ClawKit Logo
ClawKitReliability Toolkit
Back to Registry
Official Verified developer tools Safety 4/5

openclaw-remote

Set up and manage remote OpenClaw installations via SSH/tmux. Walks users through connecting to a remote machine (Tailscale or direct SSH), configuring model providers (z.ai, Anthropic, OpenAI, NVIDIA NIM, OpenRouter), setting primary/fallback models, managing auth, git-tracking config for rollback, hardening, and troubleshooting. Use when user asks to set up, configure, or manage OpenClaw on a remote server, VPS, or Mac mini.

Why use this skill?

Securely manage your remote OpenClaw agents on VPS and servers. Learn how to configure model providers, track git changes, and harden remote installations.

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/seanphan/openclaw-remote
Or

What This Skill Does

The openclaw-remote skill is a specialized agent capability designed for the secure management and orchestration of OpenClaw instances on remote infrastructure. It enforces a security-first architecture by mandating that OpenClaw never runs on a local host machine, instead leveraging remote servers, VPS units, or dedicated hardware like Mac minis via SSH and tmux. This skill provides a structured framework to connect, configure, audit, and troubleshoot remote instances without exposing API credentials or sensitive configurations.

Installation

To integrate this skill into your environment, use the command: clawhub install openclaw/skills/skills/seanphan/openclaw-remote. Once installed, ensure you have your SSH keys properly configured for the remote host. The skill relies on standard system utilities; ensure your remote environment has tmux and python3 installed to allow for the secure manipulation of configuration files via json.dump and session persistence. We recommend using Tailscale as your preferred transport layer for zero-config, encrypted remote access.

Use Cases

  • Remote Infrastructure Management: Deploying and managing OpenClaw agents on cloud VPS providers (AWS, DigitalOcean, Hetzner) without local host contamination.
  • Multi-Model Orchestration: Configuring complex environments where users need to switch between z.ai, Anthropic, OpenAI, NVIDIA NIM, and OpenRouter providers via a unified remote interface.
  • Audit and Rollback: Providing a Git-tracked configuration layer for remote nodes, allowing users to revert to previous working states if a new model or configuration change causes failures.
  • Hardening and Security: Maintaining strict access controls by isolating agent execution environments from daily-driver workstations.

Example Prompts

  • "Set up a new remote OpenClaw instance on my staging VPS at 192.168.1.50 using my existing Tailscale node."
  • "Check the health of my remote agent session; verify if the Anthropic model provider is still authenticated and currently responsive."
  • "I need to rotate my OpenRouter API key on the remote server. Please update the configuration securely and git-track the change."

Tips & Limitations

For advanced users managing large-scale fleets or those who prefer a GUI-driven experience, consider integrating with Claw Desktop. This skill excels at command-line precision but may become cumbersome for users managing dozens of concurrent agents. Always prefer python3 json.dump over shell-based heredocs to ensure JSON schema integrity and prevent injection vulnerabilities. If you encounter connectivity issues, verify the health of your SSH multiplexing before running deep-diagnostic health checks.

Metadata

Author@seanphan
Stars1054
Views0
Updated2026-02-16
View Author Profile
AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill
Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-seanphan-openclaw-remote": {
      "enabled": true,
      "auto_update": true
    }
  }
}

Tags

#openclaw#remote#tmux#ssh#models#configuration
Safety Score: 4/5

Flags: network-access, file-write, file-read, code-execution