mcpField
C
coil-compose

coil-compose

github.com/new-silvermoonDevelopment992444 installsUpdated 13w ago
Repository

About coil-compose

A skill that guides image loading in Jetpack Compose with Coil. It helps Android developers choose among Coil’s composables and apply guidance for loading states, sizing, accessibility, and performance.

Key features

  1. 01AsyncImage usage guidance
  2. 02Painter and slot API selection
  3. 03Loading and error state handling
  4. 04Image sizing and transformations
  5. 05Caching and accessibility practices

Use cases

  1. 01Displaying remote images in Compose
  2. 02Showing custom loading and error UI
  3. 03Loading images inside Canvas or Icon
  4. 04Tuning image requests for lists

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/new-silvermoon/awesome-android-agent-skills --skill coil-compose

After installing, type/coil-composein Claude Code to confirm it loaded.

Details

From the project README

Coil for Jetpack Compose

Instructions

When implementing image loading in Jetpack Compose, use Coil (Coroutines Image Loader). It is the recommended library for Compose due to its efficiency and seamless integration.

1. Primary Composable: AsyncImage

Use AsyncImage for most use cases. It handles size resolution automatically and supports standard Image parameters.

AsyncImage(
    model = ImageRequest.Builder(LocalContext.current)
        .data("https://example.com/image.jpg")
        .crossfade(true)
        .build(),
    placeholder = painterResource(R.drawable.placeholder),
    error = painterResource(R.drawable.error),
    contentDescription = stringResource(R.string.description),
    contentScale = ContentScale.Crop,
    modifier = Modifier.clip(CircleShape)
)
2. Low-Level Control: rememberAsyncImagePainter

Use rememberAsyncImagePainter only when you need a Painter instead of a composable (e.g., for Canvas or Icon) or when you need to observe the loading state manually.

Warning

rememberAsyncImagePainter does not detect the size your image is loaded at on screen and always loads the image with its original dimensions by default. Use AsyncImage unless a Painter is strictly required.

val painter = rememberAsyncImagePainter(
    model = ImageRequest.Builder(LocalContext.current)
        .data("https://example.com/image.jpg")
        .size(Size.ORIGINAL) // Explicitly define size if needed
        .build()
)
3. Slot API: SubcomposeAsyncImage

Use SubcomposeAsyncImage when you need a custom slot API for different states (Loading, Success, Error).

Caution

Subcomposition is slower than regular composition. Avoid using SubcomposeAsyncImage in performance-critical areas like LazyColumn or LazyRow.

SubcomposeAsyncImage(
    model = "https://example.com/image.jpg",
    contentDescription = null,
    loading = {
        CircularProgressIndicator()
    },
    error = {
        Icon(Icons.Default.Error, contentDescription = null)
    }
)
4. Performance & Best Practices
  • Singleton ImageLoader: Use a single ImageLoader instance for the entire app to share the disk/memory cache.
  • Main-Safe: Coil executes image requests on a background thread automatically.
  • Crossfade: Always enable crossfade(true) in ImageRequest for a smoother transition from placeholder to success.
  • Sizing: Ensure contentScale is set appropriately to avoid loading larger images than necessary.
5. Checklist for implementation
  • Prefer AsyncImage over other variants.
  • Always provide a meaningful contentDescription or set it to null for decorative images.
  • Use crossfade(true) for better UX.
  • Avoid SubcomposeAsyncImage in lists.
  • Configure ImageRequest for specific needs like transformations (e.g., CircleCropTransformation).

Comments

Markdown supported