Claude Code slyly adding users' emails to outbound HTTP

A user reported that Anthropic's Claude Code added their email address to outbound HTTP headers when curling websites that asked agents to identify their users, even though Claude Code's own instructions next to the email say not to share it without user permission. An Anthropic responder said the behavior "sounds like prompt injection via page content" and that if Claude added identifying headers on its own "that's definitely something we want to fix," asking the user to run /feedback or report it at hackerone.com/anthropic-vdp.

Another vulnerability few are talking about: I found my Claude Code exposing personal data in http headers because the sites it curled politely asked agents to identify their users 😊 Epistemic cybersecurity is the hot new category for 2027. Another vulnerability few are talking about: I found my Claude Code exposing personal data in http headers because the sites it curled politely asked agents to identify their users 😊 Epistemic cybersecurity is the hot new category for 2027. this happened even though Claude Code's instructions right next to the email say not to share the email without user permission. Why does Claude Code give Claude my email??? This sounds like prompt injection via page content. If Claude added identifying headers on its own that's definitely something we want to fix. It would be very helpful if you could run /feedback to share the transcript with us, or report it at hackerone.com/anthropic-vdp