injecting-input-and-state

What problem does it solve?

This skill solves the problem of reproducing user journeys on a real Android device or emulator by injecting input events and enforcing a predictable system + app state from the host shell.

Core Features & Use Cases

• Input event injection: Use adb shell input to tap, swipe, drag-and-drop, type text (with correct quoting/space encoding), and send hardware key events (for back/home/power/navigation, etc.).
• Display and environment control: Override screen size/density with wm, and apply hermetic system settings by disabling animation scales via settings put global window_animation_scale / transition_animation_scale / animator_duration_scale.
• Between-test state reset + Activity launching: Reset app state with pm clear (not am force-stop), optionally cold-start with am start -S -W, and launch Activities by component or deep-link intents with typed extras; also use cmd service wrappers instead of deprecated svc aliases on API 30+.

Use case: In CI, you need a stable scripted flow for a settings/debug screen across a screen-size matrix while avoiding animation flake and ensuring each test starts from a clean app + permissions state.

Quick Start

Use this skill when you want an AI to generate a host-side adb command sequence that injects taps and typed text, applies animation-hardened hermetic settings, resets the app with pm clear, and launches the target Activity with extras.