Tuesday, November 10, 2026 · Claude & Anthropic
Claude · Safety · happened Nov 8, 2026
Claude Code put users' account emails into web requests to outside sites; Anthropic says it will remove the email from the model's contextReported
Developers reported on GitHub that Claude Code sometimes placed their signed-in account email in the User-Agent header of curl and other web requests, including calls to US government sites, because the email had been added to the model's context to fill in author fields. Lydia Hallie of the Claude Code team acknowledged the behavior on Hacker News, apologized, and said the next release will drop the email from the context; version 2.1.317 already asks before a command that contains it goes out, but she said that check misses scripts Claude writes and runs itself. Anthropic hasn't published a formal advisory, and the 2.1.322 release notes on Nov. 10 don't mention the change.
Why it matters
Personal email addresses may have reached third-party server logs without users knowing, and the stopgap check doesn't cover every path.
Try it
Update Claude Code to 2.1.317 or later (the stable channel was still on 2.1.311) and keep the built-in privacy check on until the context change ships.