Claude Code sends your email address to websites

When you are building projects with AI coding assistants, you expect your personal information to stay private. A recent GitHub issue raised concerns among developers using Claude Code, Anthropic's terminal-based assistant, after discovering that web requests sent by the tool carried the user's email address in the network header.

According to the report on GitHub, this setup created an unintended trail where third-party servers could see who initiated the browsing task, leading one user to call it a "privacy violation in User-Agent header" after receiving unwanted spam. For anyone learning to code with AI agents that fetch documentation or browse online resources, this is a good reminder of how automated tools handle network requests under the hood.

As AI agents become more autonomous in browsing the web on your behalf, keeping tabs on privacy settings and tool updates is essential. Anthropic and similar platforms frequently release patches for these issues, so make sure you keep your command-line AI tools updated to the newest versions to stay safe while you build.

Community discussion: https://news.ycombinator.com/item?id=50385272