Show HN: A single command to run Claude Code inside Lima VMs (github.com)

A new tool allows users to run Claude Code, an AI coding agent, within sandboxed Linux virtual machines (VMs) using a single command. This setup enhances security, as the host system remains safe from potential exploits while providing full autonomy to the AI. Currently supporting macOS and Linux, the tool automatically installs Lima, which enables the creation of a Debian 13 VM with essential development tools, Docker, Chromium, and Claude Code. Users can easily clone this template into fresh VMs for isolated AI development, ensuring that any vulnerabilities associated with AI operations are contained away from the main system. This development is significant for the AI/ML community as it establishes a robust environment for experimenting with AI coding agents, minimizing risks linked to running powerful models with extensive permissions. By leveraging VMs for this purpose, developers avoid common pitfalls associated with Docker container limitations, benefiting from full kernel-level isolation and seamless integration of development tools. The automatic port forwarding feature also improves accessibility, making it easier for users to run and test applications without complex setups. This innovation promises to enhance developer productivity while bolstering security for AI implementations.