optional, 1536-dim)
• Semantic search for learned behaviors
• Success/failure pattern recognition
• Confidence scoring for solutions

VS Code Integration

cd vscode-extension
npm run compile
# Press F5 to launch extension development host

Output Structure

• runs/ - Test execution traces (JSON)
• screenshots/ - UI screenshots during testing
• reports/ - Analysis reports
• qdrant_data/ - Vector database storage
• Knowledge Base - Stored at http://localhost:6333

Troubleshooting

Qdrant Connection Issues:

# Check if Qdrant is running
curl http://localhost:6333/collections

# Restart Qdrant
docker restart <qdrant-container-id>

Missing Dependencies:

# Install all dependencies
npm install
cd runtime-discovery && npm install
cd ../vscode-extension && npm install