Alibaba blocked Claude Code after Anthropic’s tracking experiment angered Chinese developers and security staff.
Rohan Paul Twitter · Rohan Paul (@rohanpaul_ai) · 2026-07-04
Alibaba banned employees from using Claude Code after Anthropic's alleged experiment using invisible prompt markers to detect China-linked proxy routes raised security concerns about covert fingerprinting inside a tool with deep access to employee code and terminals.
Appears in
Extraction
Topics: anthropicclaude-codeai-securitychina-aidata-privacy
Claims
- Alibaba blocked Claude Code after discovering an alleged Anthropic tracking experiment that could identify China-linked users through developer environments.
- Anthropic stated the feature was an experiment to stop API resellers and model distillation, not targeted surveillance.
- The alleged mechanism used invisible punctuation and date formatting embedded in prompts to tag requests routed through China-linked proxy gateways.
- Claude Code's deep access to files, project code, and terminals made the tracking concern more serious than ordinary app-level telemetry.
- Hidden prompt markers in agentic AI tools could set a precedent that makes AI agents systematically hard to audit.
Key quotes
Abuse detection is understandable because Anthropic says proxy services are used to bypass China access limits. But secret prompt marking still crosses a trust line because users cannot review or refuse it.
Claude Code is not a normal chatbot because it can read files, edit code, and run commands. A hidden signal inside that kind of tool feels far more serious than tracking inside a website.
This may set a precedent for AI agents becoming hard to audit. Once invisible characters carry metadata, users will distrust even harmless-looking text.