NOFire AI released Brig, an Apache 2.0 tool that runs AI coding agents like Claude Code inside microVMs on Apple Silicon Macs and x86_64/ARM Linux. Each agent gets its own Linux kernel while the shared project stays writable, giving isolation for auto-approved agent runs.
iso (isolate) runs Claude Code and Codex inside disposable Linux VMs on Apple Silicon Macs (macOS 27+), giving agents a full dev environment without touching the host. The tool targets practitioners who want agent isolation with API keys, secrets, and the filesystem kept off the VM.
iso (isolate) runs Claude Code and Codex inside disposable Linux VMs on Apple Silicon Macs (macOS 27+), giving agents a full dev environment without touching the host. The tool targets practitioners who want agent isolation with API keys, secrets, and the filesystem kept off the VM.
Airlock runs AI coding agents like Claude Code and Codex inside a virtual machine on macOS. It records what the agent ran and changed, letting the developer review everything before merging it back into their project. It addresses the trust problem of agents having broad access to local files and networks.
Airlock runs AI coding agents like Claude Code and Codex inside a virtual machine on macOS. It records what the agent ran and changed, letting the developer review everything before merging it back into their project. It addresses the trust problem of agents having broad access to local files and networks.
SideKernel is a MicroVM-based sandbox for macOS that lets developers run AI coding agents like Claude Code in isolation instead of directly on their machine. It targets everyday developers who want agent safety without a second computer or VPS. Built as an MSc capstone project at Georgia Tech.
SideKernel is a MicroVM-based sandbox for macOS that lets developers run AI coding agents like Claude Code in isolation instead of directly on their machine. It targets everyday developers who want agent safety without a second computer or VPS. Built as an MSc capstone project at Georgia Tech.
Agent-run is a tool that lets you run a coding agent (an LLM-powered agent that writes code) inside a sandboxed environment. It solves the problem of safely executing AI-generated code without risking system integrity, targeting developers who want to experiment with or deploy coding agents.
Agent-run is a tool that lets you run a coding agent (an LLM-powered agent that writes code) inside a sandboxed environment. It solves the problem of safely executing AI-generated code without risking system integrity, targeting developers who want to experiment with or deploy coding agents.
Spens provides standardized sandboxing and observability for coding agents, using Docker, nono, and mitmproxy to create reproducible environments that capture all agent traffic. It supports pi, opencode, claude, and codex out of the box, with node and python, and can be configured for any agent or environment. It addresses the problem of agents doing unpredictable or risky actions (like deleting folders or making network requests) without visibility.
Spens provides standardized sandboxing and observability for coding agents, using Docker, nono, and mitmproxy to create reproducible environments that capture all agent traffic. It supports pi, opencode, claude, and codex out of the box, with node and python, and can be configured for any agent or environment. It addresses the problem of agents doing unpredictable or risky actions (like deleting folders or making network requests) without visibility.