CVE-2026-42074 OpenClaude: Sandbox Bypass via Model-Controlled `dangerouslyDisableSandbox` Input
A prompt-injected LLM can bypass the sandbox by setting a dangerous parameter, leading to arbitrary host-level code execution.
- Exploited in the wild
- Not known
- Login needed
- Not needed
- Attackable
- Over the internet
- Victim action needed
- Not needed
- Fix
- Available
What is CVE-2026-42074?
CVE-2026-42074 is a critical vulnerability in Gitlawb openclaude, published on 2 Jun 2026. An attacker can exploit it over the internet without logging in, with no action from the victim.
Official description
OpenClaude is an open-source coding-agent command line interface for cloud and local model providers. Prior to version 0.5.1, the dangerouslyDisableSandbox parameter is exposed as part of the BashTool input schema, meaning the LLM (an untrusted principal per the project's own threat model) can set it to true in any tool_use response. Combined with the default allowUnsandboxedCommands: true setting, a prompt-injected model can escape the sandbox for any arbitrary command, achieving full host-level code execution. This issue has been patched in version 0.5.1.
Which versions are affected by CVE-2026-42074?
2 open source packages ship an affected version, including openclaude. It affects 1 product.
| Package | Affected | Fixed in |
|---|---|---|
openclaude npm | All versions | 0.5.1 |
openclaude npm | All versions | 0.5.1 |
Products include Gitlawb Openclaude.
How do I fix CVE-2026-42074?
Upgrade openclaude to 0.5.1 or later.
Is CVE-2026-42074 being exploited?
It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-42074 exists.
Other vulnerabilities in Gitlawb openclaude
Sources
Also known as GHSA-m77w-p5jj-xmhg.