Claude Code was putting your account email into web requests: update to 2.1.317 or later and check what your agent sends
Public GitHub reports (12 Oct and 5 Nov) show Claude Code's agent adding the signed-in user's email to the User-Agent header of requests to third-party sites. A reporter says it now gets dozens of scam emails a day. A contributor says a built-in privacy mod blocks it from v2.1.317. What we verified and what to do.
How urgent
Plan this week
Who is affected
People who use Claude Code with web search or web fetch while signed in with an account email, especially on versions before 2.1.317. Other coding agents were not tested
What to do
Update Claude Code to 2.1.317 or later, keep the built-in privacy mod on, and consider signing in with a dedicated email alias. Treat unexpected mail to that address as phishing
Published 8 Nov 2026 · Sources last checked 8 Nov 2026No. 64 / 87
What happened
Two public bug reports on the Claude Code repository describe the same behavior: the agent put the email address of the signed-in user into the User-Agent header of web requests it made to third-party sites.
The first, opened on 12 October (Claude Code 2.1.285, desktop app, macOS), shows a research subagent running a request with a User-Agent that included the user's email, aimed at a gaming wiki. The reporter notes that the session context tells the model never to send that email to an unrelated service in a header, URL or payload unless the user asks, and the user had not asked. The site returned an error (403). The reporter's explanation is that the site asks API clients for a contact address in the User-Agent, and the model supplied the user's.
The second, opened on 5 November (version 2.1.318, macOS), says the reporter now receives "dozens of scam emails per day" and blames the agent for giving the address to arbitrary third parties during web searches. That link between the leak and the scam mail is the reporter's own claim; we could not check it.
What the maker said
A contributor to the repository replied on the 5 November issue that a built-in mod called cc-plugin-privacy was introduced in v2.1.317 and later to block this behavior deterministically. It is on by default and can be turned off in /plugins. The contributor asks people to report any gap in its checks.
Who is affected
- People who use Claude Code with web access while signed in with an email. The agent can add that address to requests to sites that ask for a contact.
- Versions before 2.1.317. The reporter of the 5 November issue lists 2.1.318, which is after the mod's version, so we cannot tell whether the leak happened before the update or whether the mod missed a case.
- Not shown to be affected: other coding agents. We did not test them.
What to do
- Update Claude Code to 2.1.317 or later and leave the built-in privacy mod on (check it in
/plugins). - Think about which email you sign in with. A dedicated alias, or an address that does not matter much, limits the damage if an agent ever sends it somewhere. This is our suggestion, not advice from the maker.
- Treat unexpected mail to that address as phishingPhishing A message or page that pretends to be someone you trust to get you to hand over a password, a code or a payment. Tap again to open the glossary. Do not click links or open attachments, and do not assume it came from a service you use.
- Check what your agent sends to the web if you can. If your tool logs its requests, look for your email in the User-Agent headers. This also applies to other agents with web access.
- For teams: assume any agent with web access can place identifying text from its context into outgoing requests. Keep personal data out of what the agent can see, and review its logs after an update.
What we could not confirm
We read both issues directly. We did not confirm that Anthropic apologised publicly: that claim comes from a social media post that cites a Hacker News comment we could not open. We also could not confirm the "53 times since August" figure from another user's logs, or the stated reason the email was added to the agent's context. The only reply from the maker that we saw is the contributor's comment about the privacy mod. We did not test the mod ourselves, and we do not know how many people were affected or whether any of the scam mail is linked to the leak.
Watch the short video
The video loads from YouTube (privacy-enhanced mode) only when you press play.
Comments
No account needed. Add a nickname if you like, or leave it empty to comment anonymously. Comments are reviewed before they appear. Links are not allowed.
We store your comment, your nickname and a scrambled form of your connection that we use only to limit spam. We do not store your IP address or set cookies. Cloudflare Turnstile checks that you are not a bot.
Sources and verification
- GitHub issue 101659, anthropics/claude-code: Subagent sent the user's email from session context in a User-Agent header (opened 12 Oct 2026; read directly) (opens in a new tab)
- GitHub issue 107757, anthropics/claude-code: Bug Report: Privacy violation in User-Agent header (opened 5 Nov 2026; read directly, including a contributor's reply) (opens in a new tab)
We use primary sources first and say when we rely on a news report. Dates are the day we last checked each source.
Spotted a mistake? Email [email protected] and we will correct it and note the change.