logoalt Hacker News

valleyeryesterday at 7:23 PM3 repliesview on HN

Not if the harness applies an OS-level sandbox to the process, as Codex does.

https://learn.chatgpt.com/docs/sandboxing?surface=cli


Replies

postalcoderyesterday at 7:49 PM

Unfortunately it was possible in codex too, until recently.

https://github.com/openai/codex/issues/5237

dymkyesterday at 7:39 PM

No, you need to apply the sandbox to the codex process itself. Codex does not do this.

show 1 reply