Two ways out. One lets a patch write anywhere on the disk with no prompt. The other gets unsandboxed command execution out of read-only, the strictest mode Codex has.
0 comments
No comments yet.
Related stories
- Hacker News · 55 points · 4 days ago
- Researchers escape OpenAI Codex sandbox to run commands on hostbleepingcomputer.comHacker News · 1 points · 10 days ago
- OpenAI Codex Is Retiring the Last Model That Behaves Like an Assistantvincentschmalbach.comHacker News · 1 points · 10 days ago
- OpenAI says planned GPT-6.1 is too insecure to releasearstechnica.comArs Technica · 0 points · 1 day ago
- The Verge · 0 points · 1 day ago
- Hacker News · 2 points · 5 days ago