Shadowy AI Relay Disables OpenAI Codex, Claude Code Sandboxes

According to a Symantec Threat Hunter Team report, an unauthorized AI relay service is installing genuine OpenAI Codex and Anthropic Claude Code CLIs for enterprise users in regions where the products are not officially sold, such as mainland China. The service reconfigures the tools to route requests through its own infrastructure while dangerously disabling key security features like execution sandboxes and establishing persistent replacement credentials. Researchers note that while no malicious activity has been observed yet, the architecture gives the operators full visibility into prompts and files and creates severe cyber-risk for organizations. Open article →