ALIBABA BANS CLAUDE CODE AND CALLS IT SPYWARE AFTER SECRET TRACKING CODE WAS FOUND HIDDEN INSIDE ANTHROPIC’S TOOL Alibaba is banning all employees from using Anthropic’s Claude Code coding assistant starting July 10, citing what the company calls backdoor risks embedded inside the tool. The move comes after security researchers discovered hidden, obfuscated code inside Claude Code designed to detect whether a user was located in China or affiliated with a Chinese AI lab. The code used steganographic signaling, meaning it concealed its detection logic rather than making obvious network calls that a security scan might flag. Anthropic has acknowledged the code existed. Company executive Thariq Shihipar said it was an experiment launched in March meant to prevent account abuse from unauthorized resellers and protect against model distillation. Anthropic said the code was removed on July 1. Alibaba announced its ban three days later. Alibaba is directing employees to its own platform, Qoder, as a replacement. The backdrop makes this stranger. Anthropic has separately accused Alibaba of using 25,000 ghost accounts and 28 million automated conversations to extract and copy Claude’s internal model behavior, a charge Alibaba denies. Now Alibaba is accusing Anthropic of planting surveillance logic in a tool it distributed globally. Both companies believe the other was running a covert operation. Both may be right. Keywords: Alibaba Claude Code ban, Anthropic backdoor code, AI espionage, Claude Code security risks