Design and optimize AI agent action spaces, tool definitions, and observation formatting for higher completion rates.
日本語の概要は準備中です。原文の説明を表示しています。
Clean Architecture patterns for Android and Kotlin Multiplatform projects — module structure, dependency rules, UseCases, Repositories, and data layer patterns.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Clean Architecture patterns for Android and KMP projects. Covers module boundaries, dependency inversion, UseCase/Repository patterns, and data layer design with Room, SQLDelight, and Ktor.
project/
├── app/ # Android entry point, DI wiring, Application class
├── core/ # Shared utilities, base classes, error types
├── domain/ # UseCases, domain models, repository interfaces (pure Kotlin)
├── data/ # Repository implementations, DataSources, DB, network
├── presentation/ # Screens, ViewModels, UI models, navigation
├── design-system/ # Reusable Compose components, theme, typography
└── feature/ # Feature modules (optional, for larger projects)
├── auth/
├── settings/
└── profile/
app → presentation, domain, data, core
presentation → domain, design-system, core
data → domain, core
domain → core (or no dependencies)
core → (nothing)
Critical: domain must NEVER depend on data, presentation, or any framework. It contains pure Kotlin only.
Each UseCase represents one business operation. Use operator fun invoke for clean call sites:
class GetItemsByCategoryUseCase(
private val repository: ItemRepository
) {
suspend operator fun invoke(category: String): Result<List<Item>> {
return repository.getItemsByCategory(category)
}
}
// Flow-based UseCase for reactive streams
class ObserveUserProgressUseCase(
private val repository: UserRepository
) {
operator fun invoke(userId: String): Flow<UserProgress> {
return repository.observeProgress(userId)
}
}
Domain models are plain Kotlin data classes — no framework annotations:
data class Item(
val id: String,
val title: String,
val description: String,
val tags: List<String>,
val status: Status,
val category: String
)
enum class Status { DRAFT, ACTIVE, ARCHIVED }
Defined in domain, implemented in data:
interface ItemRepository {
suspend fun getItemsByCategory(category: String): Result<List<Item>>
suspend fun saveItem(item: Item): Result<Unit>
fun observeItems(): Flow<List<Item>>
}
Coordinates between local and remote data sources:
class ItemRepositoryImpl(
private val localDataSource: ItemLocalDataSource,
private val remoteDataSource: ItemRemoteDataSource
) : ItemRepository {
override suspend fun getItemsByCategory(category: String): Result<List<Item>> {
return runCatching {
val remote = remoteDataSource.fetchItems(category)
localDataSource.insertItems(remote.map { it.toEntity() })
localDataSource.getItemsByCategory(category).map { it.toDomain() }
}
}
override suspend fun saveItem(item: Item): Result<Unit> {
return runCatching {
localDataSource.insertItems(listOf(item.toEntity()))
}
}
override fun observeItems(): Flow<List<Item>> {
return localDataSource.observeAll().map { entities ->
entities.map { it.toDomain() }
}
}
}
Keep mappers as extension functions near the data models:
// In data layer
fun ItemEntity.toDomain() = Item(
id = id,
title = title,
description = description,
tags = tags.split("|"),
status = Status.valueOf(status),
category = category
)
fun ItemDto.toEntity() = ItemEntity(
id = id,
title = title,
description = description,
tags = tags.joinToString("|"),
status = status,
category = category
)
@Entity(tableName = "items")
data class ItemEntity(
@PrimaryKey val id: String,
val title: String,
val description: String,
val tags: String,
val status: String,
val category: String
)
@Dao
interface ItemDao {
@Query("SELECT * FROM items WHERE category = :category")
suspend fun getByCategory(category: String): List<ItemEntity>
@Upsert
suspend fun upsert(items: List<ItemEntity>)
@Query("SELECT * FROM items")
fun observeAll(): Flow<List<ItemEntity>>
}
-- Item.sq
CREATE TABLE ItemEntity (
id TEXT NOT NULL PRIMARY KEY,
title TEXT NOT NULL,
description TEXT NOT NULL,
tags TEXT NOT NULL,
status TEXT NOT NULL,
category TEXT NOT NULL
);
getByCategory:
SELECT * FROM ItemEntity WHERE category = ?;
upsert:
INSERT OR REPLACE INTO ItemEntity (id, title, description, tags, status, category)
VALUES (?, ?, ?, ?, ?, ?);
observeAll:
SELECT * FROM ItemEntity;
class ItemRemoteDataSource(private val client: HttpClient) {
suspend fun fetchItems(category: String): List<ItemDto> {
return client.get("api/items") {
parameter("category", category)
}.body()
}
}
// HttpClient setup with content negotiation
val httpClient = HttpClient {
install(ContentNegotiation) { json(Json { ignoreUnknownKeys = true }) }
install(Logging) { level = LogLevel.HEADERS }
defaultRequest { url("https://api.example.com/") }
}
// Domain module
val domainModule = module {
factory { GetItemsByCategoryUseCase(get()) }
factory { ObserveUserProgressUseCase(get()) }
}
// Data module
val dataModule = module {
single<ItemRepository> { ItemRepositoryImpl(get(), get()) }
single { ItemLocalDataSource(get()) }
single { ItemRemoteDataSource(get()) }
}
// Presentation module
val presentationModule = module {
viewModelOf(::ItemListViewModel)
viewModelOf(::DashboardViewModel)
}
@Module
@InstallIn(SingletonComponent::class)
abstract class RepositoryModule {
@Binds
abstract fun bindItemRepository(impl: ItemRepositoryImpl): ItemRepository
}
@HiltViewModel
class ItemListViewModel @Inject constructor(
private val getItems: GetItemsByCategoryUseCase
) : ViewModel()
Use Result<T> or a custom sealed type for error propagation:
sealed interface Try<out T> {
data class Success<T>(val value: T) : Try<T>
data class Failure(val error: AppError) : Try<Nothing>
}
sealed interface AppError {
data class Network(val message: String) : AppError
data class Database(val message: String) : AppError
data object Unauthorized : AppError
}
// In ViewModel — map to UI state
viewModelScope.launch {
when (val result = getItems(category)) {
is Try.Success -> _state.update { it.copy(items = result.value, isLoading = false) }
is Try.Failure -> _state.update { it.copy(error = result.error.toMessage(), isLoading = false) }
}
}
For KMP projects, use convention plugins to reduce build file duplication:
// build-logic/src/main/kotlin/kmp-library.gradle.kts
plugins {
id("org.jetbrains.kotlin.multiplatform")
}
kotlin {
androidTarget()
iosX64(); iosArm64(); iosSimulatorArm64()
sourceSets {
commonMain.dependencies { /* shared deps */ }
commonTest.dependencies { implementation(kotlin("test")) }
}
}
Apply in modules:
// domain/build.gradle.kts
plugins { id("kmp-library") }
domain — keep it pure KotlinGlobalScope or unstructured coroutines — use viewModelScope or structured concurrencySee skill: compose-multiplatform-patterns for UI patterns.
See skill: kotlin-coroutines-flows for async patterns.
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Design and optimize AI agent action spaces, tool definitions, and observation formatting for higher completion rates.
日本語の概要は準備中です。原文の説明を表示しています。
Operate as an agentic engineer using eval-first execution, decomposition, and cost-aware model routing.
日本語の概要は準備中です。原文の説明を表示しています。
Engineering operating model for teams where AI agents generate a large share of implementation output.
日本語の概要は準備中です。原文の説明を表示しています。
REST API design patterns including resource naming, status codes, pagination, filtering, error responses, versioning, and rate limiting for production APIs.
日本語の概要は準備中です。原文の説明を表示しています。
Write articles, guides, blog posts, tutorials, newsletter issues, and other long-form content in a distinctive voice derived from supplied examples or brand guidance. Use when the user wants polished written content longer than a paragraph, especially when voice consistency, structure, and credibility matter.
日本語の概要は準備中です。原文の説明を表示しています。
Patterns and architectures for autonomous Claude Code loops — from simple sequential pipelines to RFC-driven multi-agent DAG systems.
日本語の概要は準備中です。原文の説明を表示しています。