@jgengine/core 0.33.0 → 0.35.0

package/skills/jgengine-world/capabilities.md

@@ -28,6 +28,10 @@ Reach for these before hand-rolling. Each row is *the thing you need* → *the p
 
 - `measureAir` (function) · `import { measureAir } from "@jgengine/core/physics/handlingProbe"`
 
+## ambience-zones — a scene declares looping sound beds keyed to world volumes; walking between them crossfades bed levels on the existing buses instead of cutting.
+
+- `AmbienceZoneDef` (interface) · `import { AmbienceZoneDef } from "@jgengine/core/audio/ambience"`
+
 ## anim-graph — data-first animation state machine with blend trees, crossfades, layers, and clip events
 
 - `createAnimGraphRuntime` (function) · `import { createAnimGraphRuntime } from "@jgengine/core/anim/animGraph"`
@@ -109,6 +113,10 @@ Reach for these before hand-rolling. Each row is *the thing you need* → *the p
 
 - `defineAttackMoveOrder` (function) · `import { defineAttackMoveOrder } from "@jgengine/core/world"`
 
+## audio-occlusion — an emitter behind blocking world geometry is attenuated and darkened by a budgeted line test, and falls back to distance-only once the tick's budget is spent.
+
+- `AudioOcclusionConfig` (interface) · `import { AudioOcclusionConfig } from "@jgengine/core/audio/occlusion"`
+
 ## audio-voice-budget — bound simultaneous sound playback with per-sound caps, priorities, and deterministic stealing
 
 - `createVoiceAllocator` (function) · `import { createVoiceAllocator } from "@jgengine/core/audio/voiceAllocator"`

package/skills/jgengine-world/reference.md

@@ -413,7 +413,7 @@ Replace behavior data through `ctx.scene.entity.update(id, { behaviors })` or en
 
 Compose, don't replace: scale `MobBrainConfig.aggroRadius` by `perceptionScale`, route `acquireTarget` scoring through `pickScored`, gate `advancePursuit` target swaps behind the reaction gate.
 
-**Drive** — `ai/driver` is the vehicle consumer of the profile: `driveStep(state, dt, pose, goal, profile, tuning, rng, obstacleAhead?)` turns a pose + goal into the shared `AxisInput` (`throttle`/`brake`/`steer`) for `tickDrivableVehicle`. The profile decides how well the car drives: the steering target is a delay-line resampled every `reactionSeconds` (easy chasers cut to where the target *was*), corners shed speed instead of orbiting at full throttle, obstacle distance is perceived late by `speed × reactionSeconds`, the wheel wobbles by `executionJitter` (resampled on an interval, so it reads as a loose wheel), and a car grinding a wall backs out with counter-steer after a tier-scaled delay (brake past standstill is the sim's reverse). `createDriverState` is one small serializable object; the caller owns the sim, collision, and all spatial queries. For streets and circuits, `pathTargetAhead(path, x, z, lookahead, loop?)` walks a road/route polyline to the pure-pursuit point ahead — feed it back in as the goal each tick and the same brain follows authored roads; give sharper tiers a longer lookahead.
+**Drive** — `ai/driver` is the vehicle consumer of the profile: `driveStep(state, dt, pose, goal, profile, tuning, rng, obstacleAhead?)` turns a pose + goal into the shared `AxisInput` (`throttle`/`brake`/`steer`) for `tickDrivableVehicle`. The profile decides how well the car drives: the steering target is a delay-line resampled every `reactionSeconds` (easy chasers cut to where the target *was*), corners shed speed instead of orbiting at full throttle, obstacle distance is perceived late by `speed × reactionSeconds`, the wheel wobbles by `executionJitter` (resampled on an interval, so it reads as a loose wheel), and a car grinding a wall backs out with counter-steer after a tier-scaled delay (brake past standstill is the sim's reverse). `createDriverState` is one small serializable object; the caller owns the sim, collision, and all spatial queries. For streets and circuits, `pathTargetAhead(path, x, z, lookahead, loop?)` walks a road/route polyline to the pure-pursuit point ahead — feed it back in as the goal each tick and the same brain follows authored roads; give sharper tiers a longer lookahead. Catch-up pacing is a `PaceBand` on the goal: the caller measures `paceGap` (positive = ahead of the reference) and the band's `margin`/`tolerance` leave a dead zone, `ease` sheds pace for a leader over `easeAt`, `push` adds it for a chaser over `pushAt`, and `scale` is the per-car share. `paceBandFactor(band, gap)` is the same rule on its own; the push never outdrives `tuning.maxSpeed`.
 
 ### Snapshot and restore on movement and visibility handles