Quick answers
- Is CVE-2026-35603 exploited in the wild?
- No. Junglewise Threat Intelligence has no report of CVE-2026-35603 being exploited in the wild.
- How severe is CVE-2026-35603?
- High severity, with a CVSS score of 7.3, published 17 April 2026.
- Which products and versions are affected?
- Anthropic Claude Code all versions before 2.1.75; @anthropic-ai/claude-code (npm) < 2.1.75, SEMVER:introduced=0,fixed=2.1.75.
- Is there a fix for CVE-2026-35603?
- Yes. @anthropic-ai/claude-code (npm): upgrade to 2.1.75 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.
- How is CVE-2026-35603 attacked?
- It is attacked with local access, in a software library.
- Am I affected by CVE-2026-35603?
- Check whether you run Anthropic Claude Code, Microsoft Windows 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 2.1.75 or later
npm install @anthropic-ai/claude-code@2.1.75
Technical details
The vulnerability is an insecure search path issue (CWE-426) where Claude Code on Windows loaded system-wide default configuration from C:\ProgramData\ClaudeCode\managed-settings.json without validating directory ownership or file permissions. Because Windows ProgramData is writable by non-administrative users by default and the ClaudeCode subdirectory was not pre-created with restricted access, an unprivileged local user could create the directory and inject a malicious managed-settings.json file. Any user subsequently launching Claude Code on the same system would automatically load and execute this attacker-controlled configuration. The attack requires local system access, low privilege level, and user interaction (victim must launch Claude Code after the malicious file is placed). Impact is high: confidentiality, integrity, and availability of the victim's system could all be compromised. The fix is available in version 2.1.75 and later.
Affected products
- Anthropic @anthropic-ai/claude-code all versions before 2.1.75
Identifiers
- CVE-2026-35603
- GHSA-5cwg-9f6j-9jvx
Timeline
- 2026-04-17: disclosed
- 2026-04-17: patched: Fix available in version 2.1.75