omniverse-simready
Skill8 starsChanged 8 months agoMITView on GitHub
NVIDIA Omniverse SimReady — OpenUSD standard for physically accurate 3D assets with metadata for robotics, digital twins, synthetic data generation
In this repository23 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- validate-skillsSkill
- alchemi-toolkit-opsSkill
- autoresearch-setupSkill
- bionemoSkill
- biosimspaceSkill
- cuda-quantumSkill
- cutileSkill
- daskSkill
- flashinferSkill
- materials-projectSkill
- mpiSkill
- nemotronSkill
- physicsnemoSkill
- polarsSkill
- prime-verifiersSkill
- qiskitSkill
- raySkill
- rdkitSkill
- slurmSkill
- tinkerSkill
- tritonSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 456
Naming: Use clear hierarchies
- Distinctive: line 462
Channel 1: Don't use additional channels
- Distinctive: line 463
Pattern alignment: Before packing
- Distinctive: line 466
Multi-component: Use multiple colliders
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.
Files
NVIDIA Omniverse SimReady Skill
Standard and ecosystem for physically accurate 3D assets built on OpenUSD with semantic labeling, physics properties, and metadata for robotics simulation, digital twins, and synthetic data generation.
Official Sources:
- SimReady Overview
- SimReady Specification
- Asset Creation Guide
- Semantic Labeling
- Physics Best Practices
- Isaac Sim Integration
What is SimReady?
Definition:
"Simulation-ready assets are physically accurate 3D objects that incorporate real-world properties, behaviors, and data bindings."
Built on OpenUSD:
- Universal Scene Description platform
- Modular, flexible composition
- Cross-platform compatibility
More Than Visual:
- Semantic labeling (WikiData integration)
- Physics properties (USDPhysics schema)
- Non-visual sensor attributes
- Dense captions for context
Core Requirements
Modeling Standards
Scale:
- Real-world scale
- Z-up orientation
- Export in meters
- Pivot at origin
Geometry:
- Optimal poly count
- Clean topology
- Named hierarchies
- Forward-facing in viewport
UVs:
- Non-overlapping islands
- UV Channel 1 only
- 0-1 UV space
- 512+ pixels/meter texel densityMaterial Standards
Workflow:
- PBR Metal-Rough mandatory
- UsdPreviewSurface for portability
- OmniPBR/OmniGlass/SimPBR for Omniverse
Textures:
- 4K or 8K maximum resolution
- One material per object
Shaders:
- UsdPreviewSurface (cross-platform)
- Omniverse MDL materials (advanced)Physics Requirements
USDPhysics Schema:
- Rigid body colliders
- Mass in kilograms
- Physical materials (friction, density, restitution)
- Collision meshes (no convex hulls)
Compatibility:
- PhysX
- MuJoCo
- Gazebo
- DrakeQuick Start
Download SimReady Assets
# NVIDIA Physical AI Dataset
# Download from: https://developer.nvidia.com/isaac/sim
# Isaac Sim Asset Packs
# 1000+ SimReady assets (conveyors, boxes, pallets)
# Available in Isaac SimLoad in Isaac Sim
import omni.isaac.core.utils.stage as stage_utils
from omni.isaac.core.utils.prims import create_prim
# Load SimReady asset
asset_path = "/Isaac/Environments/Simple_Warehouse/Props/S_TrafficCone.usd"
prim_path = "/World/Cone"
create_prim(
prim_path=prim_path,
usd_path=asset_path,
position=[0, 0, 0]
)Inspect SimReady Properties
from pxr import Usd, UsdPhysics, Semantics
# Open stage
stage = Usd.Stage.Open("asset.usd")
prim = stage.GetPrimAtPath("/World/Asset")
# Check physics
if prim.HasAPI(UsdPhysics.RigidBodyAPI):
rb_api = UsdPhysics.RigidBodyAPI(prim)
print(f"Mass: {rb_api.GetMassAttr().Get()}")
# Check semantic labels
if prim.HasAPI(Semantics.SemanticsAPI):
sem_api = Semantics.SemanticsAPI(prim)
labels = sem_api.GetSemanticLabelsRel().GetTargets()
print(f"Labels: {labels}")Semantic Labeling
WikiData Integration
# SimReady schema metadata
metadata = {
"class": "cup", # Human-readable
"hierarchy": "container/cup", # Taxonomy
"qcode": "Q81727", # WikiData unique ID
"sku": "PROD-12345" # Custom identifier
}
# WikiData: 100M+ objects, multi-language
# Access: https://wikidata.org
# QCode provides:
# - Synonyms
# - Taxonomic subtrees
# - Multi-language translationsApply Semantic Labels
from pxr import Semantics, Sdf
stage = Usd.Stage.Open("asset.usd")
prim = stage.GetPrimAtPath("/World/Asset")
# Add SemanticsAPI
if not prim.HasAPI(Semantics.SemanticsAPI):
Semantics.SemanticsAPI.Apply(prim)
sem_api = Semantics.SemanticsAPI(prim)
# Create semantic label prim
label_path = Sdf.Path("/World/Semantics/Labels/Cup")
label_prim = stage.DefinePrim(label_path, "SemanticsLabel")
# Set properties
label_prim.CreateAttribute("semantic:class", Sdf.ValueTypeNames.String).Set("cup")
label_prim.CreateAttribute("semantic:hierarchy", Sdf.ValueTypeNames.String).Set("container/cup")
label_prim.CreateAttribute("semantic:qcode", Sdf.ValueTypeNames.String).Set("Q81727")
# Link to asset
sem_api.GetSemanticLabelsRel().AddTarget(label_path)
stage.Save()Physics Configuration
Add Rigid Body
from pxr import UsdPhysics, PhysxSchema
stage = Usd.Stage.Open("asset.usd")
root_prim = stage.GetPrimAtPath("/World/Asset")
# Apply RigidBodyAPI to root
if not root_prim.HasAPI(UsdPhysics.RigidBodyAPI):
rb_api = UsdPhysics.RigidBodyAPI.Apply(root_prim)
# Set mass
mass_api = UsdPhysics.MassAPI.Apply(root_prim)
mass_api.GetMassAttr().Set(2.5) # kg
stage.Save()Add Collision Mesh
# Apply CollisionAPI to geometry
mesh_prim = stage.GetPrimAtPath("/World/Asset/Mesh")
if not mesh_prim.HasAPI(UsdPhysics.CollisionAPI):
collision_api = UsdPhysics.CollisionAPI.Apply(mesh_prim)
# PhysX extension (optional)
if not mesh_prim.HasAPI(PhysxSchema.PhysxCollisionAPI):
physx_collision = PhysxSchema.PhysxCollisionAPI.Apply(mesh_prim)
physx_collision.GetContactOffsetAttr().Set(0.02)
physx_collision.GetRestOffsetAttr().Set(0.0)
stage.Save()Add Physical Material
# Create physics material
material_path = Sdf.Path("/World/PhysicsMaterials/Metal")
material_prim = stage.DefinePrim(material_path, "PhysicsMaterial")
# Apply PhysicsMaterialAPI
physics_mat = UsdPhysics.MaterialAPI.Apply(material_prim)
physics_mat.CreateStaticFrictionAttr().Set(0.5)
physics_mat.CreateDynamicFrictionAttr().Set(0.4)
physics_mat.CreateRestitutionAttr().Set(0.3)
physics_mat.CreateDensityAttr().Set(7850.0) # kg/m^3 for steel
# Bind to collision mesh
binding_api = UsdShade.MaterialBindingAPI.Apply(mesh_prim)
binding_api.Bind(UsdShade.Material(material_prim),
bindingStrength=UsdShade.Tokens.weakerThanDescendants,
materialPurpose="physics")
stage.Save()Asset Creation Workflow
1. Modeling
Requirements:
- Real-world scale (meters)
- Z-up orientation
- Clean topology
- Optimal poly count
- Named hierarchies
- Pivot at origin
- Faces forward in viewport
Tools:
- Maya
- Blender
- 3ds Max
- Houdini2. UV Layout
Standards:
- Non-overlapping islands
- UV Channel 1 only
- 0-1 UV space
- 512+ pixels/meter texel density
- Align to material patterns
- Re-UV CAD models
Tools:
- RizomUV
- Maya UV Toolkit
- Blender UV Editor3. Materials
Workflow: PBR Metal-Rough
Required Maps:
- Base Color (albedo)
- Metallic
- Roughness
- Normal (optional)
- Opacity (if needed)
Resolution: 4K or 8K
Format: PNG, TIFF, or EXR4. Physics Setup
Colliders:
- Multiple components allowed
- No convex hull requirement
- Per-mesh configuration
Mass:
- Set in kilograms
- Real-world accurate
Materials:
- Static friction: 0-1
- Dynamic friction: 0-1
- Restitution: 0-1
- Density: kg/m^35. Semantic Labeling
Required Metadata:
- Class: Human-readable name
- Hierarchy: Taxonomy path
- QCode: WikiData identifier
- Custom: SKU, ProductID
Consistency: Use same taxonomy across datasetIntegration with Isaac Sim
Load SimReady Assets
from omni.isaac.core import World
from omni.isaac.core.utils.stage import add_reference_to_stage
world = World()
# Add SimReady asset
add_reference_to_stage(
usd_path="/Isaac/Props/Pallets/pallet.usd",
prim_path="/World/Pallet"
)
world.reset()Synthetic Data Generation
import omni.replicator.core as rep
# Setup camera
camera = rep.create.camera(position=(5, 5, 5))
# Render with semantic labels
rp = rep.create.render_product(camera, (1024, 1024))
# Annotators
rep.AnnotatorRegistry.get_annotator("rgb")
rep.AnnotatorRegistry.get_annotator("semantic_segmentation")
rep.AnnotatorRegistry.get_annotator("bounding_box_2d_tight")
rep.AnnotatorRegistry.get_annotator("distance_to_camera")
# Randomize assets
with rep.trigger.on_frame():
rep.randomizer.scatter_2d(
surface_prims=["/World/Ground"],
asset_prim="/World/Pallet"
)
rep.orchestrator.run(num_frames=1000)Available Asset Categories
Warehouse/Logistics:
- Pallets (wood, plastic, metal)
- Boxes (cardboard, plastic)
- Conveyors
- Shelving
- Racks
- Forklifts
Manufacturing:
- Machinery
- Tools
- Parts bins
- Work surfaces
Infrastructure:
- Traffic cones
- Barriers
- Signs
- Bollards
Generic Props:
- Containers
- Furniture
- EquipmentValidation
Check SimReady Compliance
from pxr import Usd, UsdPhysics, Semantics
def validate_simready(asset_path):
stage = Usd.Stage.Open(asset_path)
root = stage.GetDefaultPrim()
checks = {
"has_physics": root.HasAPI(UsdPhysics.RigidBodyAPI),
"has_mass": root.HasAPI(UsdPhysics.MassAPI),
"has_semantics": root.HasAPI(Semantics.SemanticsAPI),
"scale_in_meters": True, # Manual check
"clean_hierarchy": True # Manual check
}
# Check for colliders in children
has_collider = False
for prim in Usd.PrimRange(root):
if prim.HasAPI(UsdPhysics.CollisionAPI):
has_collider = True
break
checks["has_collider"] = has_collider
return checks
# Validate
results = validate_simready("asset.usd")
print(results)Best Practices
Modeling
- Real-world scale: Export in meters
- Optimization: Balance detail vs performance
- Naming: Use clear hierarchies
- Pivot points: Position for proper placement
UVs
- Non-overlapping: Required for baking
- Texel density: Maintain consistency
- Channel 1: Don't use additional channels
- Pattern alignment: Before packing
Physics
- Multi-component: Use multiple colliders
- Mass accuracy: Real-world values
- Material tuning: Dial in friction/restitution
- Runtime compatibility: USDPhysics schema
Semantics
- Consistent taxonomy: Across all assets
- WikiData QCodes: For interoperability
- Hierarchical labels: Enable discovery
- Custom IDs: SKUs for tracking
Common Use Cases
Robotics:
- Robot training (manipulation, navigation)
- Sensor simulation (cameras, LIDAR)
- Synthetic data generation
- Sim-to-real transfer
Digital Twins:
- Factory layouts
- Warehouse optimization
- Datacenter planning
- Infrastructure design
Autonomous Vehicles:
- Street scene simulation
- Perception training
- Scenario testing
- Sensor validation
Healthcare:
- Facility planning
- Safety analysis
- Equipment placement
- Workflow optimizationReferences
- Specification - SimReady standard, requirements, scope
- Asset Creation - Modeling, UVs, materials, workflow
- Physics Setup - USDPhysics, colliders, mass, materials
- Semantic Labeling - WikiData, taxonomy, metadata
- USD Structure - OpenUSD, composition, schemas
- Isaac Sim Integration - Loading assets, synthetic data
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
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.

