Does it support specific hardware platforms?

It provides native support for K-Bot humanoid specifications, the 3D-printed Zeroth-bot platform, and Robstride motor controls.

Can I use this for training new locomotion policies?

Yes, the skill includes specialized tools like ksim-rl for PPO/AMP training and entropy-sim2real for robust domain randomization.

What is the K-Scale skill used for?

It is a comprehensive index and toolkit for humanoid robot development, covering everything from simulation and training to hardware deployment and firmware management.

How does it handle communication with physical robots?

The skill integrates with kos-firmware, which provides gRPC services and the KOS PolicyService for low-latency control, calibration, and state monitoring.

K-Scale Robotics Ecosystem

Name: K-Scale Robotics Ecosystem
Author: plurigrid

byplurigrid

•

데이터 과학 및 ML

Provides a unified framework for humanoid robot development, reinforcement learning training, and sim-to-real deployment.

The K-Scale skill is a comprehensive open-source robotics stack designed to bridge the gap between simulation and real-world humanoid deployment. By aggregating specialized tools for reinforcement learning (RL), vision-language-action (VLA) models, and firmware orchestration, this skill enables developers to build, train, and control advanced robotic platforms like the Zeroth-bot. It simplifies the complex workflow of robot development—from URDF-to-MJCF conversion and MuJoCo scene composition to gRPC-based hardware control and entropy-driven domain randomization, making it an essential tool for robotics researchers and engineers using Claude Code.

주요 기능

01Firmware management and gRPC services for real-time robot control and monitoring

02Unified index for humanoid robot specifications, actuators, and VLA models

03Advanced sim-to-real transfer via entropy-driven domain randomization

04Integrated RL training pipelines for locomotion using PPO and AMP algorithms

05Automated URDF to MJCF conversion for rapid MuJoCo simulation setup

062 GitHub stars

사용 사례

01Deploying vision-language-action (VLA) models to real-world edge hardware via KOS firmware

02Composing complex 3D scenes in MuJoCo for multi-robot interaction or obstacle navigation testing

03Training humanoid robots for stable walking and complex locomotion in simulated environments

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add plurigrid/asi kscale

For use in Claude.ai and ChatGPT

Download Skill