usd-single-layer-namespace-source

name	usd-single-layer-namespace-source
description	Use this skill when implementing or verifying the degenerate namespace source that wraps one document-model layer for stage population.
metadata	{"author":"NVIDIA"}

usd-single-layer-namespace-source

Use this skill when implementing or verifying the degenerate namespace source that wraps one document-model layer for stage population.

Spec Sources

• specification/stage_population/README.md
• specification/document_data_model/README.md

Pinned tag / commit: v1.0.1

Provides

• A NamespaceSource capability backed by exactly one document-model Layer
• Layer field lookup for root layer metadata
• Prim input lookup by absolute prim path
• Property input lookup by property path
• Child prim names from primChildren
• Property names from properties
• Source capability declaration as single_layer
• Ordering declaration as raw single-layer child/property lists
• Diagnostics passthrough from the layer/opening boundary

Contract

This skill owns contracts/handles/single-layer-namespace-source.handle.json and implements the general namespace-source capability described by contracts/capabilities/namespace-source.json.

The point of this node is to avoid making usd-stage-population depend directly on document-model storage. Stage population consumes a namespace view; today that view is backed by one layer, and later a composition node can provide the same capability after evaluating sublayers, references, payloads, variants, inherits, specializes, relocates, and load state.

A single-layer namespace source is intentionally degenerate:

• each path has zero or one local spec opinion
• field values are authored layer values
• child and property lists come from the document-model child-list fields
• no listOp composition is performed
• no value resolution is performed
• no schema fallback properties are injected
• no inactive filtering is performed
• child and property names are raw layer lists for stage population to order

The source must declare those limits explicitly through its capability inspection surface. A generated target must not silently use the single-layer source as evidence that composition or value resolution is implemented.

Boundary Guards

Consume usd-document-model for layer/spec/field storage. Do not duplicate the layer as a separate map of JSON specs.

Consume usd-paths for all path identity and path-kind checks. Adapter string paths may be parsed at the boundary, but target code should pass path handles or equivalent path references.

Consume usd-tokens and the document-model field-token vocabulary for field names, child names, property names, and token-valued fields. Do not introduce a provider-local string identity domain.

Do not perform stage population. This skill exposes source inputs; it does not create stage nodes, apply population masks, prune inactive descendants, or build traversal/lookup indexes.

Do not perform composition, value resolution, asset resolution, schema fallback evaluation, model hierarchy traversal, or instancing.

Test Obligations

• source capability declaration reports single_layer
• root layer field lookup
• prim lookup for existing and missing paths
• property lookup for attributes and relationships
• child/property name enumeration follows document-model stored lists
• active=false is exposed as a field and not interpreted locally
• authored composition arc fields remain inert source-visible fields
• all cases in goldens/unit/usd-namespace-source/single-layer-namespace-source.json