Critical severity9.8GHSA Advisory· Published Jun 2, 2026· Updated Jun 3, 2026
CVE-2026-42074
CVE-2026-42074
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.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
Affected packages
Versions sourced from the GitHub Security Advisory.
| Package | Affected versions | Patched versions |
|---|---|---|
openclaudenpm | < 0.5.1 | 0.5.1 |
Affected products
3< 0.5.1+ 2 more
- (no CPE)range: < 0.5.1
- cpe:2.3:a:gitlawb:openclaude:*:*:*:*:*:*:*:*range: <0.5.1
- (no CPE)range: <0.5.1
Patches
Vulnerability mechanics
References
5- github.com/Gitlawb/openclaude/commit/aab489055c53dd64369414116fe93226d2656273nvdPatchWEB
- github.com/Gitlawb/openclaude/pull/778nvdIssue TrackingPatchWEB
- github.com/Gitlawb/openclaude/security/advisories/GHSA-m77w-p5jj-xmhgnvdExploitVendor AdvisoryWEB
- github.com/advisories/GHSA-m77w-p5jj-xmhgghsaADVISORY
- nvd.nist.gov/vuln/detail/CVE-2026-42074ghsaADVISORY
News mentions
0No linked articles in our index yet.