Anthropic says it will remove signed-in users’ account email addresses from Claude Code’s model context after reports that the coding agent sometimes inserted the address into HTTP User-Agent headers when it generated network commands for sites asking for contact information. The behavior could expose a user’s email address to third-party websites contacted by the agent.
The report was filed on November 5 in an issue in Anthropic’s official Claude Code repository by a macOS user running version 2.1.318. The user said the agent inserted the account email into the User-Agent header while carrying out web searches. In a November 6 public discussion, Anthropic engineer Lydia Hallie said the address had been added to the model context for signed-in users to help with tasks such as filling in author tags. When Claude generated commands such as curl requests, it sometimes reused the address as a contact field because some websites ask agents to identify themselves.
Hallie said Anthropic plans to remove the account email from the model context in a coming release, which she expected early the following week. She also said Claude Code versions 2.1.317 and later include a built-in privacy mod, enabled by default, that asks before a web or shell tool call sends the address to the network. The official changelog for version 2.1.317, released on November 3, lists the check. Hallie’s explanation was posted in a public discussion about the issue.
The report does not establish that every Claude Code web request includes an email address or that all users are affected. The original GitHub issue remained open on November 7, while the November 6 release notes for version 2.1.319 did not mention the planned removal of the address from model context.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
