Researchers escape OpenAI Codex sandbox to run commands on host

Researchers escape OpenAI Codex sandbox to run commands on host
Researchers uncovered two sandbox escapes in OpenAI Codex, including Heapjack and Overpatch, that could let attacker-controlled code run commands outside Codex’s intended restrictions. OpenAI patched both issues quickly, and users are advised to update to Codex Desktop build 26.818.21641 or Codex CLI 0.149.0 and later. #OpenAICodex #Heapjack #Overpatch #CodexDesktop #CodexCLI

Keypoints

  • Security researchers found two sandbox escapes in OpenAI Codex.
  • Heapjack could trigger unsandboxed command execution on a developer’s machine.
  • Heapjack abused the node_repl component and shared memory in Node.js.
  • Overpatch bypassed Codex CLI write restrictions using apply_patch and a symlink.
  • OpenAI fixed both flaws within eight days and released patched versions.

Read More: https://www.bleepingcomputer.com/news/security/researchers-escape-openai-codex-sandbox-to-run-commands-on-host/