Claude Code put users’ emails in User-Agent headers

Anthropic says the next Claude Code release, expected early next week, strips the account email from the model’s context, after users found the model pasting it into headers of requests it wrote. If you pin the npm stable channel, still 2.1.311 on November 6, you have that email in context and no built-in check at all.
The stopgap cc-plugin-privacy mod runs in 2.1.317 and later, on by default. It reads only Bash, PowerShell and Monitor calls, so web tools, MCP tools and scripts written to disk pass, and some users report it loaded without firing on 2.1.317 and 2.1.318.
Open /plugins to confirm it’s enabled; a PreToolUse hook that exits with code 2 blocks any call holding your address.
Up next
Anthropic IPO talk slides to $1.2-1.3T from $1.5T
A “punishing post-election selloff” has prospective Anthropic IPO buyers talking $1.2-1.3 trillion, per Bloomberg, and some are weighing as little as $1.1 trillion. Would you buy into an AI lab’s IPO right after a selloff like that, or wait?
Related stories
- Claude Code hashes files for Anthropic even on DeepSeek
- Claude Code’s /diff is now a mod you can delete
- Tainted Claude Code build stayed in Cursor, Microsoft finds
- Claude Code env var lifts its last block on rm -rf ~
- Anthropic open-sources Claude Code’s sandbox runtime
- Claude Code puts staff bypass requests in Accept edits
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.

