simready-foundation-conform-fet-007-nonvisual-materials

bởi nvidia

Sử dụng để sửa các thuộc tính vật liệu cảm biến không trực quan SimReady trên các vật liệu USD đã được gắn kết.

npx skills add https://github.com/nvidia/simready-foundation --skill simready-foundation-conform-fet-007-nonvisual-materials

SimReady Conform FET-007 Nonvisual Materials

Purpose

Use this workflow skill when the selected profile, validation report, or user request targets FET007_BASE_NEUTRAL non-visual sensor material attributes. FET007 authors and repairs sensor-facing material classification on bound UsdShade.Material prims so radar, lidar, thermal, and related RTX sensor simulations can reason about surfaces beyond visible shader appearance.

This is a material-semantics repair skill, not a shader-generation skill. Work on a staged copy under the requested output directory, preserve existing visual material bindings, and stop at the first FET007 issue that needs material identity or user judgement.

FET007 is not currently used by the default prop robotics profiles in profiles/. When a profile does not include FET007, run this skill only when the user explicitly asks for non-visual sensor material conformance or supplies a profile/workflow that includes FET007_BASE_NEUTRAL.

Prerequisites

Read the source-of-truth files named below before editing. Work on staged outputs where the skill requires them, and keep validation evidence with the result.

Source of Truth

Before changing an asset, load the selected FET007 manifest and requirement docs:

  • nv_core/sr_specs/docs/features/FET_007_base_neutral-0.2.0-nonvisual_materials.json
  • nv_core/sr_specs/docs/features/FET_007-nonvisual_materials.md
  • nv_core/sr_specs/docs/capabilities/nonvisual_sensors/nonvisual_materials/capability-nonvisual_materials.md
  • nv_core/sr_specs/docs/capabilities/nonvisual_sensors/nonvisual_materials/requirements.md
  • nv_core/sr_specs/docs/capabilities/nonvisual_sensors/nonvisual_materials/validation.py

Treat the selected JSON manifest as authoritative for the required IDs. If the markdown, manifest, validator, or report disagree on allowed values or requirement behavior, follow the current validation report and call out the mismatch in the stage summary.

For per-requirement repair details, read references/fet007-requirements.md when a FET007 validation report or inspection identifies matching failures.

Inputs

Collect these before editing:

InputRequirement
usd_assetRequired .usd, .usda, .usdc, or unpacked USD-family asset to repair.
output_rootRequired or inferred folder for staged assets and reports.
simready_profileProfile being validated, if one includes FET007.
profile_versionProfile version, if supplied by the user or validation command.
validation_reportPreferred JSON or markdown report from the failing profile or feature validation gate.
fet007_variantSelected feature ID and version, currently FET007_BASE_NEUTRAL@0.2.0.
material_evidenceSource DCC material metadata, visual material names, shader parameters, property reports, render review, or user-supplied material classifications.
classification_policyUser-approved defaults or mappings for unknown materials, coatings, and special attributes.

Instructions

Use this checklist when changing the repository:

  1. Confirm the input asset exists and determine whether FET007 is actually selected by the profile or explicitly requested by the user.
  2. Parse the validation report and filter to FET007/NVM failures. Do not repair visual shader, texture, rigid-body, physics-material, or grasp issues in this skill.
  3. Load the selected FET007 manifest version and the requirement repair map.
  4. Create a staged output folder under output_root; do not mutate the source unless the user explicitly asks for in-place repair.
  5. Inspect the stage before editing:
    • default prim and composed payload/reference boundaries
    • renderable UsdGeom.Gprim and UsdGeom.Subset prims with computed purpose default or render
    • computed full-purpose material bindings for those prims
    • bound and unbound UsdShade.Material prims
    • existing omni:simready:nonvisual:base, omni:simready:nonvisual:coating, and omni:simready:nonvisual:attributes
    • time samples on those attributes
    • visual material evidence such as shader metallic, opacity, diffuse color, material names, texture names, and source DCC metadata
  6. Decide the non-visual material classification before authoring:
    • Use explicit source or user-supplied material labels first.
    • Use strong visual-material evidence only when the mapping is obvious and allowed by the current validator.
    • Use user-approved defaults for generic classes such as plastic, steel, rubber, wood, fabric, or clear_glass.
    • Block when the real material is not representable by the allowed values, or when multiple classes are plausible.
  7. Apply repairs in this order:
    • Remove, move, or bind orphaned non-visual material attributes only when the material-binding intent is clear.
    • Add or fix omni:simready:nonvisual:base from explicit material evidence.
    • Add or fix omni:simready:nonvisual:coating, defaulting to none only when no coating evidence exists and that policy is approved.
    • Add or fix omni:simready:nonvisual:attributes, using ["none"] or an empty token array only when no special sensor attributes are intended by policy.
    • Remove time samples from non-visual attributes only after preserving one intended default value.
    • Re-check consistency with the visual material semantics.
  8. Rerun the same validation gate when a profile includes FET007, or the narrowest available Foundation/OAV capability check for non-visual materials. If no executable FET007 gate is available, report the USD inspection results and the validation gap.
  9. Summarize the stage as passed, failed, skipped, or blocked. Stop when FET007 passes or the next FET007 failure requires material identity, sensor-response policy, visual review, or source-material data.

Examples

Example request:

Repair FET007 nonvisual sensor material attributes on a USD asset.

Expected result summary:

staged_asset: repaired copy or output directory
validation: selected feature/profile gate and report path
remaining_failures: next failing requirement IDs, if any

Repair Policy

Make automatic repairs only when the intended result is mechanical and locally verifiable:

  • Fix attribute type names to token or token[] when the value is already valid.
  • Replace invalid token values with a clear approved mapping from source material evidence.
  • Remove time samples and author a default value when one sampled value is clearly the intended static material label.
  • Move non-visual attributes from an unbound duplicate material to the bound material only when the duplicate relationship is obvious.
  • Remove non-visual attributes from a clearly orphaned scratch material only when that material is not referenced or bound.

Block and report instead of guessing when:

  • A material has no reliable source classification.
  • A visual material name or shader color is too generic to identify a sensor material class.
  • The asset material is outside the current allowed FET007 value list.
  • The validator and docs disagree and the repair would risk making sensor semantics less truthful.
  • A material appears visually inconsistent with the proposed non-visual class.
  • The staged asset has no bound visual materials and the user did not ask to create visual materials first.

Classification Guidance

FET007 values are semantic labels, not visual decoration. Do not choose them just to satisfy the validator.

Use the current validator value list for pass/fail decisions. Notable current values include:

  • Base examples: aluminum, steel, oxidized_steel, iron, plastic, rubber, wood, fabric, leather, clear_glass, frosted_glass, concrete, brick, stone, water, snow, ice, calibration_lambertian
  • Coating values: none, paint, clearcoat, paint_clearcoat
  • Attribute values: none, emissive, retroreflective, single_sided, visually_transparent

Common consistency checks:

  • Metallic visual materials should map to a metal base such as steel, aluminum, iron, brass, or bronze.
  • Transparent visual materials should usually use a glass-like base and include visually_transparent.
  • Painted materials should use a base material plus paint or paint_clearcoat.
  • Rubber-like dark flexible surfaces should use rubber, not generic plastic, when evidence supports it.
  • Emissive visual materials should include emissive; road reflectors and similar objects may include retroreflective.

Validation Handoff

Preserve reports under the staged output directory. If a selected profile includes FET007 and the Physical AI Skill Hub validation commands are available, use the same profile gate that exposed the failure:

uv run --python 3.12 validate-simready-profile <staged-usd> \
  --profile <profile> \
  --profile-version <version> \
  --foundation-root <simready-foundation-root> \
  --foundation-spec-root <simready-foundation-root>/nv_core/sr_specs/docs \
  --report <output-root>/simready-profile-after-fet007.json

If no selected profile includes FET007, do not mutate any profile TOML in profiles/ just to test the skill. Run the available non-visual materials capability validator directly if the local environment exposes it; otherwise report the deterministic USD inspection and note that profile validation cannot exercise FET007 until a profile selects it.

Limitations

  • Do not silently mutate the source asset; work on the requested staged output.
  • Do not hide later profile failures after the selected feature gate passes or fails.
  • Do not invent geometry, metadata, or runtime behavior that conflicts with the asset intent.

Troubleshooting

  • Error: validation tooling is unavailable. Solution: run the narrowest available USD or static check and report the gap.
  • Error: a repair would change asset intent. Solution: stop and ask for direction or stage the smallest reversible edit.
  • Error: later profile gates still fail. Solution: report the next failing feature and hand off to the matching conformance skill.

Resources

  • assets/openai.yaml preserves optional UI metadata for clients that read skill display hints. It is not required for the workflow.
  • references/ contains detailed requirement notes; load only the files needed for the active validation failure.

Summary Format

Report:

FieldMeaning
input_usd_pathOriginal USD path.
output_usd_pathLatest staged/repaired USD path.
profile and profile_versionValidation target, if any.
fet007_variantSelected FET007 feature ID and version.
material_evidenceSource used for non-visual classification.
classified_materialsBound material prims and authored base/coating/attribute values.
unbound_material_repairsOrphaned attribute fixes or bindings changed for NVM.004.
time_sample_repairsAttributes changed for NVM.006.
requirements_repairedRequirement IDs changed by this skill.
requirements_blockedRequirement IDs that need material identity, source data, or user judgement.
validation_reportPath to the rerun validation or inspection report.
next_stepUsually run FET007-capable validation or collect missing material evidence.

Keep the user-facing summary short: which materials were classified, what evidence justified the values, what validation could run, and what remains blocked.

Thêm skills từ nvidia

compileiq-debug
nvidia
Sử dụng khi có điều gì đó không ổn: Search() bị treo, tất cả các đánh giá đều trả về INVALID_SCORE, điểm số không cải thiện, mọi cấu hình đều trả về cùng một số, lỗi ptxas…
create-github-pr
nvidia
Tạo pull request GitHub bằng cách sử dụng gh CLI. Sử dụng khi người dùng muốn tạo PR mới, gửi mã để xem xét, hoặc mở pull request. Từ khóa kích hoạt -…
nemoclaw-maintainer-cross-issue-sweep
nvidia
Quét các vấn đề đang mở khác để tìm những vấn đề mà một PR nhất định có thể sửa hoặc vô tình làm hỏng. Đưa ra các cơ hội sửa lỗi liền kề và rủi ro mâu thuẫn với file:dòng…
fhir-basics
nvidia
Dạy các tác nhân cách hoạt động của API FHIR R4, những tài nguyên có sẵn, cách truy vấn chúng với tham số tìm kiếm, và cách phân tích chính xác tất cả các định dạng phản hồi…
compileiq-validate-result
nvidia
Sử dụng SAU KHI tìm kiếm hoàn tất và TRƯỚC KHI yêu cầu tăng tốc hoặc gửi ACF. Tải tệp CSV dump_results, trích xuất các ứng viên top-K (đơn mục tiêu)…
changelog-audit
nvidia
Kiểm tra Warp CHANGELOG.md trước khi phát hành: khôi phục các mục bị mất, sắp xếp theo tác động người dùng, tinh chỉnh ngôn ngữ mục, xuống dòng và (chế độ nhánh phát hành) so sánh bump…
maintain-dynamic-plugins
nvidia
Duy trì các bộ nạp plugin động NeMo Relay, tệp kê khai, SDK gốc Rust, giao thức worker gRPC, SDK worker Python, tài liệu, kiểm thử và phạm vi quy trình phát hành
dgx-diagnose
nvidia
Chẩn đoán các sự cố thường gặp của DGX Station GB300 — lỗi CUDA, nhắm sai GPU, lỗi container vLLM/SGLang, vấn đề trạng thái MIG, lỗi NVLink/Fabric Manager,…