agentleFS
rinkuniksrinkuniks/android-claude-skill/AGENTS.md

android-claude-skill

AGENTS.md3 starsChanged 3 months agoMITView on GitHub

Build Android applications following Google's official architecture guidance, as demonstrated in the NowInAndroid reference app. Also audit and refactor existing Android/Compose/Kotlin Multiplatform (KMP) code down to zero technical debt. Use this file when the task involves: creating Android projects/screens/ViewModels/repositories/feature modules, Android architecture questions, or auditing/reviewing/refactoring existing Android/Compose/KMP code (quality, ANR risk, memory leaks, security, accessibility, test coverage). This same knowledge is packaged as a Claude Code Skill in SKILL.md — this file carries the identical guidance for any other agent…

  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 31,843 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repositoryOne other file it gives its agents.

The file

We split this AGENTS.md into the rules it gives an agent

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use this file when the task involves: creating Android projects/screens/ViewModels/repositories/feature modules, Android architecture …

    line 5
  2. Distinctive:

    Reactive streams: Use Kotlin Flow for all data exposure

    line 46
  3. Distinctive:

    Testable by design: Use interfaces and test doubles, no mocking libraries

    line 48
  4. Very specific:

    Create feature:myfeature:api module with navigation key

    line 87
  5. Very specific:

    Create feature:myfeature:impl module with:

    line 88
  6. Very specific:

    Use convention plugins in build-logic/ for consistent configuration:

    line 201

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

Android Development & Audit — Agent Instructions

Build Android applications following Google's official architecture guidance, as demonstrated in the NowInAndroid reference app. Also audit and refactor existing Android/Compose/Kotlin Multiplatform (KMP) code down to zero technical debt.

Use this file when the task involves: creating Android projects/screens/ViewModels/repositories/feature modules, Android architecture questions, or auditing/reviewing/refactoring existing Android/Compose/KMP code (quality, ANR risk, memory leaks, security, accessibility, test coverage).

This same knowledge is packaged as a Claude Code Skill in SKILL.md — this file carries the identical guidance for any other agent (Codex CLI, Cursor, GitHub Copilot, Antigravity, etc.) that reads a root AGENTS.md. See README.md for per-tool install steps.

Quick Reference

TaskReference File
Project structure & modulesmodularization.md
Architecture layers (UI, Domain, Data)architecture.md
Jetpack Compose patternscompose-patterns.md
Gradle & build configurationgradle-setup.md
Testing approach (writing new tests)testing.md
Auditing/refactoring existing codeaudit-checklist.md

Workflow Decision Tree

Creating a new project? → Read modularization.md for project structure → Use templates in assets/templates/

Adding a new feature? → Create feature module with api and impl submodules → Follow patterns in architecture.md

Building UI screens? → Read compose-patterns.md → Create Screen + ViewModel + UiState

Setting up data layer? → Read data layer section in architecture.md → Create Repository + DataSource + DAO

Auditing, reviewing, or refactoring existing Android/Compose/KMP code? → Read audit-checklist.md → Run its Context Graph → Scope Sizing → TODO Matrix → Permission Setup → Refactor → Verify workflow (10 dimensions: style, memory, performance/ANR, security, UI integrity, architecture, dead code, bugs, test coverage, accessibility) — reads the whole project first to build a module/dependency graph, asks once before doing bulk file edits instead of per-file, and attaches the graph in the final report alongside the TODO matrix

Core Principles

  1. Offline-first: Local database is source of truth, sync with remote
  2. Unidirectional data flow: Events flow down, data flows up
  3. Reactive streams: Use Kotlin Flow for all data exposure
  4. Modular by feature: Each feature is self-contained with clear boundaries
  5. Testable by design: Use interfaces and test doubles, no mocking libraries

Architecture Layers

┌─────────────────────────────────────────┐
│              UI Layer                    │
│  (Compose Screens + ViewModels)          │
├─────────────────────────────────────────┤
│           Domain Layer                   │
│  (Use Cases - optional, for reuse)       │
├─────────────────────────────────────────┤
│            Data Layer                    │
│  (Repositories + DataSources)            │
└─────────────────────────────────────────┘

Module Types

app/                    # App module - navigation, scaffolding
feature/
  ├── featurename/
  │   ├── api/          # Navigation keys (public)
  │   └── impl/         # Screen, ViewModel, DI (internal)
core/
  ├── data/             # Repositories
  ├── database/         # Room DAOs, entities
  ├── network/          # Retrofit, API models
  ├── model/            # Domain models (pure Kotlin)
  ├── common/           # Shared utilities
  ├── ui/               # Reusable Compose components
  ├── designsystem/     # Theme, icons, base components
  ├── datastore/        # Preferences storage
  └── testing/          # Test utilities

Creating a New Feature

  1. Create feature:myfeature:api module with navigation key
  2. Create feature:myfeature:impl module with:
    • MyFeatureScreen.kt - Composable UI
    • MyFeatureViewModel.kt - State holder
    • MyFeatureUiState.kt - Sealed interface for states
    • MyFeatureNavigation.kt - Navigation setup
    • MyFeatureModule.kt - Hilt DI module

Standard File Patterns

ViewModel Pattern

@HiltViewModel
class MyFeatureViewModel @Inject constructor(
    private val myRepository: MyRepository,
) : ViewModel() {

    val uiState: StateFlow<MyFeatureUiState> = myRepository
        .getData()
        .map { data -> MyFeatureUiState.Success(data) }
        .stateIn(
            scope = viewModelScope,
            started = SharingStarted.WhileSubscribed(5_000),
            initialValue = MyFeatureUiState.Loading,
        )

    fun onAction(action: MyFeatureAction) {
        when (action) {
            is MyFeatureAction.ItemClicked -> handleItemClick(action.id)
        }
    }
}

UiState Pattern

sealed interface MyFeatureUiState {
    data object Loading : MyFeatureUiState
    data class Success(val items: List<Item>) : MyFeatureUiState
    data class Error(val message: String) : MyFeatureUiState
}

Screen Pattern

@Composable
internal fun MyFeatureRoute(
    onNavigateToDetail: (String) -> Unit,
    viewModel: MyFeatureViewModel = hiltViewModel(),
) {
    val uiState by viewModel.uiState.collectAsStateWithLifecycle()
    MyFeatureScreen(
        uiState = uiState,
        onAction = viewModel::onAction,
        onNavigateToDetail = onNavigateToDetail,
    )
}

@Composable
internal fun MyFeatureScreen(
    uiState: MyFeatureUiState,
    onAction: (MyFeatureAction) -> Unit,
    onNavigateToDetail: (String) -> Unit,
) {
    when (uiState) {
        is MyFeatureUiState.Loading -> LoadingIndicator()
        is MyFeatureUiState.Success -> ContentList(uiState.items, onAction)
        is MyFeatureUiState.Error -> ErrorMessage(uiState.message)
    }
}

Repository Pattern

interface MyRepository {
    fun getData(): Flow<List<MyModel>>
    suspend fun updateItem(id: String, data: MyModel)
}

internal class OfflineFirstMyRepository @Inject constructor(
    private val localDataSource: MyDao,
    private val networkDataSource: MyNetworkApi,
) : MyRepository {

    override fun getData(): Flow<List<MyModel>> =
        localDataSource.getAll().map { entities ->
            entities.map { it.toModel() }
        }

    override suspend fun updateItem(id: String, data: MyModel) {
        localDataSource.upsert(data.toEntity())
    }
}

Key Dependencies

// Gradle version catalog (libs.versions.toml)
[versions]
kotlin = "1.9.x"
compose-bom = "2024.x.x"
hilt = "2.48"
room = "2.6.x"
coroutines = "1.7.x"

[libraries]
androidx-compose-bom = { group = "androidx.compose", name = "compose-bom", version.ref = "compose-bom" }
hilt-android = { group = "com.google.dagger", name = "hilt-android", version.ref = "hilt" }
room-runtime = { group = "androidx.room", name = "room-runtime", version.ref = "room" }

Build Configuration

Use convention plugins in build-logic/ for consistent configuration:

  • AndroidApplicationConventionPlugin - App modules
  • AndroidLibraryConventionPlugin - Library modules
  • AndroidFeatureConventionPlugin - Feature modules
  • AndroidComposeConventionPlugin - Compose setup
  • AndroidHiltConventionPlugin - Hilt setup

See gradle-setup.md for complete build configuration.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.