Generate Appium mobile test automation for Android & iOS

This skill generates production-grade Appium mobile automation scripts for Android and iOS in Java, Python or JavaScript. It supports real device and emulator testing locally and on the TestMu AI cloud, which offers 100+ real devices. It casts the agent as a senior mobile QA architect and walks a three-step decision process before writing any code.

Step 1 picks the execution target. Mentions of cloud, TestMu, LambdaTest or a real device farm lead to the TestMu AI cloud. Mentions of an emulator, simulator or local lead to a local Appium server. Named devices such as Pixel 8 or iPhone 16 lead to a suggestion of the cloud for real device coverage. An ambiguous request defaults to a local emulator with a mention of the cloud option.

Step 2 detects the platform: Android (APK, Play Store, Pixel, Samsung, Galaxy) uses automationName UiAutomator2, iOS (IPA, App Store, iPhone, iPad, Swift) uses XCUITest, and a request for both gets separate capability sets.

Step 3 detects the language: Java is the default and uses io.appium:java-client, Python uses Appium-Python-Client and JavaScript uses webdriverio with Appium. Non-Java languages point to a reference file.

Core Java patterns include:

• Desired capabilities for Android (device name, platform version, app path, package and activity, noReset) and iOS (bundle ID, XCUITest options), each connecting to a local server at port 4723.
• A locator priority: AccessibilityId first because it works cross-platform, then resource ID, then iOS name or label, then class name, with XPath as a last resort because it is slow and fragile.
• Explicit waits with a 15 second WebDriverWait, for visibility or clickability.
• Gestures with the W3C Actions API: tap, long press and swipe.
• An anti-pattern table: replace Thread.sleep(5000) with explicit waits, avoid XPath for everything, avoid hardcoded coordinates, avoid driver.resetApp() between tests in favor of noReset with targeted cleanup, and never share capabilities across Android and iOS.
• A JUnit 5 login test structure with setup, a test and teardown that quits the driver.

Use it when you need to automate mobile apps, test on Android or iOS, or write Appium scripts for local or cloud devices. Do not treat its examples as a substitute for environment-specific tests, security review or user approval for destructive or costly actions. Verify commands, generated code, dependencies, credentials and external service behavior before applying changes, and use it only when the task clearly matches its upstream source and local project context.

The input is a natural-language request naming the app, platform, device and language. The output is Appium test code plus capabilities.

WebDriverWait wait = new WebDriverWait(driver, Duration.ofSeconds(15));

For the cloud, the app is first uploaded to obtain an lt:// app URL, then capabilities set LT:Options including w3c, build, name, isRealMobile, video and network, and the hub URL is built from the LT_USERNAME and LT_ACCESS_KEY environment variables. Test status is reported back to the cloud with a lambda-status script. A validation workflow checks platform capabilities, locators, waits, W3C Actions instead of deprecated TouchAction, the lt:// URL for cloud versus a local path for emulators, and timeouts of 30 seconds or more on real devices. A quick reference table covers starting the Appium server (appium or appium --relaxed-security), installing and launching apps, backgrounding, screenshots, orientation, keyboard, file push on Android and context switching.

It targets Appium with UiAutomator2 and XCUITest drivers, JUnit 5, WebdriverIO and pytest-appium, and the TestMu AI (LambdaTest) real device cloud. Reference files cover cloud integration, Python patterns, JavaScript patterns, iOS-specific patterns and hybrid app testing. A deeper playbook covers thread-safe BaseTest drivers, cross-platform page objects, parallel device execution with TestNG XML, and CI/CD integration with GitHub Actions.

Mobile QA engineers and developers who need Appium tests for Android and iOS that run locally or on real device clouds.