dpconde/claude-android-skill

Overview

This skill helps create production-quality Android applications using Google's architecture guidance and NowInAndroid best practices. It guides project modularization, Jetpack Compose UI patterns, MVVM ViewModels, Hilt DI, Room databases, and reactive data flows with Kotlin Flow. Use it to scaffold features, screens, repositories, and multi-module layouts that are testable and maintainable.

How this skill works

The skill inspects requests to create Android projects, feature modules, screens, ViewModels, repositories, and architecture decisions. It provides concrete file patterns, module layouts, and sample code snippets for UI, domain, and data layers. It suggests Gradle convention plugins, dependency versions, and testing strategies to produce consistent, offline-first, unidirectional-data-flow apps.

When to use it

• Starting a new Android app and designing module boundaries
• Adding a new feature module with screen, navigation, and DI
• Implementing ViewModel + UiState patterns for Compose screens
• Designing an offline-first repository with Room + network sync
• Setting up Gradle, version catalogs, or common build conventions
• Creating testable components and test utilities across modules

Best practices

• Follow feature-based modularization: api and impl submodules per feature
• Make local database the source of truth and expose data as Kotlin Flow
• Use unidirectional data flow: events down, state up via UiState and StateFlow
• Prefer Hilt for DI with explicit modules and scoped bindings
• Keep UI, Domain, and Data layers separate and small for testability
• Use shared design system and core libraries to avoid duplication

Example use cases

• Scaffold a new feature module: create api/impl, screen, ViewModel, DI module, and navigation key
• Implement a Compose screen that collects ViewModel StateFlow and renders Loading/Success/Error states
• Create an offline-first repository: Room DAO as local source and Retrofit network sync
• Set up Gradle convention plugins and libs.versions.toml for consistent builds across modules
• Add testing utilities and test doubles to core/testing to enable unit and integration tests