rl-isaaclab

Manage end-to-end IsaacLab reinforcement learning workflows on the Euler cluster.

1|Updated Feb 3, 2026

Search Tags:#mlops #simulation #reinforcement learning #robotics #isaaclab #cluster computing

Ready to run

npx skills add https://github.com/Idate96/codex_skills --skill rl-isaaclab

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: rl-isaaclab
Source: https://github.com/Idate96/codex_skills/tree/main/skills/rl-isaaclab
Command: npx skills add https://github.com/Idate96/codex_skills --skill rl-isaaclab

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill streamlines the entire Reinforcement Learning workflow within the IsaacLab environment, from local testing to cluster deployment and result synchronization.

Core Features & Use Cases

Local Smoke Testing: Quickly validate RL scripts on a smaller scale before cluster submission.
Cluster Job Submission & Monitoring: Deploy and manage RL training jobs on the Euler cluster.
Result Synchronization: Consolidate experiment results from the cluster.
Policy Playback: Locally run trained policies for evaluation.
Use Case: Train a robot arm policy for a manipulation task, starting with a quick local test, submitting a large-scale training run to the cluster, monitoring its progress, and finally playing back the learned policy to visualize its performance.

Quick Start

Run a local smoke test for the excavation3d_w_cabin task with 4 environments for 3 iterations.

Frequently Asked Questions about rl-isaaclab

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

FAQPage Schema

How do I manage an end-to-end reinforcement learning workflow for IsaacLab?▼

An end-to-end reinforcement learning workflow for IsaacLab is managed by running local smoke tests, submitting jobs to a cluster, monitoring progress, synchronizing results, and playing back policies. This ensures validation before scaling to large cluster deployments.

What is the best way to validate reinforcement learning scripts before cluster submission?▼

Validating reinforcement learning scripts before cluster submission is best done through local smoke testing. You can quickly run IsaacLab tasks on a smaller scale, such as using 4 environments for 3 iterations, to ensure scripts function correctly before committing cluster resources.

How do I deploy and monitor reinforcement learning training jobs on the Euler cluster?▼

Deploying and monitoring reinforcement learning training jobs on the Euler cluster requires specific shell wrappers and environment variables. Once configured, the workflow handles job submission and tracks training progress directly on the cluster computing environment.

Can I run trained reinforcement learning policies locally for evaluation in IsaacLab?▼

Trained reinforcement learning policies can be run locally for evaluation in IsaacLab using the policy playback feature. This allows you to visualize the performance of a learned policy, such as a robot arm manipulation task, after synchronizing results from the cluster.

Does the IsaacLab reinforcement learning workflow require specific environment variables for execution?▼

The IsaacLab reinforcement learning workflow does require specific shell wrappers and environment variables for execution on the Euler cluster. These configurations are necessary to properly manage cluster job submission, result synchronization, and debugging operations.