Skip to content
Back to skills

Humanoid Gym

ASecurity

"Guide Humanoid-Gym users through PPO training, XBot-L environment

  • 247 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 8, 2026
devopspythonbashbackend

Security analysis

A100/100

Pro scans all 20 files and shows the line behind each finding

Scanned September 8, 2026

npx -y skills add VectorSpaceLab/AREX-Skill --skill humanoid-gym --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Humanoid Gym?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Humanoid Gym
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/vectorspacelab-humanoid-gym/badge)](https://www.skillsdirectory.com/skills/vectorspacelab-humanoid-gym)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: humanoid-gym
description: "Guide Humanoid-Gym users through PPO training, XBot-L environment
  customization, and MuJoCo sim-to-sim deployment."
disable-model-invocation: true
metadata:
  disco-role: operating
license: NO_LICENSE
---

# Humanoid-Gym

Use this repo skill when a future agent needs to work with Humanoid-Gym: training or evaluating the bundled `humanoid_ppo` task, changing the XBot-L environment, or validating/exporting a policy for MuJoCo sim-to-sim.

## Read this first
- `references/repo-provenance.md` if you need to compare the current checkout against the snapshot used to build this skill.
- `references/installation-and-backends.md` for Python, torch, Isaac Gym, MuJoCo, and CUDA readiness.
- `references/troubleshooting.md` when setup, import, backend, or checkpoint selection fails.

## Main routes
- `sub-skills/training-and-evaluation/` — PPO training, evaluation, checkpoint loading, and TorchScript export.
- `sub-skills/environment-customization/` — XBot-L config, rewards, terrain, and new robot/environment setup.
- `sub-skills/sim2sim-deployment/` — MuJoCo deployment of exported policies and asset/policy validation.

## Verified package facts
- Distribution name: `humanoid`
- Version: `1.0.0`
- Registered public task in this checkout: `humanoid_ppo`
- XBot-L policy dimensions: 705 observations, 219 privileged observations, 12 actions
- The public `play.py` export path writes `logs/<experiment_name>/exported/policies/policy_1.pt`

## Minimal install and check
Use a Python 3.8 environment for the repo's documented backend stack.

1. Install this checkout in editable mode from a local clone.
2. Install the runtime stack required by the selected route.
3. Confirm `import humanoid` and `python -m pip check`.
4. If you need train/play execution, confirm Isaac Gym Preview 4 is installed and importable.

Suggested smoke check:

```bash
python -c "import humanoid; print(humanoid.LEGGED_GYM_ROOT_DIR)"
```

If Isaac Gym is missing, stay in static/command-building mode and use the bundled helpers instead of claiming native training or evaluation succeeded.

## Helper script
- `scripts/inspect_humanoid_gym_install.py` — run this when you want a safe root-level install/resource/backend summary without launching Isaac Gym.

## Notes for future agents
- `humanoid/scripts/train.py` and `humanoid/scripts/play.py` are the user-facing PPO entry points, but they require the Isaac Gym backend for real execution.
- `humanoid/scripts/play.py` exports the actor only and hard-codes render/export side effects unless the source file is edited.
- `humanoid/scripts/sim2sim.py` uses MuJoCo viewer rollout and is best validated with the bundled sim2sim sub-skill.

Files in this skill

  • SKILL.md2.7 KB
  • references/installation-and-backends.md2.1 KB
  • references/repo-provenance.md1.4 KB
  • references/repo-routing-metadata.json524 B
  • references/troubleshooting.md2.3 KB
  • scripts/inspect_humanoid_gym_install.py9.2 KB
  • sub-skills/environment-customization/SKILL.md3 KB
  • sub-skills/environment-customization/references/configuration.md6.2 KB
  • sub-skills/environment-customization/references/task-registration.md2.2 KB
  • sub-skills/environment-customization/references/terrain-and-rewards.md5.2 KB
  • sub-skills/environment-customization/references/troubleshooting.md3.4 KB
  • sub-skills/environment-customization/scripts/solve_gait_coefficients.py3.7 KB
  • sub-skills/environment-customization/scripts/summarize_xbot_config.py13.3 KB
  • sub-skills/sim2sim-deployment/SKILL.md1.8 KB
  • sub-skills/sim2sim-deployment/references/policy-and-assets.md3.9 KB
  • sub-skills/sim2sim-deployment/references/troubleshooting.md3 KB
  • sub-skills/sim2sim-deployment/references/workflow.md2.7 KB
  • sub-skills/sim2sim-deployment/scripts/build_sim2sim_command.py1.6 KB
  • sub-skills/sim2sim-deployment/scripts/validate_sim2sim_assets.py17.3 KB
  • sub-skills/training-and-evaluation/SKILL.md2 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.