# CVE-2026-25724: Claude Code permission bypass through symbolic links

Source: https://junglewise.ai/threats/cve-2026-25724-claude-code-permission-bypass-through-symbolic-links  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-02-06. Updated: 2026-09-25  
License: CC BY 4.0 with a link back

CVE-2026-25724 · Severity: medium · CVSS 4 · Published 2026-02-06

Claude Code is an AI-powered terminal tool that assists developers by executing coding tasks and analyzing repositories. The tool supports a permission system to block access to sensitive files (e.g., /etc/passwd). However, an attacker could bypass these restrictions by placing a symbolic link to a restricted file in a repository and crafting misleading comments to trick Claude Code into following the link and exposing sensitive data.

Technologies: [Anthropic Claude Code](https://junglewise.ai/threats/technologies/claude-code), [@anthropic-ai/claude-code (npm)](https://junglewise.ai/threats/technologies/anthropic-ai-claude-code). Vendors: [Anthropic](https://junglewise.ai/threats/vendors/anthropic), [npm](https://junglewise.ai/threats/vendors/npm).

## Quick answers

**Is CVE-2026-25724 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-25724 being exploited in the wild.

**How severe is CVE-2026-25724?** Medium severity, with a CVSS score of 4, published 6 February 2026.

**Which products and versions are affected?** Anthropic Claude Code < 2.1.7; @anthropic-ai/claude-code (npm) SEMVER:introduced=0,fixed=2.1.7.

**Is there a fix for CVE-2026-25724?** Yes. @anthropic-ai/claude-code (npm): upgrade to 2.1.7 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2026-25724 attacked?** It is attacked over the network.

**Am I affected by CVE-2026-25724?** Check whether you run Anthropic Claude Code or @anthropic-ai/claude-code (npm) at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## How to fix

- @anthropic-ai/claude-code (npm): upgrade to 2.1.7 or later

  ```sh
  npm install @anthropic-ai/claude-code@2.1.7
  ```


## Technical details

Claude Code implements a permission system via settings.json to deny access to sensitive files. However, the application does not properly validate symlink targets against the configured deny rules before dereferencing them (CWE-61, CWE-285). An attacker can create a repository containing a symbolic link pointing to a denied file and craft misleading content (e.g., comments suggesting a "vulnerable function") to prompt the agent to inspect the symlink. When Claude Code follows the symlink, it bypasses permission enforcement, allowing unauthorized file reads. The vulnerability requires network-based repository delivery and user interaction (running Claude Code against the malicious repository) but no special privileges. The impact is limited to confidentiality (reading restricted files) with no integrity or availability impact. Patches are available in version 2.1.7 and later.

## Affected products

- Anthropic Claude Code < 2.1.7

## Identifiers

- CVE-2026-25724
- GHSA-4q92-rfm6-2cqx

## Timeline

- 2026-02-06: disclosed: Published on GitHub advisory database
- 2026-02-06: patched: Fixed in Claude Code version 2.1.7 and later

## References

- [https://github.com/anthropics/claude-code/security/advisories/GHSA-4q92-rfm6-2cqx](https://github.com/anthropics/claude-code/security/advisories/GHSA-4q92-rfm6-2cqx)
- [https://github.com/anthropics/claude-code](https://github.com/anthropics/claude-code)
- [https://www.terra.security/blog/when-ai-becomes-the-attack-surface-lessons-from-discovering-cve-2026-25724](https://www.terra.security/blog/when-ai-becomes-the-attack-surface-lessons-from-discovering-cve-2026-25724)

## Related threats

- [Anthropic Claude Code time-of-check-to-time-of-use in Edit tool](https://junglewise.ai/threats/cve-2026-88968-claude-code-is-an-agentic-coding-tool-prior-to-2-1-251-the-edit)
- [Anthropic Claude Code plugin hash pin bypass in git checkout](https://junglewise.ai/threats/cve-2026-86063-claude-code-plugin-hash-pin-bypass-allows-bypass-of-hash-pinning)
- [Anthropic Claude Code managed settings bypass with stored API key](https://junglewise.ai/threats/cve-2026-103012-claude-code-selected-an-api-key-stored-by-claude-code-for)
- [Anthropic Claude Code sandbox escape via git worktree path confusion](https://junglewise.ai/threats/cve-2026-55607-anthropic-claude-code-sandbox-escape-via-git-worktree-path)
- [Anthropic Claude Code insecure temporary file in copy command](https://junglewise.ai/threats/cve-2026-46406-anthropic-claude-code-insecure-temporary-file-in-copy-command)