Sign in
Deutsch English Español Polski
Medium

@anthropic-ai/claude-code has an Insecure Temporary File in /copy Command that Enables Response Disclosure and Symlink-Based File Write

  • Disclosed: 25.06.2026
  • Updated: 11.10.2026 10:22

The Claude Code /copy command wrote responses to a hardcoded, predictable path (/tmp/claude/response.md) without UID isolation, randomness, or symlink protection. The file was created world-readable (0644) in a world-traversable directory (0755), allowing any local user to read a privileged user's Claude response, which could contain secrets or credentials. Additionally, because the path was static and predictable, a local attacker could pre-create the directory and plant a symlink at the expected file path, causing the privileged process to follow the symlink and overwrite an attacker-chosen file with the response text. Exploiting this required a local unprivileged user on the same system and a privileged user to run the /copy command.

Users on standard Claude Code auto-update have received this fix already. Users performing manual updates are advised to update to the latest version.

Claude Code thanks hackerone.com/c_h4ck_0 for reporting this issue.

Affected software

Click a technology to see all vulnerabilities affecting it.

PlatformPackageVulnerable versionsFixed in
npm package @anthropic-ai/claude-code >= 2.1.59, < 2.1.128 2.1.128

References

CVSS score -
CVSS vector
CVSS:4.0/AV:L/AC:L/AT:P/PR:L/UI:A/VC:L/VI:L/VA:H/SC:N/SI:N/SA:N
Weakness (CWE)
CWE-200, CWE-377, CWE-59
Platform
npm
Package
@anthropic-ai/claude-code
Disclosed
25.06.2026
Updated
11.10.2026 10:22

Sources and licenses

OSV / GitHub Advisory Database - GHSA-4vp2-6q8c-pvq2 - CC BY 4.0

Vulnerability history: @anthropic-ai/claude-code

All published vulnerabilities of this package (npm package), newest first. Shows whether this is a one-off slip or a package that needs watching.

All (29) in the archive
Total entries
29
First → latest
06.2025 10.2026
Without a fix
0
Exploited in attacks
0

Severity breakdown

  • High 23
  • Medium 4
  • Low 2

Disclosures per year

Severity:
  1. Claude Code: Plugin Hash Pin Bypass Allows Bypass of Hash Pinning
    High CVE-2026-86063 < 2.1.179 2.1.179
  2. Claude Code: Sandbox Escape via Git Worktree Path Confusion Allows Unsandboxed Code Execution
    High CVE-2026-55607 >= 2.1.38, < 2.1.163 2.1.163
  3. @anthropic-ai/claude-code has an Insecure Temporary File in /copy Command that Enables Response Disclosure and Symlink-Based File Write this entry
    Medium CVE-2026-46406 >= 2.1.59, < 2.1.128 2.1.128
  4. Claude Code: Trust Dialog Bypass via Git Worktree Spoofing Allows Arbitrary Code Execution
    High CVE-2026-40068 >= 2.1.63, < 2.1.84 2.1.84
  5. Claude Code: Sandbox Escape via Symlink Following Allows Arbitrary File Write Outside Workspace
    High · 10.0 CVE-2026-39861 < 2.1.64 2.1.64
  6. Claude Code has a Workspace Trust Dialog Bypass via Repo-Controlled Settings File
    High CVE-2026-33068 < 2.1.53 2.1.53
  7. Claude Code has Sandbox Escape via Persistent Configuration Injection in settings.json
    High CVE-2026-25725 < 2.1.2 2.1.2
  8. Claude Code Vulnerable to Command Injection via Piped sed Command Bypasses File Write Restrictions
    High CVE-2026-25723 < 2.0.55 2.0.55
  9. Claude Code Vulnerable to Command Injection via Directory Change Bypasses Write Protection
    High CVE-2026-25722 < 2.0.57 2.0.57
  10. Claude Code has a Command Injection in find Command Bypasses User Approval Prompt
    High CVE-2026-24887 < 2.0.72 2.0.72
  11. Claude Code has a Path Restriction Bypass via ZSH Clobber which Allows Arbitrary File Writes
    High CVE-2026-24053 < 2.0.74 2.0.74
  12. Claude Code has a Domain Validation Bypass which Allows Automatic Requests to Attacker-Controlled Domains
    High CVE-2026-24052 < 1.0.111 1.0.111
  13. Claude Code Command Validation Bypass Allows Arbitrary Code Execution
    High CVE-2025-66032 < 1.0.93 1.0.93
  14. @anthropic-ai/claude-code has Sed Command Validation Bypass that Allows Arbitrary File Writes
    High CVE-2025-64755 < 2.0.31 2.0.31
  15. Claude Code vulnerable to command execution prior to startup trust dialog
    High CVE-2025-65099 < 1.0.39 1.0.39
  16. Claude Code can execute commands prior to the startup trust dialog
    High CVE-2025-59536 < 1.0.111 1.0.111
  17. Claude Code Vulnerable to Arbitrary Code Execution via Plugin Autoloading with Specific Yarn Versions
    High CVE-2025-59828 < 1.0.39 1.0.39
  18. Claude Code vulnerable to arbitrary code execution caused by maliciously configured git email
    High CVE-2025-59041 < 1.0.105 1.0.105
  19. Claude Code rg vulnerability does not protect against approval prompt bypass
    High CVE-2025-58764 < 1.0.105 1.0.105
  20. Claude Code Vulnerable to Arbitrary Code Execution Due to Insufficient Startup Warning
    High ghsa-ph6w-f82w-28w6 < 1.0.87 1.0.87
Show more (4) Showing 20 of 24

Know which vulnerabilities affect your application

With the DockRay integration and software inventory enabled, DockRay checks reported component versions against known vulnerabilities and notifies you about confirmed matches. Coverage depends on the supported technology and available advisory data.

Connect your application
DockRay newsletter

Application security in your inbox

Important vulnerabilities, actionable guidance and practical advice for your applications. Choose your technologies and follow what matters to your project.

  • A daily digest of new vulnerabilities, critical ones included.
  • Notifications about important updates.
  • Practical advice on application security and monitoring.

Leave empty to receive updates for every technology.

Privacy policy

Free to subscribe. Unsubscribe with one click in any email.