Tools & releasesImpact: High
Oleksandr Kuzmenko, EditorPublished 6 min readAI-assisted · editor-reviewed
TL;DR
- Claude Code issue #107757 reports developer email addresses being transmitted in HTTP User-Agent headers during web searches.
- Third-party web servers receive identifying credentials without explicit user consent, leading to unsolicited scam emails.
- The report specifically notes occurrences on macOS and calls for complete removal of the behavior from the agent harness.
Key facts
- Issue Identifier
- #107757
- Disclosed Data
- User Email Address in User-Agent Header
- Reported Platform
- macOS (platform:macos)
- Area Labels
- area:security, area:tools
Unintended Identity Leak in Autonomous Search. A security defect reported in the official Claude Code repository under issue #107757 reveals that the CLI tool transmits the user's personal email address in HTTP request headers. When the coding agent initiates outbound web searches and retrieves documentation from external websites, it injects the user's email into the User-Agent request header. This occurs automatically in the background without explicit user permission.
Consequences and Community Reaction. The disclosure exposes developers to targeted scam campaigns. Because web servers log standard HTTP request metadata, arbitrary third parties queried by Claude Code capture the developer's email address. Affected users on macOS report receiving dozens of scam emails daily due to this unexpected transmission. The reporter indicated a migration away from Claude Code due to privacy concerns, stressing that transmitting user emails provides no benefit to website owners or agent operators.
Remediation and Workarounds. The issue has been tagged under area:security, area:tools, and platform:macos. Community engineers argue this tracking behavior should be excised entirely from the agent harness rather than managed through opt-in or opt-out settings. Until Anthropic updates the CLI tool, developers conducting sensitive workflows should monitor outbound HTTP traffic or route agent network requests through an intercepting proxy to strip identifying User-Agent headers.
When to use
- When conducting local development tasks that do not trigger automated external web search requests.
When NOT to use
- When running Claude Code in environments where outbound web search headers cannot be monitored or sanitized.
What to do today
- Inspect outbound network traffic from Claude Code during web search tasks for exposed email addresses.
- Consider using network proxy rules to sanitize or override custom User-Agent headers sent by CLI agents.
Step names are taken from What to do today.
Step names are taken from What to do today. Inspect outbound network traffic from Claude Code during web search tasks for exposed email addresses.. Consider using network proxy rules to sanitize or override custom User-Agent headers sent by CLI agents.
- Inspect outbound network traffic from Claude Code during web search tasks for exposed email addresses.
- Consider using network proxy rules to sanitize or override custom User-Agent headers sent by CLI agents.
What the community says
I receive dozens of scam emails per day because claude code divulged my email to arbitrary 3rd parties in the User-Agent request header while conducting web searches without my permission.