simiancraft-skills

Automated Android UI/integration testing specialist; the agent that drives a real Android app on a headless emulator under WSL/Linux and gates on what it observes. Use when the task is "run the app on an emulator", "smoke-test a screen", "drive the Android UI", "reproduce a tap-and-crash / ANR", "automate an Android flow", "set up Android UI testing", "write a Maestro flow", "verify a change on Android without a physical device", or "test the camera/segmentation/mask on Android". Owns the boot → install → drive → assert loop: KVM AVD bring-up, APK acquisition (local gradle or EAS artifact), Maestro driving (tap-by-testID/label, wait, assert), and logcat + screenshot gating. Knows its own ceiling: it validates no-crash, navigation, wiring, flows, and effect toggles, and it does NOT validate real FPS/perf, true camera/mic fidelity, or network quality; those stay device passes and it says so plainly. Loads the android-emulator-harness skill (and android-emulator-mask-testing for camera work) before acting.

Android native specialist for Kotlin, Java, Gradle/AGP, the Jetpack libraries, JNI/NDK, OpenGL ES and camera pipelines (Camera2, CameraX, MediaCodec, MediaPipe, ML Kit), and React Native / Expo Modules native bridging. Use when the task touches `android/**` (`*.kt`, `*.java`, `*.gradle` / `*.gradle.kts`, `AndroidManifest.xml`, `libs.versions.toml`, ProGuard / R8 rules), JNI / `CMakeLists.txt`, RN legacy bridge or TurboModules / Fabric / Codegen, the Expo Modules Kotlin DSL, Kotlin coroutines / Flow and threading primitives, manifest-merger conflicts, `minSdk` / `compileSdk` / `targetSdk` drift, R8 reflection breakage, AGP / Kotlin Gradle Plugin / JDK version mismatches, autolinking under Expo + React Native, or "compiles locally but fails on EAS" symptoms. Two modes: implement (write/edit Kotlin or Gradle, verify with `compileDebugKotlin`) or review (audit for idiom, lifecycle, threading, and resource-management correctness). Loads upstream docs (kotlinlang.org, developer.android.com, the React Native and Expo Modules guides) when prior knowledge is insufficient.

Bring up an Android app in a headless emulator on Linux/WSL and drive it for automated integration testing, the Android analog of Playwright for web. Boots an AVD under KVM, installs an APK (standalone or dev-client+Metro), drives the UI with Maestro (resilient tap-by-label/wait/assert), and gates on logcat + screenshots that come back for inspection. Use for ANY Android/Expo/React-Native project (not just one app) when the task is "run the app on an emulator", "drive the Android UI", "smoke-test a screen", "reproduce a tap-and-crash", "automate an Android flow", or "set up Android UI testing". Project-agnostic base; layer a domain skill on top for specialized inputs (see android-emulator-mask-testing for camera/segmentation). Validated on Linux/WSL with KVM, Maestro 2.x, and JDK 17.

Specialization of android-emulator-harness for CAMERA / segmentation testing: get a real PERSON in front of the emulator camera so MediaPipe / ML Kit selfie segmentation produces an actual mask, then verify background-replacement / blur / shader effects and tune mask threshold by vision. Use when the task is "test the mask", "verify segmentation", "tune mask threshold", "check a camera/background effect on Android", or any time an app must SEE a human through the emulator camera without a physical device. Read android-emulator-harness FIRST for bring-up, drive, and assert; this skill only overrides the camera-specific pieces. Validated on Linux/WSL with a 32-bit x86 emulator image and MediaPipe/ML Kit segmentation.

Plans self-destruct when shipped (the Inspector Gadget Rule). Use this skill to draft tactical, hand-off-ready planning docs with a 150-word Goal cap, atomic commit steps with verification gates, before/after file trees, and a two-key handshake before deletion. Trigger on "plan this", "write a plan for X", "draft a plan", "make a plan", "plan a refactor", "plan a migration", "plan a deprecation", "design the implementation", "write a PLAN.md", or when scoping multi-commit feature work that needs hand-off. Skip for: subscription plans, query execution plans, quarterly roadmaps, billing plans, sprint plans, test plans (QA matrices).

Specialization of playwright-harness for CAMERA / segmentation testing in the browser: put a real PERSON in front of getUserMedia so MediaPipe / selfie segmentation produces an actual mask, then verify background-replacement / blur / shader effects by vision. Use when the task is "test the mask on web", "verify segmentation in the browser", "feed the camera a person fixture", "check a camera/background effect in Chrome", or any time a web app must SEE a human through the webcam without physical hardware. Read playwright-harness FIRST for script conventions and execution; this skill only adds the camera-feed pieces. The web analog of android-emulator-mask-testing. Validated on Linux/WSL, headless Chromium, against MediaPipe Selfie Segmentation.

Specialization of playwright-harness for capturing an ANIMATED GIF of web content: drive a page/canvas/WebGL animation, grab a sequence of frames, and encode a looping GIF. Use when the task is "make a gif of this", "grab a gif of the demo/animation/shader", "record the canvas as a gif", or "capture a looping clip for a PR/README/issue"; any time a still screenshot is not enough. Read playwright-harness FIRST for script conventions and execution. For the quality/size dials (preventing deep-fried color, janky frame rate, or oversized files) read references/gif-optimization.md rather than guessing. Validated on Linux/WSL, headless Chromium + ANGLE, against WebGL canvas content.