akitaonrails/ai-jail
A sandbox wrapper that runs AI coding agents (Claude Code, GPT Codex, OpenCode, Crush) under OS-level access restrictions on Linux and macOS.
Collecting fresh signals — velocity needs a few days of history.
collecting data…
star history
ai-jail wraps AI coding agents in operating-system-level sandboxes (bubblewrap on Linux, sandbox-exec on macOS) to restrict their file-system and process access to user-defined allowed paths. It launches agents as child processes and enforces constraints at the OS level, providing a security layer for untrusted or experimental AI coding sessions without requiring the agent itself to implement restrictions.
Frequently asked
- What is akitaonrails/ai-jail?
- A sandbox wrapper that runs AI coding agents (Claude Code, GPT Codex, OpenCode, Crush) under OS-level access restrictions on Linux and macOS.
- Is ai-jail open source?
- Yes — akitaonrails/ai-jail is open source, released under the GPL-3.0 license.
- What language is ai-jail written in?
- akitaonrails/ai-jail is primarily written in Rust.
- How popular is ai-jail?
- akitaonrails/ai-jail has 1k stars on GitHub.
- Where can I find ai-jail?
- akitaonrails/ai-jail is on GitHub at https://github.com/akitaonrails/ai-jail.