agentleFS
chrisgliddonchrisgliddon/bevy-skills/skills/bevy-migration-0-17-to-0-18/SKILL.md

bevy-migration-0-17-to-0-18

Skill29 starsChanged 13 days agoMITView on GitHub

Use when upgrading from Bevy 0.17 to Bevy 0.18, when an LLM writes `EventReader`/`EventWriter`/`Trigger<E>` instead of `MessageReader`/`MessageWriter`/`On<E>`, when `mesh.insert_attribute` fails to compile, when `AmbientLight` no longer works as a resource, or when `Camera { target: ... }` errors on the `target` field. Index of every breaking 0.17→0.18 change.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

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

    SimpleExecutor was removed: Add .before(), .after(), or .chain() when a later system must observe commands from an earlier one in the …

    line 117
  2. Very specific:

    Reflect attribute syntax tightened: Only #[reflect(...)] (parens) is accepted.

    line 137
  3. Very specific:

    AnimationTarget split into two components: Use (AnimationTargetId(id), AnimatedBy(player)).

    line 138

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill 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
SKILL.md

Bevy 0.17 → 0.18 — Migration cheat sheet

Released 2026-01-13. Apply top-down — earlier items break later builds if skipped.

When to use this skill

  • Compiler errors after a bevy = "0.17" → bevy = "0.18" bump.
  • LLM emits 0.17-era API names from training data.
  • Auditing a PR for hidden 0.18-incompatible patterns.
  • Hot-fixing a downstream crate that hasn't shipped 0.18 yet.

The renames you'll hit first

Observers: Trigger<E> → On<E>

// 0.18 (also 0.17)
fn on_hit(on: On<Damage>) { let _e = on.event(); }

Buffered events: Event → Message

// 0.18
#[derive(Message)] struct Tick;
// app.add_message::<Tick>();
fn read(mut reader: MessageReader<Tick>) { for _ in reader.read() {} }
fn write(mut w: MessageWriter<Tick>) { w.write(Tick); }

Event is now reserved for observed, entity-targeted events (#[derive(EntityEvent)]).

Camera { target: ... } → RenderTarget component

// 0.18
commands.spawn((Camera3d::default(), RenderTarget::Image(handle.into())));

AmbientLight is no longer a resource

// 0.18: global default
app.insert_resource(GlobalAmbientLight { brightness: 2000.0, ..default() });
// Per-camera override:
commands.spawn((Camera3d::default(), AmbientLight { brightness: 5000.0, ..default() }));

MaterialPlugin<M> config moved to trait methods

// 0.18
impl Material for M {
    fn enable_prepass() -> bool { false }
    fn enable_shadows() -> bool { false }
}
app.add_plugins(MaterialPlugin::<M>::default());

AsBindGroup::label() is now required

impl AsBindGroup for MyMaterial {
    fn label() -> &'static str { "MyMaterial" }
}

AssetLoader must #[derive(TypePath)]

#[derive(TypePath)]
struct MyLoader;
impl AssetLoader for MyLoader { /* ... */ }

Mesh::insert_attribute is panicky — prefer try_insert_attribute

// 0.18 safe form — returns Result<(), MeshAccessError>
mesh.try_insert_attribute(Mesh::ATTRIBUTE_POSITION, positions)?;

State::set always triggers a transition

// 0.18: always fires OnExit + OnEnter. Guard with:
next_state.set_if_neq(GameState::Playing);

glTF coordinate conversion replaced

// 0.18
GltfPlugin {
    convert_coordinates: GltfConvertCoordinates {
        rotate_scene_entity: true,
        rotate_meshes: true,
    },
    ..default()
}

Schedule executor

SimpleExecutor was removed. Add .before(), .after(), or .chain() when a later system must observe commands from an earlier one in the same schedule. The remaining executors use those dependencies to place ApplyDeferred; ambiguity detection remains separately configurable. See references/schedule-renames.md.

Topics

TopicReference
On<E>, MessageReader/MessageWriter, Entity::index, tick moves, Resource 'static, reflect syntax, hierarchy helpersreferences/ecs-renames.md
RenderTarget component, GlobalAmbientLight, MaterialPlugin, AsBindGroup::label, gizmos, bevy_gizmos_render splitreferences/render-renames.md
AssetLoader + TypePath, LoadContext::asset_path, asset_bytes reader pattern, AssetSource channel, try_insert_attributereferences/asset-renames.md
SimpleExecutor removal, ScheduleBuildError variants, State::set always-fires, or combinator, glTF coordinatesreferences/schedule-renames.md
animation → gltf_animation, picking-backend renames, documentation → reflect_documentation, input opt-inreferences/cargo-feature-renames.md

Gotchas

  • bevy_gizmos was split. Render-side code now lives in bevy_gizmos_render. Update import paths if you used internals.
  • Resource no longer allows non-'static lifetimes. A Resource with a borrowed field will fail to compile.
  • Reflect attribute syntax tightened. Only #[reflect(...)] (parens) is accepted; brace/bracket forms are rejected.
  • AnimationTarget split into two components. Use (AnimationTargetId(id), AnimatedBy(player)).
  • Atmosphere now references a Handle<ScatteringMedium> — what was a flat struct is now indirect.

See also

  • bevy-migration-0-18-to-0-19 — the next release's rename catalogue.
  • bevy-core-concepts — schedule executor change in detail.
  • bevy-cargo-features — feature rename table.
  • bevy-ecs-systems — Message/MessageReader/MessageWriter ergonomics.

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.