SpatialML App Workflow

Add SpatialML to an existing Kotlin Spatial SDK app, then validate the integration. SpatialML is not a separate project type and this skill must not replace Kotlin SDK setup. Unity SpatialML is owned by the PICO Unity Agentic Tools plugin, not by this workflow.

Recognize SpatialML Intent

Runtime service logs, zero/stale outputs, global tensor probes, debug-flag changes, and readback permission failures start with spatialml-debugging, including requests that also mention tensor mappings or execution order. Resume this skill when diagnosis identifies a graph or package edit.

Activate this workflow for either explicit SpatialML artifacts or an implicit spatial-inference feature. An implicit feature normally combines: a PICO/spatial app, sensor or tensor input, an ML operation, and an app-visible or pipeline-visible result. Examples include detecting objects in VST camera frames, segmenting the room image, estimating pose from camera input, classifying audio from the microphone, or running a custom .tflite model on device.

Do not activate SpatialML for an adjacent capability by itself: