🎭 Playwright AI Agentic Testing Framework

A production-ready Quality Engineering test framework leveraging Playwright, Model Context Protocol (MCP), VS Code, and GitHub Copilot AI Agents for unified UI and REST API automation.

🏗️ Architecture & Core Components

• Agentic Loop (Plan → Generate → Heal): Autonomous test creation, specification generation, and self-healing broken locators via MCP.

• Hybrid Test Capabilities: Cross-browser UI automation + Schema-validated REST API contract testing via Zod.

• VS Code Native MCP Integration: Connects Copilot Chat directly to Playwright's local browser runtime context via standard I/O.

Related MCP server: RunAutomation MCP Server

🚀 Quickstart Guide

Prerequisites

• Node.js 18+ installed

• VS Code with GitHub Copilot & Playwright extensions enabled

Setup Commands

# 1. Install dependencies & browsers
npm install
npx playwright install

# 2. Initialize Playwright MCP Agents
npm run agents:init

🤖 Using Playwright AI Agents

In VS Code Copilot Chat, interact with the agentic loop:

• Planner Agent: @playwright-test-planner draft test specs for specs/ui-shopping-flow.md using tests/seed/seed.spec.ts

• Generator Agent: @playwright-test-generator write TypeScript tests for specs/ui-shopping-flow.md into tests/ui/checkout.spec.ts

• Healer Agent: @playwright-test-healer diagnose and repair failing tests in tests/ui/checkout.spec.ts

🧪 Running Tests

# Run all tests (UI + API)
npm run test

# Run UI tests only
npm run test:ui

# Run API tests only
npm run test:api

# View HTML Report
npm run test:report