name: android-ui-test description: Demonstration skill showing how to use uiautomator2-mcp-server for AI-driven Android UI testing. Also serves as the project's own automated UI test suite for continuous validation. license: MIT compatibility: Requires connected Android device with USB debugging enabled and ADB server running metadata: author: uiautomator2-mcp-server team version: 1.0.0 purpose: dual-purpose demonstration and production testing

Android UI Test Automation (Demo & Production)

This skill serves a dual purpose:

1. Demonstration: Shows how to build AI-driven Android UI testing skills
2. Production Testing: Acts as the project's own automated UI test suite

Goals

• Demonstrate best practices for Android UI automation skills
• Provide comprehensive testing for uiautomator2-mcp-server functionality
• Enable AI agents to validate Android device interactions
• Serve as a reference implementation for similar skills

Prerequisites

• Connected Android device with USB debugging enabled
• ADB server running (adb start-server)
• uiautomator2-mcp-server properly configured

Usage Scenarios

Scenario 1: Learning Example

When developers want to understand how to create Android UI testing skills:

• Study the structured approach in this skill
• Learn proper YAML metadata formatting
• Understand test case organization and execution flow
• See how to handle device-specific considerations

Scenario 2: Production Testing

When validating uiautomator2-mcp-server functionality:

• Execute complete test suite automatically
• Generate detailed pass/fail reports
• Validate all core device interaction capabilities
• Ensure consistent performance across device variations

Test Execution Flow

Phase 1: Device Setup and Validation

1. List connected devices via device_list
2. Select first available device
3. Initialize device via init tool
4. Verify device connectivity and basic info

Phase 2: Comprehensive Functionality Tests

1. Device Operations: Info, screenshot, UI hierarchy dump
2. Touch Actions: Click, long press, double click at various positions
3. Gesture Operations: Swipe, drag, system key presses
4. Application Management: List apps, launch apps, get app info
5. Element Interactions: Wait for elements, get bounds, text extraction, element clicks
6. Input Operations: Text input, keyboard management
7. Clipboard Operations: Read/write clipboard content

Phase 3: Intelligent Reporting

Generate comprehensive test report with:

• Device information and specifications
• Detailed pass/fail/skip status for each test case
• Error messages and troubleshooting guidance
• Performance metrics and recommendations

Key Features Demonstrated

✅ Structured Skill Organization: Clear directory structure and file naming ✅ Proper Metadata: Standard YAML frontmatter with complete information ✅ Progressive Disclosure: Light metadata, detailed content on demand ✅ Error Handling: Graceful handling of device-specific limitations ✅ Comprehensive Testing: Covers all major uiautomator2-mcp-server capabilities ✅ Self-Documentation: Clear instructions for both learning and execution

Expected Outcomes

For Learning:

• Clear understanding of skill structure and best practices
• Working example of Android UI automation implementation
• Reference for creating similar skills

For Testing:

• ✅ All core functionality tests pass
• ✅ Device operations work reliably
• ✅ UI interactions execute correctly
• ✅ Comprehensive test report generated

Best Practices Highlighted

• Modular Design: Separate concerns into scripts, references, and documentation
• Clear Documentation: Self-explanatory structure and comprehensive guides
• Robust Error Handling: Anticipate and handle common failure scenarios
• Scalable Architecture: Easy to extend with additional test cases
• Cross-Platform Considerations: Handle device manufacturer variations

Extension Points

This skill can be easily extended for:

• Additional device types and manufacturers
• Custom test scenarios and workflows
• Integration with CI/CD pipelines
• Advanced reporting and analytics
• Performance benchmarking

Release Workflow

This skill is also used during the release process to validate uiautomator2-mcp-server functionality:

Pre-Release Testing

Before each release:

1. Ensure all 69 unit tests pass
2. Run doctor command: u2mcp doctor -v
3. Execute Android UI tests on physical device
4. Verify all 77 MCP tools are registered

Release Process

The project uses automated CI/CD pipeline:

1. Version Tag: Create a PEP 440 compliant tag (e.g., v0.3.3)

2. CI Build: GitHub Actions automatically:

   • Runs lint, type check, and tests
   • Builds distribution packages
   • Publishes to PyPI
   • Creates GitHub Release with CHANGELOG notes
   • Updates server.json version automatically
   • Publishes to MCP Registry

3. server.json Update:

   # CI automatically does this during release:
   jq --arg v "0.3.3" '.version = $v | .packages[0].version = $v' server.json > server.json.tmp
   mv server.json.tmp server.json
   

4. MCP Registry Publish:

   • Uses GitHub OIDC for authentication
   • Validates package ownership via PyPI
   • Publishes server.json to official MCP registry

For implementation details, see the accompanying scripts and documentation.