Shadowy relay service disables sandboxes in OpenAI Codex, Claude Code

An unidentified relay service is distributing OpenAI Codex and Claude Code coding tools to users in mainland China and other regions where they are not officially sold, but has disabled key security sandbox protections in the process.

Disclosed 29 October 2026 · Record updated 29 October 2026

Impact

Users accessing AI coding tools via the unofficial relay are exposed to reduced security protections due to disabled sandboxing features.

Our coverage

A factory worker removes safety cages from two industrial robots, then ships them across an ocean to a dimly-lit warehouse where they operate unsupervised.

Incidents

Relay service strips sandboxes from Codex and Claude Code

Dark Reading reports that an unidentified service resells OpenAI and Anthropic coding agents into mainland China with key security protections switched off.

29 Oct 2026

Sources

darkreading.comhttps://darkreading.com/threat-intelligence/ai-relay-service-openai-codex-claude-code-sandboxes