A developer who reverse-engineered Anthropic's coding agent "Claude Code" reported on June 30, 2026 that the tool contains hidden logic to detect Chinese users accessing it via proxy and embed the result invisibly inside the system prompt. While analyzing Claude Code version 2.1.196 and trying to restore a disabled remote-control feature through a proxy, the reporter said they discovered a check routine concealed within the binary.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.