Use this skill to fix scroll jank, lost item state, and broken animateItem() animations in LazyColumn, LazyRow, LazyVerticalGrid, and LazyHorizontalGrid. Covers stable item keys, contentType for mixed-type feeds, Modifier.animateItem() requirements, hoisting modifier chains and painters out of the items lambda, and validating item composable stability. Use when the developer mentions LazyColumn jank, dropped frames while scrolling, items losing scroll state on insert/remove/reorder, mixed feeds of cards/headers/ads feeling sluggish, animateItem() not animating, RecyclerView view-type analog, key parameter, or contentType parameter. The prefetch-window tuning lives in a sibling skill.
npx skills add https://github.com/skydoves/compose-performance-skills --skill optimizing-lazy-layouts
Lazy layouts compose only what's visible, but two things still cost: re-composition of items that should have been reused (missing key), and per-item allocation that compounds with scroll velocity (missing contentType, modifier chains created inside items { }). Both have a one-line fix. This skill teaches Claude how to apply that fix correctly and to validate that item composables are themselves skippable. Prefetch tuning is a separate concern — see ../configuring-lazy-prefetch/SKILL.md.
LazyColumn, LazyRow, LazyVerticalGrid, or LazyHorizontalGrid.Modifier.animateItem() was added but no animation runs on inserts or removals.unstable/non-skippable, or @TraceRecomposition shows item composables recomposing on every scroll tick.../configuring-lazy-prefetch/SKILL.md.List<Foo>, Flow<Foo>, a domain var) → first run ../../stability/diagnosing-compose-stability/SKILL.md and then ../../stability/stabilizing-compose-types/SKILL.md.state.value in Composition phase, recomposing the row every frame → use ../../recomposition/deferring-state-reads/SKILL.md.firstVisibleItemIndex == 0) is the hot path → use ../../recomposition/choosing-derivedstateof/SKILL.md.Modifier.animateItem() (the GA replacement for the experimental animateItemPlacement).org.jetbrains.kotlin.plugin.compose applied. Strong Skipping is on by default; non-skippable item composables become amplified at scroll speed.../../measurement/generating-baseline-profiles/SKILL.md when ready to measure.items(...) call. Walk every LazyListScope.items(list), items(count), itemsIndexed(list), and the LazyGridScope equivalents. For each, decide: does each element have a stable identity that outlives a single composition? If yes — and it almost always does — supply key = { it.id } using a server-side stable ID. MUST NOT use the list index, UUID.randomUUID() evaluated per emission, or hashCode() of a mutable object.// WRONG
LazyColumn { items(snacks) { snack -> SnackRow(snack) } }
// WRONG because: index-based identity → insert/remove discards composition state and breaks animateItem().
// RIGHT
LazyColumn {
items(
items = snacks,
key = { it.id },
contentType = { it::class },
) { snack ->
SnackRow(snack, Modifier.animateItem())
}
}
contentType for heterogeneous lists. Lazy layouts maintain a per-type composition cache analogous to RecyclerView's view-type. When item N + 1 has the same contentType as a recycled slot, the cached composition is reused; otherwise it is discarded and rebuilt. For homogeneous lists Compose infers a single content type and contentType is optional. For mixed feeds (cards, headers, ads, carousels, dividers) MUST supply a stable type discriminator.../../stability/diagnosing-compose-stability/SKILL.md. If the item composable accepts an unstable parameter, no amount of key/contentType work will help — the row recomposes on every scroll-driven snapshot tick anyway. Fix with ../../stability/stabilizing-compose-types/SKILL.md before tuning further.BorderStroke instances built inside the lambda are reallocated each pass. Hoist constants and remember-based caches above the LazyColumn or to the call site. Modifier chains are themselves cheap because Compose deduplicates them structurally — hoist a Modifier only when profiling proves it matters.Modifier.animateItem() for visual continuity. Pair with a stable key. The animation runs on inserts, removals, and reorders; without key the animation cannot bind to identity and silently no-ops. The default fade-in / fade-out / placement spring is usually correct; tune with fadeInSpec, fadeOutSpec, placementSpec only when the design system requires it.painterResource(...), MaterialTheme.colorScheme.surface, RoundedCornerShape(...) resolutions on every item composition add up. Hoist to the screen-level composable and pass down, or remember once at the LazyColumn parent.@TraceRecomposition and Layout Inspector. During a controlled scroll, expect each item composable to recompose at most once per real state change — not per scroll tick. Layout Inspector → Recomposition Counts column should plateau, not climb monotonically.key// WRONG
LazyColumn {
items(snacks) { snack -> SnackRow(snack) }
}
// WRONG because: items default to index-based identity. On insert/remove/reorder, every position past the change point has a different "identity", composition state and scroll-restoration are lost, and Modifier.animateItem() has nothing to animate from.
// RIGHT
LazyColumn {
items(snacks, key = { it.id }) { snack -> SnackRow(snack) }
}
// WRONG
items(snacks, key = { UUID.randomUUID() }) { snack -> SnackRow(snack) }
// WRONG because: a fresh key on every recomposition guarantees the cached composition is discarded every time — strictly worse than no key.
// WRONG
items(snacks, key = { it.hashCode() }) { snack -> SnackRow(snack) }
// WRONG because: hashCode() of a mutable type changes when fields mutate, breaking identity continuity for the same logical item.
// RIGHT
items(snacks, key = { it.id }) { snack -> SnackRow(snack) }
contentType// WRONG
items(feed, key = { it.id }) { item ->
when (item) {
is FeedItem.Card -> CardRow(item)
is FeedItem.Ad -> AdRow(item)
is FeedItem.Header -> HeaderRow(item)
}
}
// WRONG because: cached compositions of one type are discarded when scrolled into a different type's slot — every row crossing a type boundary is a fresh build instead of a recycled update.
// RIGHT
items(
items = feed,
key = { it.id },
contentType = { it::class },
) { item ->
when (item) {
is FeedItem.Card -> CardRow(item)
is FeedItem.Ad -> AdRow(item)
is FeedItem.Header -> HeaderRow(item)
}
}
Modifier.animateItem() without a stable key// WRONG
items(snacks) { snack ->
SnackRow(snack, Modifier.animateItem())
}
// WRONG because: animateItem() binds animation state to the item's key. With no key, identity is index-based, so an insert at position 0 looks like every-row-changed and nothing animates correctly.
// RIGHT
items(snacks, key = { it.id }) { snack ->
SnackRow(snack, Modifier.animateItem())
}
// WRONG
items(snacks, key = { it.id }) { snack ->
val placeholder = painterResource(R.drawable.snack_placeholder)
val border = BorderStroke(1.dp, MaterialTheme.colorScheme.outline)
Card(border = border) {
AsyncImage(snack.imageUrl, placeholder = placeholder)
}
}
// WRONG because: painterResource resolution and BorderStroke allocation happen on every item composition; at high scroll velocity these compound into measurable allocation pressure.
// RIGHT
@Composable
fun SnackList(snacks: ImmutableList<Snack>) {
val placeholder = painterResource(R.drawable.snack_placeholder)
val border = BorderStroke(1.dp, MaterialTheme.colorScheme.outline)
LazyColumn {
items(snacks, key = { it.id }, contentType = { it::class }) { snack ->
Card(border = border) {
AsyncImage(snack.imageUrl, placeholder = placeholder)
}
}
}
}
Note: Compose deduplicates structurally-equal Modifier chains internally, so reallocating Modifier.fillMaxWidth().padding(16.dp) per item is a *micro*-optimization. Hoist a Modifier only when profiling identifies it as the bottleneck — premature remember { Modifier.… } adds noise without measurable benefit.
// WRONG
@Composable
fun SnackRow(snack: Snack, tags: List<String>) { /* ... */ }
// Caller:
items(snacks, key = { it.id }) { snack ->
SnackRow(snack, tags = snack.tags)
}
// WRONG because: List<String> is an unstable parameter under inference; every scroll-driven recomposition recomposes the row body even though the snack didn't change.
// RIGHT
@Immutable
data class Snack(val id: Long, val name: String, val tags: ImmutableList<String>)
@Composable
fun SnackRow(snack: Snack) { /* ... */ }
items(snacks, key = { it.id }, contentType = { it::class }) { snack ->
SnackRow(snack)
}
Cross-reference: ../../stability/stabilizing-compose-types/SKILL.md.
LazyVerticalGrid with mixed spans// RIGHT — keys + contentType apply to grids identically
LazyVerticalGrid(columns = GridCells.Fixed(2)) {
items(
items = feed,
key = { it.id },
contentType = { it::class },
span = { item -> if (item is FeedItem.Header) GridItemSpan(maxLineSpan) else GridItemSpan(1) },
) { item ->
when (item) {
is FeedItem.Header -> HeaderRow(item, Modifier.animateItem())
is FeedItem.Card -> CardCell(item, Modifier.animateItem())
}
}
}
key for every items(...) block where item identity outlives a single composition (effectively: every list backed by domain objects).UUID.randomUUID() evaluated per emission, MUST NOT use hashCode() of a mutable object.contentType for heterogeneous lists (cards + headers + ads, etc.). Use a stable type discriminator such as it::class or a sealed enum.Modifier.animateItem() without a stable key — the animation silently no-ops.../../stability/diagnosing-compose-stability/SKILL.md before blaming the lazy layout. An unstable item parameter cancels every gain from key/contentType.items { } in extra inline composable wrappers (Row { items { } }) hoping to "force" skippability — Row/Column/Box are NOT restartable/skippable to begin with (skydoves hot take #3).../configuring-lazy-prefetch/SKILL.md for high-velocity scroll surfaces only after item-level fixes are in place.Modifier.animateItem() runs the expected fade and placement animation on inserts and removals.@TraceRecomposition on the item composable shows recompositions only on real state changes, not on every scroll-driven invalidation.composables.txt) shows the item composable as restartable skippable with all parameters stable or runtime.Generate breadboard circuit mockups and visual diagrams using HTML5 Canvas drawing techniques. Use when asked to create circuit layouts, visualize electronic component placements, draw breadboard diagrams, mockup 6502 builds, generate retro computer schematics, or design vintage electronics projects. Supports 555 timers, W65C02S microprocessors, 28C256 EEPROMs, W65C22 VIA chips, 7400-series logic gates, LEDs, resistors, capacitors, switches, buttons, crystals, and wires.
> Use when a HyperFrames composition needs seek-safe 2D/3D keyframes, GSAP timelines, CSS keyframes, Anime.js, WAAPI, FLIP, paths, masks, SVG morph/draw, text trails, 3D depth, or `hyperframes keyframes` diagnostics. Don't use for broad scene strategy, brand design, media sourcing, captions, or general video planning.
Analyze images, websites, and Figma files to extract their design and generate a `design.md` with token system, component inventory, and reconstruction notes. Use this skill whenever the user wants to understand, document, replicate, or audit the design of something visual: a screenshot, a URL, a Figma link, a Pinterest reference, a mockup, a competitor's site, a component, a dashboard, a landing page. Also when they ask 'extract the design system from X', 'document the style of Y', 'analyze this visually', 'convert this image into tokens', 'help me replicate this design', 'what palette does this site use', 'how is this built'. Also for single elements: 'copy this navbar', 'recreate this illustration', 'give me a prompt to regenerate this graphic' — element mode outputs a focused element.md, with token-grounded image-model prompts when the element is visual art. If the user brings any visual source and wants to understand it at a design level — this skill should activate.
Premium brand-kit image generation skill for creating high-end brand-guidelines boards, logo systems, identity decks, and visual-world presentations. Trained for minimalist, cinematic, editorial, dark-tech, luxury, cultural, security, gaming, developer-tool, and consumer-app brand systems. Optimized for intentional logo concepting, refined composition, sparse typography, strong symbolic meaning, premium mockups, art-directed imagery, and flexible grid layouts.
Elite mobile app image-generation skill for creating premium, app-native screen concepts and flows. Designed for iOS, Android, and cross-platform mobile products. Prioritizes clean hierarchy, comfortably readable text, strong multi-screen consistency, controlled color palettes, non-generic creative direction, textured surfaces, image-led composition, tasteful custom iconography, and clean phone mockup framing. By default, screens should be shown inside a subtle premium iPhone or similar phone mockup with a visible frame, while the main focus stays on the app content itself. This skill generates images only. It does not write code.
Optimize web performance: bundle size, images, caching, lazy loading, and overall page speed. Use when site is slow, reducing bundle size, fixing layout shifts, improving Time to Interactive, or optimizing for Lighthouse scores. Triggers on: web performance, bundle size, page speed, slow site, lazy loading. Do NOT use for Core Web Vitals-specific fixes (use core-web-vitals), running Lighthouse audits (use perf-lighthouse), or Astro-specific optimization (use perf-astro).
| Premium brand-kit image generation skill for creating high-end brand-guidelines boards, logo systems, identity decks, and visual-world presentations. Trained for minimalist, cinematic, editorial, dark-tech, luxury, cultural, security, gaming, developer-tool, and consumer-app brand systems. Optimized for intentional logo concepting, refined composition, sparse typography, strong symbolic meaning, premium mockups, art-directed imagery, and flexible grid layouts.
| Official GSAP skill for performance — prefer transforms, avoid layout thrashing, will-change, batching. Use when optimizing GSAP animations, reducing jank, or when the user asks about animation performance, FPS, or smooth 60fps.
Take skydoves/optimizing-lazy-layouts from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.