Does it generate test reports?

Yes, it automatically generates detailed Markdown reports in your project's docs/test/report/ directory, including summaries, detailed results, and error logs.

Do I need to install external dependencies?

Yes, you need ADB installed on your system and the @midscene/android package installed in your project for visual-driven capabilities.

What is the difference between visual and fast execution modes?

Visual mode uses AI vision to understand the screen layout dynamically, while fast mode uses direct ADB shell commands for speed, precise timing, and coordinate-based interactions.

What devices does this skill support?

It supports physical Android devices and emulators connected via ADB on macOS and Linux systems.

How does the document-driven testing work?

You can provide a Markdown file (like a bug report or test plan) to the skill; it parses the steps, filters out passed cases, executes the remaining ones, and updates the status.

Android UI Automation Test

Name: Android UI Automation Test
Author: adzcsx2

byadzcsx2

0•

Mobile Development

Automates Android UI testing by intelligently combining Midscene vision-driven actions with high-speed ADB command execution.

This skill empowers developers to perform sophisticated Android UI automation directly from Claude. It features a smart execution engine that switches between Midscene's visual recognition (GLM-4.6V) for dynamic UI elements and direct ADB shell commands for millisecond-precision operations. It excels at document-driven testing, allowing users to execute test cases from Markdown reports or bug lists, automatically skipping passing tests and generating comprehensive timestamped execution reports with screenshots and logs.

Características Principales

01Document-driven testing that parses Markdown reports and bug lists

02Automated generation of detailed test reports with screenshots and logs

03Vision-driven UI recognition using Midscene and GLM-4.6V integration

040 GitHub stars

05Hybrid mode combining visual positioning with rapid command sequences

06High-speed ADB execution for precise timing and race condition testing

Casos de Uso

01Executing regression tests directly from an existing Markdown checklist

02Testing complex UI race conditions using millisecond-level ADB sequences

03Verifying bug reports by automating reproduction steps from a .md file

Características Principales

01Document-driven testing that parses Markdown reports and bug lists

02Automated generation of detailed test reports with screenshots and logs

03Vision-driven UI recognition using Midscene and GLM-4.6V integration

040 GitHub stars

05Hybrid mode combining visual positioning with rapid command sequences

06High-speed ADB execution for precise timing and race condition testing

Casos de Uso

01Executing regression tests directly from an existing Markdown checklist

02Testing complex UI race conditions using millisecond-level ADB sequences

03Verifying bug reports by automating reproduction steps from a .md file