AGENT SKILL

asset-import-export

Use Newton URDF, MJCF, USD, schema resolver, mesh, remesh, heightfield, and asset import/export workflows safely.

License 许可
Apache-2.0
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
asset-import-export
license
Apache 2.0
metadata
{ "disco-role": "operating" }
description
Use Newton URDF, MJCF, USD, schema resolver, mesh, remesh, heightfield, and asset import/export workflows safely.
disable-model-invocation
true

02 资源文件 4

03 SKILL.md 正文

Newton asset import and export

Use this sub-skill when a task involves loading assets into Newton, converting URDF/MJCF/USD models, diagnosing optional importer dependencies, handling mesh quality/remeshing, using schema resolvers, or preparing scene artifacts for USD/File viewers.

Route here for

  • ModelBuilder.add_urdf(), add_mjcf(), and add_usd() arguments and workflow choices.
  • Optional extras for importers, MuJoCo/MJCF, USD schemas, mesh processing, and remeshing.
  • USD schema resolver behavior, custom attributes, deformable imports, and returned deformable maps.
  • Visual-vs-collider parsing, fixed-joint collapse, self-collision flags, asset path resolvers, and mesh hull limits.
  • newton.utils mesh validation, solidification, remeshing, heightfield, and texture utilities.
  • Export-style artifacts through ViewerUSD or file-based viewer routes, before switching to sensor/viewer details.

Route elsewhere

  • Building primitive scenes from scratch: use ../modeling-simulation/SKILL.md.
  • Choosing a solver/contact path for an imported model: use ../solvers-contacts/SKILL.md after import succeeds.
  • Robot controllers, IK, target arrays, and policies: use ../robotics-control/SKILL.md.
  • Live viewers, example CLI, screenshots, Rerun/Viser/RTX details: use ../sensors-visualization/SKILL.md.

Read order

  1. references/import-export-workflows.md for URDF/MJCF/USD decision paths and public signatures.
  2. references/mesh-and-usd-reference.md for mesh utilities, USD schema resolver notes, and deformable limitations.
  3. references/troubleshooting.md for optional dependency, path, mesh, and parser failures.
  4. scripts/check_import_extras.py to check importer/remesh optional modules without loading assets.

Dependency rule

Base pip install newton is enough for primitive model construction and many public APIs. Asset workflows often need extras:

  • newton[sim] for MuJoCo/MJCF and SolverMuJoCo support.
  • newton[importers] for USD, mesh processing, URI resolution, and schema packages.
  • newton[remesh] for remeshing utilities.
  • newton[examples] when running built-in examples that combine importers, viewers, and policy dependencies.

Install the smallest extra that matches the requested format; do not install all extras just to inspect a simple primitive scene.

Safe diagnostic

From this sub-skill directory:

python scripts/check_import_extras.py

The script reports optional modules and the likely Newton extra that provides each one. It performs no downloads and does not open asset files.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-triton-opt

Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-solve

END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-skill

PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-pattern-fusion

Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-orchestrate

Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-feedback

Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).

★ 351