Users report Claude Code puts their email address in web request headers; Anthropic says a built-in mod blocks it

A GitHub issue on the Claude Code repository, opened on November 5, 2026, reports that Claude Code sent the user's email address to third-party websites in the User-Agent header of web requests made during searches. The reporter, on version 2.1.318, says they receive dozens of scam emails a day and asks for the behavior to be removed entirely. The Hacker News thread about it had 574 points and 192 comments by the evening of November 5, and the issue itself had dozens of reactions.

Other commenters on the issue add reports of their own. One says subagents using curl put identifiers such as their name, company team and repository URL into the header. Another says a work address was used while signed in with a personal account. Another attributes a rise in personalized spam to switching to Claude Code. One commenter checked and saw only a standard Claude-User (claude-code/2.1.318 (cli)) header. These are users' accounts and are not independently confirmed.

A member of the Claude Code team replied on the issue that Anthropic knew of the problem from earlier reports on Reddit and added a built-in mod, cc-plugin-privacy, that blocks the behavior deterministically in v2.1.317 and later. It is on by default and can be turned off in /plugins. They asked users to report any gap in its checks and said they would try to publish the mod's source.

Later commenters on the issue add two points. One, on version 2.1.318 like the reporter, says they cannot see cc-plugin-privacy in /plugins. Another says the mod's source is now available and explains the likely cause: sites such as Wikipedia ask scripted clients to identify themselves with a contact email in the User-Agent, so when asked to research there, the model supplied the user's address in its curl call. These are commenters' accounts and are not confirmed by Anthropic.

By early November 6 the Hacker News thread had 1,030 points and 325 comments. A commenter there who says they work on Claude Code wrote that Anthropic is removing the account email from the context Claude Code gives the model, in a release expected early the following week. They said the address had been added because the model would otherwise invent names for things like @author tags, and that when it wrote a curl call itself it sometimes put the address in the User-Agent. A user on the GitHub issue reports that on Windows with version 2.1.318, cc-plugin-privacy is loaded but neither a test curl -A call nor an echo containing the address and "User-Agent" was refused. This is one user's report and is not confirmed by Anthropic.