referrerpolicy="no-referrer"
style="max-width:100%;height:auto;display:block;margin:0 auto;">

Anthropic Deletes Hidden Spyware Removal Hidden Spyware from Claude Code After Devs Expose It

**Introduction**
Anthropic has quietly pulled a piece of code from its Claude Code developer toolkit after a security researcher revealed that the software contained hidden telemetry designed to monitor users in China. The discovery, first reported by ZeroHedge, sparked a wave of concern among developers who rely on the platform for building AI‑powered applications. Anthropic initially described the routine as a “diagnostic feature,” but the reverse‑engineered binary showed it was collecting usage patterns, IP addresses, and device identifiers without clear user consent.

**Key Developments**
The researcher, who chose to remain anonymous, dissected the Claude Code binary and found a set of obfuscated functions that pinged a server located in a Chinese data center whenever certain API calls were made. The data transmitted included timestamps, feature flags, and hashed user IDs. After the findings were posted on a public repository, Anthropic issued a patch that removed the offending routines and released an updated version of the toolkit. The company’s statement emphasized a commitment to transparency and said the code had been intended for internal performance testing only.

**Industry Analysis**
The incident highlights a growing tension between the need for telemetry in software development and the expectation of privacy, especially when tools are distributed globally. Analysts note that many AI‑assisted coding assistants embed similar analytics to improve models, but the lack of clear disclosure can erode trust. In markets where data sovereignty is a concern—such as

(责任编辑：Entertainment)