NVIDIA-Omniverse
- 362 skills
- 0 followers
- 245 repo stars
- last week last updated
- ▌ Content Workflow Mesh Segmentation 2 · nvidia-omniverse bundleSegment one fused triangular USD mesh into logical mesh prims by identifying visible parts, freezing deterministic geometry fragments, choosing per part between a registered image-mask seed produced through the image-generation skill and direct agentic signed fragment selection, and falsifying each provisional result with signed confuser evidence and held-out multi-view review before locking it. Use for fused-mesh semantic or material segmentation when Codex must control fallible image-generated masks and refine whole-fragment labels.
- ▌ Content Workflow Scene Decomposition 2 · nvidia-omniverse bundleDecompose a large OpenUSD scene into finalized processable representatives, instance/prototype/payload family mappings, optional extracted assets, and a sealed Workflow 1 handoff. Use for the decomposition phase of a large-scene run, for repeated or heavily instanced scenes, or when material and physics tasks need explicit scene partitions.
- ▌ Content Workflow Asset Task Processing 2 · nvidia-omniverse bundlePlan and execute one or more domain tasks over finalized decomposed scene assets, with an immutable eligible work matrix, independent per-item results, agent-chosen ordering, durable decision memory, and a sealed Workflow 2 handoff. Use after scene decomposition for material, physics, articulation, geometry, labeling, or mixed asset processing.
- ▌ Content Workflow Physics External Tuning 2 · nvidia-omniverseUse when a long-running coding agent owns the outer loop of external-runtime (BYOR) physics refinement — requesting budgeted qualification-gated tune-external sweeps through the wrapper broker, visually judging rendered rollout evidence against a behavior goal, revising the active parameter search, and writing digest-bound accept/revise/stop decisions.
- ▌ Brev CLI · nvidia-omniverse bundleManage Brev instances safely from the Brev CLI. Use when a workflow needs to list, create, access, copy to, execute on, port-forward, stop, or delete Brev instances for Content Agents dependency endpoints.
- ▌ Fixed Pipeline · nvidia-omniverse bundleUse when the user explicitly requests a fixed-pipeline Content Agent CLI, service client, deployment, benchmark, or maintenance workflow under apps/: material-agent-cli, material-agent-client, physics-agent-cli, physics-agent-client, joint-agent-cli, joint-agent-client, joint-agent-validation, texture-agent-cli, texture-agent-client, validation-agent-cli; deploy-collection, deploy-material-agent-docker, deploy-material-agent-brev, deploy-physics-agent-docker, deploy-physics-agent-brev, deploy-texture-agent-docker, deploy-texture-agent-brev, deploy-ovrtx-docker, deploy-embeddings-brev, deploy-image-gen-brev, or deploy-qwen-vlm-brev.
- ▌ Usd CLI · nvidia-omniverseRun explicit low-level USD scene inspection, editing, rendering, material binding, physics authoring, and raw validation with the in-tree usd-cli package. Use when the user asks for direct scene mechanics on a USD/USDA/USDC/USDZ asset or asks to set up usd-cli. Domain workflows must route to their content-workflow-* skill; usd-cli never owns workflow policy, orchestration, artifacts, recovery, or acceptance.
- ▌ Image Generation · nvidia-omniverse bundleGenerate or edit raster images for agentic workflows through the coding agent's companion image generator by default or an explicitly configured World Understanding image backend. Use when a workflow needs text-to-image, conditioned image editing, or a recorded generated-image artifact with provider, prompt, input, and output provenance.
- ▌ Content Workflow CLI · nvidia-omniverseDefault launcher for supported Content Agent tasks from the repository root, including deterministic Geometry handoff, conversion, material or physics authoring, texture, segmentation, validation, large scenes, durable articulation, composed assets, review, and resume. Use fixed-pipeline app CLIs only when the user explicitly requests fixed control flow, YAML configuration, REST, benchmarking, or a named fixed-pipeline command.
- ▌ Content Texture Scope · nvidia-omniverseInspect an explicit USD material or prim scope, existing bindings, and UV readiness before texture generation. Use for bounded Texture workflow inspection and UV-preparation decisions.
- ▌ Content Sdf Operations · nvidia-omniverse bundleUse when a content agent needs bounded signed- or unsigned-distance-field operations such as mesh rasterization, field booleans, offsets, filtering, sampling, topology conversion, surface extraction, or admitted field output through an in-process backend.
- ▌ Content Workflow Asset · nvidia-omniverse bundleRun one prompt-specific asset workflow through a sole outer reasoner that supplies and freezes a typed public-leaf dependency graph. Use for selective, resumable asset composition without a built-in domain or stage order. Explicit compatibility-fixed mode still composes Geometry preparation, articulation review, material assignment, bounded texture generation, physics authoring, validation, and final USDZ packaging as one durable single-asset workflow.
- ▌ Geometry Source Intake · nvidia-omniverse bundleClassify and prepare one geometry.source.v1 bundle or CAD, mesh, USD, URDF, MJCF, GLTF, GLB, or 3MF artifact while preserving units, provenance, editability, representation roles, and source fidelity. Use before Geometry, repair, segmentation, material, physics, or SimReady workflows need canonical geometry input.
- ▌ Content Texture Publish · nvidia-omniverseValidate a Texture decision chain and publish a portable scoped USD result. Use after all bounded generation and visual-review steps reach a terminal decision.
- ▌ Content Texture Quality · nvidia-omniverseCollect matched Texture evidence, optionally request advisory critique, and record the outer multimodal review. Use after explicit candidate generation.
- ▌ Content Workflow Physics · nvidia-omniverse bundleOwn agent-driven physics planning, policy, component decisions, guarded repair, authoring, optional attested VoMP mass properties, simulation-backed validation, refinement, acceptance, and durable artifacts for a USD asset, using usd-cli only for low-level scene operations.
- ▌ Content Workflow Texture · nvidia-omniverse bundleUse one outer Codex or Claude reasoner to select and invoke focused, digest-bound Texture capabilities without a nested Texture workflow controller.
- ▌ Geometry Evidence Review · nvidia-omniverse bundleGather and interpret geometry inspection, OVRTX render, topology, USD, projection, source-fidelity, and protected-feature evidence for one candidate. Use when an interactive session or workflow must decide whether geometry is reviewable, what additional evidence is needed, which defect is present, or whether to accept, refine, repair, or block a candidate.
- ▌ Product Image Decomposer · nvidia-omniverse bundleDecompose product reference images into CAD-ready component crops, semantic part-plan evidence, exploded-view evidence, six-view prompts, and a reassembly manifest before geometry generation or reconstruction.
- ▌ Content Texture Candidate · nvidia-omniverseGenerate, compare, and preview one bounded texture candidate set for exact texture-unit IDs. Use for initial generation or failed-unit regeneration without unrelated mutation.
- ▌ Content Workflow Geometry · nvidia-omniverse bundleComplete a provider-neutral Geometry handoff from CAD, mesh, generated, or existing USD. Use when a workflow needs source lineage, prepared and optimized geometry, OVRTX and USD evidence, semantic parts or parameters, and content_agents_manifest.json without claiming final physics, runtime behavior, or SimReady status.
- ▌ Content Workflow Material · nvidia-omniverse bundleDefault material-authoring workflow for supported, unqualified USD material tasks. Use with usd-cli when a long-running coding agent assigns visual or non-visual materials, uses libraries, performs iterative visual review, and produces canonical artifacts and validation evidence; use material-agent-* only for explicit fixed pipeline requests.
- ▌ Content Workflow Simready · nvidia-omniverse bundleUse after USD conversion, material and physics authoring, or Joint Agent publication to preflight SimReady Foundation, formally validate the current staged USD or USDZ, conditionally conform it from the failed validation report, revalidate the conformed output, run Gate 3B validation, and preserve durable non-blocking remediation evidence.
- ▌ Content Articulation Review · nvidia-omniverseSelect, reject, or safely edit every articulation candidate through a digest-bound decision patch. Use before human review or accepted-only authoring.
- ▌ Content Workflow Validation · nvidia-omniverse bundleValidate USD assets through deterministic preparation, one decision-only Codex or Claude planning child, outer-validated exact adapters, independent assessment/review, and non-mutating terminal receipts. Use for agentic standalone Validation; use focused commands for explicit compatibility operations and the fixed-pipeline umbrella for config-driven Validation Agent runs.
- ▌ Content Workflow Large Scene · nvidia-omniverse bundleCoordinate a large OpenUSD scene through decomposition, per-asset domain processing, and original-topology collection using durable run state and deterministic handoff gates. Use when a scene is too large or repetitive to process monolithically, when material/physics/other tasks must run over decomposed representatives, or when resuming or repairing a three-phase large-scene run.
- ▌ Content Articulation Proposal · nvidia-omniversePropose bounded revolute or prismatic articulation candidates from exact inspection evidence. Use to reason about endpoints, axes, limits, confidence, and native readiness before review.
- ▌ Content Workflow Articulation · nvidia-omniverseRun durable Joint Agent articulation-v1 workflows that bind usd-cli inspection evidence, pause for explicit review, resume without repeating completed phases, author only approved revolute or prismatic joints, and verify exact saved USDZ readback. Use when prompt-driven joint, drawer, hinge, slider, mechanism, or articulation work needs the isolated agentic workspace.
- ▌ Content Articulation Authoring · nvidia-omniverseAuthor only approved articulation candidates, then validate saved-stage identity and joint-graph readback. Use after agent and required human review are durably bound.
- ▌ Content Articulation Inspection · nvidia-omniverseInspect bounded USD hierarchy, existing articulation, and usd-cli evidence before proposing joints. Use for exact Articulation candidate preparation and targeted reinspection.
- ▌ Content Workflow Convert To Usd · nvidia-omniverseUse before USD-based content workflows when a source asset must be routed to OpenUSD through the existing content-workflow-cli conversion workflow, including dependency preflight, existing USD passthrough, CAD/mesh, URDF, MuJoCo/MJCF, durable reports, blocked results, and downstream USD handoff.
- ▌ Content Workflow Geometry Repair · nvidia-omniverse bundleComplete a source-preserving Geometry repair workflow for imported CAD, mesh, USD, URDF, or MJCF geometry with bounded topology work, collision separation, and explicit evidence. Use when a user or parent workflow needs a durable certified, conditional, rejected, blocked, failed, or cancelled geometry-repair outcome rather than one isolated diagnosis or edit.
- ▌ Content Workflow Scene Collection · nvidia-omniverse bundleCollect validated decomposed-asset task results back onto original OpenUSD topology through domain-specific projection, representative propagation, conflict harmonization, independent domain layers, composition, and a sealed Workflow 3 handoff. Use after asset-task processing or when repairing collection for material, physics, articulation, geometry, or mixed-domain outputs.
- ▌ Content Workflow Mesh Segmentation · nvidia-omniverse bundleSegment one fused triangular USD mesh into logical mesh prims by identifying visible parts, freezing deterministic geometry fragments, choosing per part between a registered image-mask seed produced through the image-generation skill and direct agentic signed fragment selection, and falsifying each provisional result with signed confuser evidence and held-out multi-view review before locking it. Use for fused-mesh semantic or material segmentation when Codex must control fallible image-generated masks and refine whole-fragment labels.
- ▌ Content Workflow Scene Decomposition · nvidia-omniverse bundleDecompose a large OpenUSD scene into finalized processable representatives, instance/prototype/payload family mappings, optional extracted assets, and a sealed Workflow 1 handoff. Use for the decomposition phase of a large-scene run, for repeated or heavily instanced scenes, or when material and physics tasks need explicit scene partitions.
- ▌ Content Workflow Asset Task Processing · nvidia-omniverse bundlePlan and execute one or more domain tasks over finalized decomposed scene assets, with an immutable eligible work matrix, independent per-item results, agent-chosen ordering, durable decision memory, and a sealed Workflow 2 handoff. Use after scene decomposition for material, physics, articulation, geometry, labeling, or mixed asset processing.
- ▌ Content Workflow Physics External Tuning · nvidia-omniverseUse when a long-running coding agent owns the outer loop of external-runtime (BYOR) physics refinement — requesting budgeted qualification-gated tune-external sweeps through the wrapper broker, visually judging rendered rollout evidence against a behavior goal, revising the active parameter search, and writing digest-bound accept/revise/stop decisions.
- ▌ Combining Scene With Runtime Prims · nvidia-omniverseLoading a base USD scene with additional runtime prims in tests. Use when a doc test needs to combine a scene file with RenderProducts or other prims that aren't in the original scene.
- ▌ Basic Workflow · nvidia-omniverseCreate an ovphysx instance, attach an ovstage scene, step simulation, and clean up. Use when starting a new ovphysx project, testing basic integration, or learning the minimal workflow.
- ▌ Clone Environments · nvidia-omniverseClone one USD environment subtree into many runtime-only PhysX environments for RL-style batched simulation. Use when a caller needs replicated physics environments, clone API ordering, or target transform guidance.
- ▌ Ovphysx Output Read · nvidia-omniverse bundleRead ovphysx simulation output with the ovstage-native output-read API when results must retain ovstage identity or write back without rebuilding paths or repacking tensors. Covers Python `PhysX.read()` / `PhysX.read_tokens()`, C `ovphysx_query()` / `ovphysx_read()`, emitted attribute tokens, borrowed groups, and closed-loop physics-to-ovstage write-back.
- ▌ Tensor Bindings Cpu · nvidia-omniverseCreate tensor bindings to read and write physics simulation data on CPU using numpy arrays. Use when you need to exchange simulation state (poses, velocities, joint targets) with your application via tensors.
- ▌ Tensor Bindings Gpu · nvidia-omniverseRead and write physics simulation data on GPU using CUDA device pointers and DLPack tensors. Use when you need GPU-to-GPU tensor exchange (no CPU staging) for high-throughput RL or robotics workloads.
- ▌ Ovphysx Usd Authoring · nvidia-omniverse bundleAuthor USD physics content that ovphysx can load and simulate (rigid bodies, colliders, mass, physics scene). Use when creating or editing .usda or USD scenes for ovphysx, not when calling the runtime step or tensor API. Keywords USD authoring, UsdPhysics, PhysxSchema, RigidBodyAPI, CollisionAPI, MassAPI, PhysicsScene, .usda, simulate scene.
- ▌ Search · nvidia-omniverse bundleSearch USD Search for 3D assets. One skill, three input modes: text query, reference image, or both combined. Hits POST /search_hybrid + GET /images. Iterates on intermediate matches for visual-similarity refinement when the user wants "more like this". Use when: "search assets", "find a 3D model", "find a USD", "find similar", "more like this", "visually similar".
- ▌ Quickstart · nvidia-omniverseGet a first-time user to a real 3D asset in seconds. Asks one question — "where do you want to search?" — with three backend lanes (NVIDIA-hosted default, your own URL, or a local deployment), then hands off to /search for the actual query. The journey is the SVG at docs/quickstart-user-journey.svg: Stage 1 (entry) + Stage 2 (backend pick), then /search drives Stages 3-5. Use when: "quickstart", "try usd search", "show me what this does", "getting started", "what is usd search", "tour", "first time".
- ▌ Inspect Asset · nvidia-omniverseDeep-inspect a single 3D asset by URL. Fetches multi-view thumbnails, per-plugin indexing status, USD scene summary (prim counts, geometry, scale, up-axis, default prim), prim hierarchy, forward + inverse dependency graphs, and optionally runs VLM relevance validation against a description. Use when the user asks to inspect / deep-dive / describe / learn about / report on a specific asset, or asks "what's in this USD", "tell me about this asset", "where is this asset used", "what does this asset reference", "is this asset indexed".
- ▌ Search In Scene · nvidia-omniverseSpatial and scene-graph queries inside a USD scene via the Asset Graph Service. Answers: what's near a point or prim (radius), what's inside a bounding box, what prims of a given type or property exist in this scene, what does this scene reference, where is this asset used, what's the prim hierarchy. Optionally chains to /search_hybrid with image_similarity_search to find assets similar to a prim's source asset. Use when the user has a scene URL plus a spatial, structural, or dependency question.
- ▌ Deploy Usdsearch · nvidia-omniverse bundleStand up a USD Search deployment the user controls. Branches on local docker-compose vs. helm on Kubernetes. Local branch covers the full quickstart compose stack with GPU/VLM/storage-backend questionnaires. Helm branch covers the usdsearch chart at helm/usdsearch. Use when: "deploy usdsearch", "run the stack myself", "helm install usdsearch", "start the stack", "run usdsearch locally", "host my own".
- ▌ Usd Property Catalog · nvidia-omniverse bundleDiscover which USD properties a USD Search deployment's corpus actually carries, and turn that into local config: hits GET /search/stats/usd_properties, then builds a ranked usd_property_catalog.yaml (the grounding source for the built-in LLM query parser) plus ready-to-use search_fields.generated.yaml filter stanzas. Has a P0 "gap audit" mode that reports which target properties are present/absent and pulls sample assets for parsing validation. Use when: "what USD properties can I filter on", "discover usd properties", "build a property catalog", "which metadata is in the index", "audit P0 metadata", "ground the query parser on real properties".
- ▌ Explorer Remote Proxy · nvidia-omniverseRun the locally-built Explorer UI against a REMOTE, fully-hosted USD Search instance by proxying every API call through the local gateway. No local deepsearch-api / OpenSearch / SigLIP2 / Asset Graph Service — only the Explorer + gateway run locally, so search, thumbnails, dependency graphs, and downloads all work exactly as on the hosted instance (which already has a populated index + AGS + storage). Use when: you want to test/demo the local Explorer with WORKING dependency graphs and downloads but don't want to (or can't) build a local index + Asset Graph Service. Default target is the public staging instance (api.staging.deepsearch-horde.nvidia.com).
- ▌ Perf Triage · nvidia-omniverseEntry point for any Omniverse Kit or OpenUSD performance question. Triggers on "my scene is slow", "low FPS", "scene takes forever to load", "GPU crash", "out of memory", "performance problem", "how do I speed up", "optimize my scene". Classifies the problem and routes to the matching specialist skill.
- ▌ Usdrt Advisor · nvidia-omniverseMigrate Kit code from Python USD traversal to USDRT stage queries, and from TfNotice callbacks to USDRT change tracking or UsdWatcher. Triggers on "USDRT", "Fabric", "stage traversal slow", "TfNotice", "stage query", "change tracking", "UsdWatcher", "Python performance in Kit". Provides complete runnable code samples.
- ▌ Ujitso Manager · nvidia-omniverseUJITSO derived-data cache configuration and troubleshooting — processors, deployment models, cache locations, "first load is slow" expectations. Triggers on "UJITSO", "derived data cache", "first load slow", "scene loading cache", "cache warmup", "cache miss", "DDCS", "OmniHub".
- ▌ Profiling Guide · nvidia-omniverseProfiling Omniverse Kit applications — F8/Profiler Window, Tracy, the activity timeline for load-time analysis, Chrome Tracing, Nsight Systems, and the RTX performance heatmap. Triggers on "how to profile", "where is the bottleneck", "Tracy", "F8", "GPU usage", "instrument code", "performance traces", "activity timeline". Use this before tuning any settings.
- ▌ Rendering Tuner · nvidia-omniverseRTX rendering performance — RT 2.0 vs Path Tracing, DLSS, glass, bounces, multi-GPU. Triggers on "render settings", "RTX", "DLSS", "RT 2.0 vs path tracing", "bounces", "samples per pixel", "refraction", "glass performance", "multi-GPU", "improve FPS", "scaling resolution".
- ▌ Fsd Configurator · nvidia-omniverseFabric Scene Delegate (FSD) configuration — the small set of settings partners actually need to consider, with the mergeMaterials default (off by default since Kit 108) context and the deprecated instanceCompactTransforms warning. Triggers on "FSD", "Fabric Scene Delegate", "Fabric", "mergeMaterials", "enableRendererInstancing", "populateAllAuthoredAttributes", "useFabricSceneDelegate", "instanceCompactTransforms".
- ▌ Memory Optimizer · nvidia-omniverseGPU and system memory, threading, async rendering, and texture sizing. Triggers on "GPU memory", "VRAM", "out of GPU memory", "RAM usage", "system memory", "memory leak", "thread count", "8K texture", "DDS", "texture compression", "viewport memory".
- ▌ Cloud Perf Advisor · nvidia-omniverseCloud deployment performance — caching architecture, pre-warming strategy, container cold starts, NVCF, OKAS, UCC, GXCache, DDCS. Triggers on "cloud performance", "container startup", "K8s profiling", "cloud caching", "cold start", "NVCF", "GXCache", "OKAS", "UCC", "deployment performance", "Kit App Streaming".
- ▌ Usd Scene Structure · nvidia-omniverseUSD scene-structure performance — payloads, layers, instancing strategy, file formats, prim count reduction, composition arc selection. Triggers on "payload", "sublayer vs reference", "layer count", "instanceable", "point instancer", "USD structure", "prim count", "scene is heavy", "USDZ", "USDA vs USDC".
- ▌ Content Prep Advisor · nvidia-omniverseContent preparation, validation, CAD import, materials pipeline, and physics — preventing performance issues at the source. Triggers on "import CAD", "CAD to USD", "validate USD", "asset validator", "usdchecker", "material pipeline", "primitive collider", "mesh collider", "physics performance", "PhysX", "DDS texture".
- ▌ Scene Optimizer Guide · nvidia-omniverseBottleneck-driven Scene Optimizer workflow — operations, performance validators, trade-offs, and when not to merge meshes. Triggers on "Scene Optimizer", "optimize scene", "merge meshes", "decimate", "deduplicate", "reduce triangles", "performance validator", "asset validator scene".
- ▌ Troubleshooting Router · nvidia-omniverseCommon performance scenarios with step-by-step resolution and routing back to the relevant specialist skill. Triggers on "GPU crash", "device-lost", "scene takes 30 minutes to open", "performance regression after Kit upgrade", "FPS drops to single digits", "instanced scene uses too much memory", "scene looks wrong after optimization", "Kit upgrade broke performance".
- ▌ Logging · nvidia-omniverseRoute ovstage's diagnostic log messages (and messages from its USD support layer) to a process-global callback with a severity threshold and a RUST_LOG-style per-channel filter, and force delivery with flush_log. Use when the user asks about ovstage logging, a log callback, log severity or verbosity, channel filtering, or capturing/silencing ovstage/USD runtime log output. This is distinct from error handling (return codes) — see the error-handling skill for that.
- ▌ Runtime Loop · nvidia-omniverseHeadless load -> populate -> read -> update -> read loop with no renderer attached: open a USD scene into the ovstage runtime table, read prims back to confirm, then update the live stage two ways — write straight into the ovstage table (e.g. animate a transform), or edit the USD source and propagate it through. Use when the user wants to drive an ovstage scene headlessly, populate USD into the runtime stage, or see runtime-table vs USD-source edits reflected in reads.
- ▌ Path Dictionary · nvidia-omniverseUsing the OVX path dictionary — the shared interning layer for tokens, prim paths, and prim-path lists across OV libraries. Use when user asks about interning strings/paths, ovx_token_t / ovx_primpath_t / ovx_primpath_list_t, or sharing paths with ovstage.
- ▌ Exporting To Usd · nvidia-omniverseExport selected ovstage runtime state to USD from C or Python. Use for saved typed hierarchies, sparse overlays, reusable USD destinations, async export completion, or changed-property export. Prefer the metadata-driven typed-hierarchy workflow for ordinary USD-populated subtrees and explicit rules only for intentional mapping differences.
- ▌ Cpu Ahead Gpu Async · nvidia-omniverseovstage's asynchronous, ordinal-keyed submit/observe model: enqueue returns an op id before queued execution completes, and the CPU can run ahead. Covers the current synchronous submission, serialized backend-lane, and overlapping read/write/map limitations. Use when user asks about async operations, op ids, waiting/polling, timeouts, overlap failures, non-blocking submission, or CPU-ahead execution.
- ▌ Clone Subtree Multienv · nvidia-omniverseClone the subtree under a source prim to one or more new target paths in a single ordinal-keyed call — including the multi-environment pattern (stamp out N copies of a prototype). Use when the user asks to clone, duplicate, or copy a prim/subtree, or to spawn per-environment instances (e.g. one scene/robot per RL environment).
- ▌ Dlpack Tensor Exchange · nvidia-omniverseHow ovstage attribute data crosses the boundary as DLPack DLTensors — copy-in writes, copy-out reads, staging-backed map/unmap writes, CPU vs CUDA residency, cuda_sync GPU sync, and sparsity (index_map/mask). Use when user asks about DLTensor, tensor data exchange, GPU residency, CUDA sync, zero-copy writes, or mapping attribute storage.
- ▌ Usd To Ovstage Migration · nvidia-omniversePort code written against the USD API (pxr) to the ovstage data plane: map DefinePrim / CreateAttribute+Set / Get / RemovePrim / Sdf.ChangeBlock onto INSERT-mode writes, write_attribute, read_attributes, delete_attributes, and the batched write_attributes — and make the vectorization mental shift from one call per prim to one call per column over a prim set. Use when the user wants to migrate, port, or translate USD-API authoring code to ovstage, or asks what the ovstage equivalent of a USD call is.
- ▌ Ovstorage User Read Bytes · nvidia-omniverseUse when loading one bounded object into memory to parse, inspect, hash, or hand its bytes to another tool.
- ▌ Ovstorage User Materialize · nvidia-omniverseUse when an object is large, needs random access, or must be handed to a tool as a stable local path.
- ▌ Ovstorage User Write Safely · nvidia-omniverseUse when writing a new object without clobbering or updating an existing object with optimistic concurrency.
- ▌ Ovstorage User Delete Safely · nvidia-omniverseUse when removing objects safely, especially recursive directories or targets that need a dry-run first.
- ▌ Ovstorage User Handle Errors · nvidia-omniverseUse when an ovstorage tool returns ok false and you need to interpret the envelope error and choose the next action.
- ▌ Ovstorage User Connect Via MCP · nvidia-omniverseUse when connecting an agent client to ovstorage over MCP stdio and mapping common CLI workflows to MCP tools.
- ▌ Ovstorage User Getting Started · nvidia-omniverseUse when starting a new ovstorage session or recovering from NoRoute, NotConfigured, or Unsupported errors by inspecting configured backends and routes.
- ▌ Ovstorage Operator Debug Broker · nvidia-omniverseUse when triaging a misbehaving ovstorage-broker - covers startup failures, authz denies, listener authn mismatches, streaming write aborts, reload misfires, and drain hangs.
- ▌ Ovstorage User Choose A Backend · nvidia-omniverseUse when selecting the right ovstorage backend plugin for a task or deciding which route to configure.
- ▌ Ovstorage Operator Deploy Broker · nvidia-omniverseUse when deploying the ovstorage-broker daemon with listener authentication, authorization policy, TLS, and storage connections.
- ▌ Ovstorage User List And Paginate · nvidia-omniverseUse when enumerating objects under a prefix without buffering the whole result set into memory.
- ▌ Ovstorage Operator Monitor Broker · nvidia-omniverseUse when wiring observability for a running ovstorage-broker - covers the Prometheus /metrics surface, the 11 metric families, the tracing span fields, and the audit-safe diagnostic shape.
- ▌ Ovstorage User Authenticate To Backend · nvidia-omniverseUse when configuring credentials or driving interactive authentication for a direct ovstorage backend connection.
- ▌ Ovstorage Contributor Modify Authz Layer · nvidia-omniverseUse when changing or reviewing the built-in authentication and authorization Layer or its TOML policy engine.
- ▌ Ovstorage Contributor Run Broker Locally · nvidia-omniverseUse when starting ovstorage-broker locally for development or integration testing with the current Stack and listener-auth configuration.
- ▌ Ovstorage Contributor Verify Before Merge · nvidia-omniverseUse before opening or merging an ovstorage PR to run the project verification gate and inspect risky diffs.
- ▌ Ovstorage Contributor Regenerate C Headers · nvidia-omniverseUse when Rust changes affect the ovstorage plugin ABI header and the checked-in generated copies must be refreshed.
- ▌ Ovstorage Operator Configure Broker Policy · nvidia-omniverseUse when authoring or rotating the broker's built-in authorization policy.
- ▌ Ovstorage Contributor Author Storage Plugin · nvidia-omniverseUse when adding or reviewing a storage backend plugin for ovstorage.
- ▌ Ovstorage Contributor Services Client Conformance · nvidia-omniverseUse when checking the Storage API wire contract that ovstorage-plugin-services-client depends on.
- ▌ Ecr · nvidia-omniverseGenerate Terraform for AWS ECR repositories and EKS pull access needed by KAS/NVCF. Use after eks-cluster or eks-nodes and before ecr-mirror; this skill creates repositories only and does not mirror images.
- ▌ Ovstorage · nvidia-omniverse bundleDeploy and develop against NVIDIA Omniverse Storage APIs on any Kubernetes cluster (MicroK8s, EKS, AKS, GKE, bare metal). Two storage adapters: (1) Example Storage Adapter — Python filesystem reference implementation + Discovery for learning and custom adapter development; (2) S3/Azure Production Storage Adapter — NVIDIA pre-built S3/Azure adapter + Discovery as minimum, expandable with Event Aggregation, Event Consumer, RabbitMQ, Envoy Auth Extension, Storage Navigator, and Contour ingress. Can deploy on the developer's behalf (generating scripts + .env) or guide manual deployment. Custom adapter development covers three APIs: Storage, Notifications, and Permissions. Use when a developer asks about deploying, configuring, validating, troubleshooting, or building any component of the Omniverse Storage APIs stack.
- ▌ Nvcf Gxcache · nvidia-omniverse bundleUse for installing, debugging, or modifying NVCF GXCache across self-managed Kubernetes environments. Trigger on explicit GXCache, helm-gxcache, gxcache-init, gxcache-webhook, distributed shader cache, NVCA cache feature gates, or GXCache provider adaptation for Kubernetes clusters.
- ▌ Acr · nvidia-omniverseGenerates a bash or Terraform snippet that creates an Azure Container Registry (Standard tier) in the cluster's resource group and grants the AKS kubelet identity `AcrPull` on it. Output enables passwordless image pulls from the cluster — no `imagePullSecrets` needed. Does NOT populate the registry — see the `acr-mirror` skill for the bulk NGC → ACR mirror. Use as Stage 04 of `aks-bootstrap`, after `aks-cluster`. Trigger when the user mentions "create ACR", "container registry", "AcrPull", "passwordless pull from AKS".
- ▌ Nvcf Lls Install · nvidia-omniverse bundleUse for installing, planning, or debugging NVCF LLS across self-managed Kubernetes environments. Trigger on explicit NVCF LLS, gdn-streaming, streaming-proxy, TURN, LLS WebRTC ingress, UDP 5004 media ingress, TCP 3478 TURN ingress, or LLS provider adaptation for Kubernetes clusters. Do not use for unrelated generic streaming media work.
- ▌ AWS Infra · nvidia-omniverse<!-- SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. -->
- ▌ Eks Nodes · nvidia-omniverse<!-- SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. -->
- ▌ Iam Vpc · nvidia-omniverseGenerates a bash script or Terraform module that provisions the Azure foundation for an AKS deployment — Resource Group, VNet, node subnet, API server subnet (for VNet integration), and a user-assigned managed identity for the cluster control plane. No NSG attached to subnets by default (AKS auto-manages a NIC-level NSG). Use as Stage 01 of `aks-bootstrap`, or standalone when laying down networking before any cluster work. Trigger when the user mentions "VNet", "resource group", "subnets", "managed identity", "network setup", "iam-vpc", or asks to "set up the Azure foundation".
- ▌ Ecr Mirror · nvidia-omniverseMirror every KAS/NVCF manifest image and optional Helm OCI chart from NGC into AWS ECR using generated regsync config.