
Decrypted: The UK Cyber Briefing · September 20 · 3 min
Codex's sandbox escapes: the guard was inside the cell
0:00-3:56
transcript
show notes
Two now-patched sandbox escapes in OpenAI's Codex coding agent show what happens when enforcement sits inside the thing it restrains. What UK teams should do, plus a gang spat between extortion crews and Brussels' child-safety proposal.
Sources
- Escaping the OpenAI Codex sandbox, twice (Accomplish AI)
- Researchers escape OpenAI Codex sandbox to run commands on host (BleepingComputer)
- Prompt injection is not SQL injection (NCSC)
- ShinyHunters hacks Clop leak site, threatens to extort ransomware gang (BleepingComputer)
- European Commission set to push social media restrictions, safety requirements into law (The Record)
- EU Proposes 'Kids Act' Banning Social Media Under 13 (Health Policy Watch)
links7
- Escaping the OpenAI Codex sandbox, twice (Accomplish AI)accomplish.ai
- Researchers escape OpenAI Codex sandbox to run commands on host (BleepingComputer)bleepingcomputer.com
- Prompt injection is not SQL injection (NCSC)ncsc.gov.uk
- ShinyHunters hacks Clop leak site, threatens to extort ransomware gang (BleepingComputer)bleepingcomputer.com





