IS

isaac-sim-asset-healer

Creates safe Python scripts for Isaac Sim scene building.

Install

mkdir -p .claude/skills/isaac-sim-asset-healer && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/11141" && unzip -o skill.zip -d .claude/skills/isaac-sim-asset-healer && rm skill.zip

Installs to .claude/skills/isaac-sim-asset-healer

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Use when creating, reviewing, or repairing Isaac Sim 5.1 scene-generation Python for GUI Script Editor workflows, especially from floor-plan images or local USD/USDC asset inventories. The skill forbids direct MCP script execution and produces English-only code snippets that clean old prims, validate assets, preserve navigation clearance, and apply collision APIs.
366 chars✓ has a “when” triggerlonger than Claude Code's old 250-char listing cap (fine on current versions)
Advanced

Key capabilities

  • Generate Isaac Sim 5.1 Python scene snippets
  • Validate 3D asset geometry and navigation clearance
  • Apply UsdPhysics.CollisionAPI to mesh assets
  • Clean target root prims before rebuilding
  • Create standalone SimulationApp runners

How it works

The skill generates Python code for Isaac Sim 5.1 that constructs scenes using data tables for walls, doors, and slabs. It validates geometry and navigation clearance before providing snippets for the GUI Script Editor or standalone execution.

Inputs & outputs

You give it
Floor-plan image or coordinate data
You get back
Python scene-generation code and validation report

When to use isaac-sim-asset-healer

  • Building simulation scenes
  • Repairing Isaac Sim scripts
  • Validating 3D asset integration

About this skill

Isaac Sim Asset Healer

Purpose

Generate reliable Isaac Sim 5.1 Python for two safe offline execution modes:

  • GUI Script Editor snippets that the user can paste into Isaac Sim.
  • Standalone SimulationApp runners executed with the Isaac Sim Python interpreter.

The generated code must build or repair a scene without relying on MCP execute_script transport.

Required Inputs

  • Floor-plan image, coordinate data, or a precise verbal layout.
  • Scale source, such as a dimension label, grid size, or user-confirmed meters-per-cell value.
  • Local asset directory when USD/USDC assets are requested.

If the request lacks both visual layout and coordinate data, ask for a top-down screenshot or a dimensioned sketch before generating code.

Hard Rules

  • Do not call Isaac Sim MCP execute_script to create the scene.
  • Generated Python code must use ASCII-only comments, strings, variable names, and print output.
  • Prefer standalone SimulationApp({"headless": False}) when the user wants Isaac Sim to open and build the scene automatically.
  • Prefer Script Editor snippets when Isaac Sim is already open and the user wants to paste code manually.
  • Never invent local asset filenames.
  • Check real local inventory before referencing USD or USDC files.
  • Use bundle-safe paths for packaged assets: os.path.join(asset_root, folder_name, folder_name).
  • Delete the target root prim before rebuilding it.
  • Keep (0, 0, 0) clear for robot spawning with at least 1.5m radius.
  • Keep doorways and node transitions clear with at least 1.5m radius.
  • Keep furniture separated enough for robot navigation.
  • Keep furniture inside wall bounds.
  • Build outer walls from the actual footprint, not a simplified rectangle.
  • Build doorway walls with exactly two full-height wall pieces and no header.
  • Use data tables as source of truth for walls, doors, slabs, windows, and placements.
  • Apply UsdPhysics.CollisionAPI to generated obstacles and referenced mesh assets.

Workflow

  1. Calibrate scale.
  2. Extract topology as nodes, edges, corners, walls, doors, floor slabs, windows, and optional furniture zones.
  3. Validate geometry: axis alignment, duplicate walls, door gaps, floor continuity, and navigation clearance.
  4. Generate a Script Editor snippet with self-contained data tables and helper functions.
  5. Include a concise validation report before or after the code.

For standalone mode, wrap the same generated scene builder with:

  • from isaacsim import SimulationApp
  • simulation_app = SimulationApp({"headless": False})
  • normal Isaac Sim imports after SimulationApp starts
  • an update loop using while simulation_app.is_running(): simulation_app.update()

Output Shape

When generating a scene, return:

  • Corner Points Table
  • Wall Segments Table
  • Doorway Table
  • Floor Slabs Table
  • Script Editor Python snippet
  • Optional standalone SimulationApp runner
  • Validation notes

Implementation Preference

Use UsdGeom.Cube with translate and scale for primitive geometry. This is more stable in Isaac Sim 5.1 than FixedCuboid in environments where type conversion errors occur.

Use FixedCuboid(..., color=np.array(...)) only when the target environment is known to support it.

Run standalone scripts with the Isaac Sim Python executable, for example:

/home/iy/miniconda3/envs/isaacsim-5.1/bin/python /tmp/generated_scene.py

When not to use it

  • When the request lacks both visual layout and coordinate data
  • When direct MCP script execution is required

Limitations

  • Requires top-down screenshot or dimensioned sketch if layout data is missing
  • Must use ASCII-only comments and strings

How it compares

It provides structured, validated Python code for offline execution instead of relying on direct MCP script transport.

Compared to similar skills

isaac-sim-asset-healer side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
isaac-sim-asset-healer (this skill)03moReviewAdvanced
dev26moReviewAdvanced
examples-auto-run23moReviewIntermediate
lora-manager-e2e16moReviewAdvanced

Try saying

Example prompts that trigger this skill in your AI assistant.

You might also like

dev

atopile

LLM-focused workflow for working in this repo: compile Zig, run the orchestrated test runner, consume test-report.json/html artifacts, and discover/debug ConfigFlags.

28

examples-auto-run

openai

Run python examples in auto mode with logging, rerun helpers, and background control.

23

lora-manager-e2e

willmiao

End-to-end testing and validation for LoRa Manager features. Use when performing automated E2E validation of LoRa Manager standalone mode, including starting/restarting the server, using Chrome DevTools MCP to interact with the web UI at http://127.0.0.1:8188/loras, and verifying frontend-to-backend functionality. Covers workflow validation, UI interaction testing, and integration testing between the standalone Python backend and the browser frontend.

13

migrate

alirezarezvani

Migrate from Cypress or Selenium to Playwright. Use when user mentions "cypress", "selenium", "migrate tests", "convert tests", "switch to playwright", "move from cypress", or "replace selenium".

13

mobile-use-setup

minitap-ai

Interactive setup wizard for Minitap mobile-use SDK. USE WHEN user wants to set up mobile automation, configure mobile-use SDK, connect iOS or Android devices, or create a new mobile testing project.

12

create-environments

PrimeIntellect-ai

Create or migrate verifiers environments for the Prime Lab ecosystem. Use when asked to build a new environment from scratch, port an eval or benchmark from papers or other libraries, start from an environment on the Hub, or convert existing tasks into a package that exposes load_environment and installs cleanly with prime env install.

11

Search skills

Search the agent skills registry