MCP工具：AI驱动的移动应用测试服务器

Platform prerequisites

#### Web - No additional setup after installing dependencies and Chromium.

#### Android - Android SDK installed - ADB available on PATH - ANDROID_HOME configured - At least one connected Android device or emulator

Useful checks:

npm run doctor
npm run devices

#### iOS - macOS host for local iOS automation - Xcode + Command Line Tools - Booted iOS Simulator (or compatible cloud/mobile device provider)

Quick Setup for Mobile Test Automation (One Command)

For a fast web-ready setup:

npm install && npx playwright install chromium

Then run the default web test:

npm run test:web

Optional environment setup:

cp .env.example .env

On Windows PowerShell:

Copy-Item .env.example .env

10. Example MCP client configuration

Use this in an MCP-capable client configuration file:

{
   "mcpServers": {
      "mobile-testing-ai-agent": {
         "command": "node",
         "args": ["mcp/server.js"],
         "cwd": "."
      }
   }
}

Contribution guidelines

• Prefer small, reviewable pull requests.
• Avoid unrelated refactors in the same PR.
• Preserve script and folder naming conventions.
• Document new commands and agent behaviors in this README.

Core NPM packages

• mobilewright ^0.0.33
• @playwright/test ^1.59.1
• playwright ^1.59.1
• @modelcontextprotocol/sdk ^1.29.0
• zod ^4.4.3
• openai ^6.37.0
• dotenv ^17.4.2
• fs-extra ^11.3.5
• chalk ^5.6.2
• axios ^1.16.0
• cheerio ^1.2.0

6. Run orchestration pipelines for mobile automation

#### Android pipeline

npm run orchestrate:android

#### iOS pipeline

npm run orchestrate:ios

7. Run complete web AI pipeline

npm run pipeline:web

This executes Planner -> Generator -> Healer for web.

11. CI/CD pipeline (GitHub Actions)

This repository now includes a CI/CD workflow at .github/workflows/ci-cd.yml.

CI behavior: - Runs on pushes to main and pull requests targeting main. - Installs dependencies with npm ci. - Installs Playwright Chromium. - Executes the web suite with npm run test:web. - Uploads Playwright report artifacts when available.

CD behavior: - Runs after CI on pushes to main. - Triggers deployment only if DEPLOY_WEBHOOK_URL is configured as a repository secret. - Safely skips deployment when the secret is not set.

To enable deployment, add this repository secret: - DEPLOY_WEBHOOK_URL

Project Overview: Mobile Testing, Mobile Automation, and Mobile QA

This repository provides an AI-powered framework for mobile testing and mobile test automation (plus web validation) using:

• Planner Agent: explores a target app or site and generates a structured Markdown test plan for mobile QA.
• Generator Agent: converts the Markdown test plan into executable Playwright or Mobilewright tests for mobile automation workflows.
• Healer Agent: runs tests, detects failures, and attempts automated repair cycles to improve mobile test automation stability.
• Android and iOS Orchestrators: run Planner -> Generator -> Healer as end-to-end pipelines for each platform and mobile testing lifecycle.
• MCP Server: exposes the framework as callable MCP tools over stdio for AI clients and automation platforms.

The framework is designed to accelerate mobile test authoring, improve test stability, and provide an extensible baseline for AI-assisted mobile QA workflows.

Dependencies for Mobile Automation and Mobile QA

Usage Guide for Mobile Testing and Mobile QA

7b. Run complete mobile QA pipelines

For end-to-end mobile test automation on each platform:

npm run orchestrate:android
npm run orchestrate:ios

Mobile Testing FAQ

How does this repository help with mobile QA?

This project supports mobile QA with AI-driven planning, test generation, and self-healing execution, plus Android and iOS orchestration flows.

Questions?

If you have any questions: - 💬 Open a GitHub Discussion - 🐛 Report bugs via GitHub Issues - 📧 Email: padmaraj dot nidagundi at gmail.com

Response time: Typically 24-48 hours.

---