lima-devbox

🥈Silver

Lima Devbox is a Claude Code skill that automates the setup of Lima VMs as isolated development environments on macOS. It enables safe execution of AI coding agents in a sandboxed VM, preventing system damage or data exposure. Ideal for operations teams running AI agents in YOLO mode.

1600Updated 1mo ago

Intermediate30min to implementautomation

Saves ~45 min per use

Quick InstallView Source

git clone https://github.com/recodelabs/lima-devbox.git

Works with:

Claude

Overview

About This Skill

Lima-devbox is a Claude Code skill that automates the creation and configuration of Lima virtual machines on macOS, providing isolated sandboxed environments for AI coding agents like Claude Code, Gemini CLI, and Codex CLI. It uses Lima (Linux on Mac) to create lightweight Linux VMs with automatic file sharing and port forwarding, ensuring that only explicitly shared directories are accessible to the VM. The skill includes a guided wizard that walks you through VM configuration, mount setup, git configuration, language installation via mise, GitHub CLI setup, and AI agent configuration. Operations teams running AI agents in autonomous mode benefit from this approach by preventing unintended system modifications or data exposure while maintaining near-native performance on Apple Silicon and Intel Macs. After setup, you can run commands directly in the VM using limactl or SSH, with built-in support for container tools like nerdctl.

How to Use

Install lima-devbox via Claude Code plugins (`/plugins` > Add marketplace > recodelabs/lima-devbox) or manually with `claude plugin add https://github.com/recodelabs/lima-devbox`. Invoke it with `/lima-devbox` or prompt Claude to use the skill. The guided wizard will walk you through VM configuration, directory sharing, git setup, language selection, and AI agent installation. After setup, access your VM with `limactl shell dev` or manage it with standard limactl commands.

Use Cases

Running Claude Code with full autonomy in a sandboxed VM without risking host system damage

Safely testing and executing AI coding agents (Gemini CLI, Codex CLI, OpenCode) with controlled filesystem access

Setting up isolated development environments with pre-configured language runtimes (Node.js, Go, Rust, Python)

Teams needing to run multiple AI agents simultaneously with separate, contained workspaces

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/recodelabs/lima-devbox

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Set up a Lima VM for [COMPANY] as an isolated development environment on macOS. The VM should be configured with [SOFTWARE_STACK] and [CONFIGURATION]. Ensure it's ready for running AI coding agents in YOLO mode. Provide step-by-step instructions and any necessary commands.

Example Output

# Lima VM Setup for [COMPANY]

## Configuration Details
- **Software Stack**: Python 3.9, Node.js 16, Docker
- **Configuration**: 4 vCPUs, 8GB RAM, 50GB Disk

## Setup Instructions
1. **Install Lima**: Run `brew install lima` in your terminal.
2. **Create VM**: Execute `lima _create_ [COMPANY]-devbox --cpus 4 --memory 8 --disk 50`
3. **Install Software**: SSH into the VM and run the following commands:
   ```bash
   curl -sSL https://install.python-poetry.org | python3 -
   brew install node@16
   brew install --cask docker
   ```
4. **Verify Setup**: Check the installed versions and configurations.

## Notes
- Ensure Docker is running and accessible within the VM.
- Regularly update the software stack to maintain security.