Claude Code was sending users' email addresses in web requests

A user opened an issue on 5 November in the public repository for Anthropic’s Claude Code coding agent, titled “Bug Report: Privacy violation in User-Agent header”. The reporter wrote that they received “dozens of scam emails per day because claude code divulged my email to arbitrary 3rd parties in the User-Agent request header while conducting web searches without my permission”, and asked that the behaviour be removed from the tool rather than made a setting.

The causal claim, that the header was the route by which scammers obtained the address, is the reporter’s own and the thread does not establish it. The thread does show that the account email could appear in the header of outgoing web requests.

A commenter labelled “Contributor” on the repository, writing as “we”, replied the same day. The reply said Claude Code had introduced a built-in mod, cc-plugin-privacy, “to deterministically block this behavior in v2.1.317 and later”, that it is enabled by default and can be switched off through the /plugins command, and that the commenter would try to make the mod’s source available. It asked to be told “ASAP” if the checks had a gap.

By 9 November the thread had a follow-up that did. A user testing version 2.1.319 on macOS reported that the mod did not stop a shell command that carried the account email in the header: the test was a curl request to a closed port on the local machine, which Claude asked permission to run and then ran. The commenter noted it was unclear whether the mod covers only built-in web tools or skips local addresses, and this is a single report that the thread does not show Anthropic confirming or answering. The captured thread was incomplete, with further comments not visible.

The episode is small, but it joins a run of 2026 reports about what coding agents do on users’ machines, among them the tampered extension on a Cursor marketplace the week before.