skills.rest
NVIDIA
NVIDIAOFFICIAL
Explore all skills by this author →
WebsiteGitHub

i4h-workflow-dataset-teleop

Record teleoperation episodes for agentic robotics environments into HDF5 datasets.

3.2k|370|Updated Feb 25, 2026
Search Tags:#simulation#robotics#HDF5#teleoperation#isaac-for-healthcare#imitation learning#demonstration recording
One-click install
npx skills add https://github.com/NVIDIA/skills --skill i4h-workflow-dataset-teleop
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: i4h-workflow-dataset-teleop
Source: https://github.com/NVIDIA/skills/tree/main/skills/i4h-workflow-dataset-teleop
Command: npx skills add https://github.com/NVIDIA/skills --skill i4h-workflow-dataset-teleop

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Recording human teleoperation demonstrations for robotics agentic environments requires coordinating environment configs, device-specific controls, and HDF5 recording paths, and mistakes like relative output paths or unsupported devices silently produce empty or orphaned datasets.

Core Features & Use Cases

  • Teleop Recording Workflow: Launches arena/run.sh --teleop for Isaac for Healthcare agentic environments with keyboard, SO-ARM leader, or VR devices, writing episodes to HDF5.
  • Device Keybinding Surfacing: Extracts the device-specific keybinding table printed at startup from the teleop log so the human operator can actually drive the simulation.
  • Recording Verification: Validates that the HDF5 file exists and that the log reports the requested number of successful episodes before claiming success.
  • Use Case: A robotics engineer asks the agent to record 5 keyboard teleop demos for the scissor pick-and-place task; the skill resolves the i4h-workflows repo, creates a timestamped run directory, launches teleop, surfaces controls, and verifies the saved HDF5.

Quick Start

Ask your agent to record 5 keyboard teleop demos for the scissor pick-and-place task into an HDF5 dataset.

Frequently Asked Questions about i4h-workflow-dataset-teleop

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I record teleoperation demos for Isaac for Healthcare environments?▼

Run workflows/agentic/arena/run.sh with --teleop, an env id, a supported --teleop-device, and an absolute --record-to HDF5 path. Episodes are started with B, marked successful with N, and discarded with R.

Which teleop devices does each i4h agentic environment support?▼

Device support is env-specific: scissor_pick_and_place supports keyboard and so101_leader, while locomanip_tray_pick_and_place and locomanip_push_cart support keyboard_23d. For other envs, check arena/run.sh --env <env> --help.

Why did my teleop recording write the HDF5 file to the wrong location?▼

The recorder resolves relative --record-to paths against the workflows/agentic/arena working directory, creating a nested orphan directory. Always pass an absolute path built from the repository root.

Why does my teleop run report 0/N episodes succeeded?▼

A 0/N result means no human operator completed episodes, so the HDF5 is an empty artifact rather than a usable dataset. Record successful demos with an operator before proceeding to mimic, convert, or finetune steps.

Can I run teleop recording headless without a human operator?▼

No. Teleop is a GUI and hardware handoff; without an operator you may only launch briefly to verify the recorder and controls, then stop cleanly and report that recording is waiting on a human.

Related Skills

Explore More Software Engineering Skills→
tangzheng202202tangzheng202202
1

github-python-install

Clone GitHub Python repositories, validate version requirements, and set up virtual environments.

Community
Intermediate
GonglitianGonglitian
2

data-pipeline-check

Validate dataset formats, schema consistency, and data quality for ML pipelines.

Community
Intermediate
FilenCloudDiensteFilenCloudDienste
34

argent-create-flow

Record and replay MCP tool call sequences as YAML flow files.

Official
Intermediate
slysikslysik

databricks-synthetic-data-generation

Generate realistic synthetic data with Faker and Spark for Databricks Volumes.

Community
Advanced