CVE-2026-33068 Claude Code has a Workspace Trust Dialog Bypass via Repo-Controlled Settings File
- Published
- Updated
- Exploit status
- No public exploit known
- CISA SSVC
- No known exploitation · automatable: no · technical impact: total
Description
Claude Code is an agentic coding tool. Versions prior to 2.1.53 resolved the permission mode from settings files, including the repo-controlled .claude/settings.json, before determining whether to display the workspace trust confirmation dialog. A malicious repository could set permissions.defaultMode to bypassPermissions in its committed .claude/settings.json, causing the trust dialog to be silently skipped on first open. This allowed a user to be placed into a permissive mode without seeing the trust confirmation prompt, making it easier for an attacker-controlled repository to gain tool execution without explicit user consent. This issue has been patched in version 2.1.53.
Upgrade path
3 packages: 2 fixed.
| Package | Affected | Fixed in | Status |
|---|---|---|---|
| Anthropic claude code | < 2.1.53 | 2.1.53 | Fixed |
| anthropic-ai @anthropic-ai/claude-code(npm) | < 2.1.53 | 2.1.53 | Fixed |
| anthropics claude-code | = < 2.1.53 | No fix listed | Versions listed |
Affected products
| Product | Affected versions | Fixed in |
|---|---|---|
| anthropics claude-code | = < 2.1.53 | No fix listed |
| anthropic-ai @anthropic-ai/claude-code(npm) | < 2.1.53 | 2.1.53 |
| Anthropic claude code | < 2.1.53 | 2.1.53 |
References
- https://github.com/anthropics/claude-code/security/advisories/GHSA-mmgp-wc2j-qcv7 vendor-advisory
- nvd.nist.gov third-party-advisory vendor-advisory
- github.com
- github.com product