[Feature Request] AI-native automated UI testing for mobile apps via codebase analysis

Bug Description
Feature Request (not a bug): AI-Native Automated UI Testing for Claude Code

I'm a Flutter/Android developer and a Claude Code user. I'd like to suggest a feature that I believe could set Claude Code apart from all existing developer tools.

--- The Problem ---
Current mobile UI testing tools (Maestro, Appium, Flutter integration_test) all require developers to manually write configuration files or test scripts describing every screen, input rule, and navigation step. This is time-consuming and breaks every time the UI changes.

--- The Idea ---
Claude Code, as an AI that already reads and understands source code, could potentially:

1. Analyze the codebase (widget trees, navigation routes, state management) to understand the app structure automatically
2. Autonomously plan and execute UI tests on an emulator or real device — no YAML or test scripts needed
3. Offer variable execution speeds: slow for human visual review, and maximum speed for crash/regression detection
4. Report issues with direct references to the relevant source code

--- Why Claude Code specifically ---
This is only possible because Claude understands code context. Existing tools require human-authored instructions because they have no code comprehension. Claude Code could eliminate that entirely.

I think this would be a compelling and unique capability for mobile developers.

Environment Info

• Platform: darwin
• Terminal: androidstudio
• Version: 2.1.77
• Feedback ID: e94dcc39-42d1-4ee7-b98c-02de66aae113

Errors

[]