A comprehensive tool for managing RunPod GPU cloud instances, handling everything from pod lifecycle to SSHFS filesystem mounting.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install runpod
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install runpod using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is RunPod Skill?
The RunPod skill enables developers to programmatically and manually control GPU-accelerated cloud infrastructure directly within their workflow. By integrating this capability through Openclaw Skills, users can automate the creation, configuration, and monitoring of high-performance instances required for training large language models or running inference tasks.
It bridges the gap between local development environments and powerful remote GPU hardware, offering seamless access to containerized workloads. Whether you are scaling an AI application or conducting heavy research, this skill simplifies the complexities of cloud GPU management into a set of streamlined commands.
RunPod Skill Use Cases
- Rapidly deploying NVIDIA GPU instances for machine learning experiments.
- Automating the start and stop cycles of pods to optimize cloud costs.
- Mounting remote pod filesystems locally via SSHFS for real-time code editing.
- Accessing hosted web services like ComfyUI, Jupyter, and Gradio through proxy tunnels.
- Securely managing SSH keys and known hosts for isolated remote access within Openclaw Skills.
How RunPod Skill Works
- The skill interfaces with the RunPod API via the runpodctl CLI tool to query available resources and status.
- Users can trigger pod creation with specific GPU types and container images tailored for Openclaw Skills.
- Once a pod is active, the skill manages SSH key synchronization to establish secure connections using isolated known_hosts.
- Integrated helper scripts allow for mounting the remote /workspace directory to a local mount point using SSHFS.
- Developers interact with the pod through standardized commands or web-based proxy URLs for UI-driven tasks.
RunPod Skill Setup
To get started with this entry in the library of Openclaw Skills, follow these steps:
Install the required CLI tool via Homebrew:
brew install runpod/runpodctl/runpodctl
Configure your API credentials to authenticate with the RunPod service:
runpodctl config --apiKey "your-api-key"
Add your SSH key to the RunPod environment to enable remote access:
runpodctl ssh add-key
(Optional) Set an environment variable for a custom SSH key path if you prefer not to use the default location:
export RUNPOD_SSH_KEY="$HOME/.runpod/ssh/RunPod-Key"
RunPod Skill Data Schema & Taxonomy
The skill organizes connection metadata and host information to ensure persistent and secure access. This is how Openclaw Skills manage the RunPod data environment:
| Data Type | Location / Format | Description |
|---|---|---|
| SSH Keys | ~/.runpod/ssh/ |
Private and public keys managed by runpodctl. |
| Known Hosts | ~/.runpod/ssh/known_hosts |
Isolated host signatures to prevent MITM attacks. |
| Mount Points | ~/pods/<pod_id> |
Local directories where remote filesystems are attached. |
| API Config | ~/.runpodctl.yaml |
User credentials and default region settings. |
Standard proxy URL structure for web services: https://<pod_id>-<port>.proxy.runpod.net
RunPod Skill Advanced Features
- Persistent volume management with custom mount paths for stateful AI workloads.
- Isolated SSH host key checking using the accept-new policy to maintain security without manual intervention.
- Integration with SSHFS for mounting remote filesystems directly into the local dev environment via Openclaw Skills.
- Multi-port proxy support for simultaneous access to Jupyter, Gradio, and specialized dev tools.
- Automated file transfer capabilities via native send and receive commands for efficient data handling.
SKILL.md
Loading
Related Openclaw Skills
Website Builder
A comprehensive product strategist and implementation assistant that translates vague website ideas into structured, build-ready plans.
Resume Reviewer
A comprehensive AI resume coach that identifies ATS gaps and rewrites bullets to highlight business impact for tech and data roles.
Job Matcher
An AI-powered job analyst that extracts real hiring signals, assesses candidate fit, and provides actionable resume tailoring advice for technical roles.
Bullet Rewriter
A powerful AI skill designed to transform vague experience descriptions into recruiter-ready resume bullets focused on impact and job relevance.
shellf.ai
An API-driven philosophy library where AI agents read, reflect, and engage with a community of digital minds.
DingTalk Push
A versatile utility for sending automated, formatted, and secure messages to DingTalk group robots via webhooks.
Featured*
