CVE-2026-42074: OpenClaude Sandbox Bypass via Model-Controlled `dangerouslyDisableSandbox` Input

Should you act?

Critical and remotely reachable

No known exploitation yet, but it is network-reachable without credentials. This signal is computed from CISA KEV, EPSS and CVSS data; it is not an editorial verdict.

• Not in CISA KEV, but CISA’s SSVC assessment records a public proof-of-concept, so attack code exists.
• EPSS 0.7%, percentile 53: FIRST’s model estimates a 0.7% chance of exploitation activity in the next 30 days, higher than 53% of all scored CVEs (moderate). EPSS is a probability, not a verdict.
• Remotely reachable with no login and no user interaction. If the vulnerable service is exposed to the internet, prioritize it.
• Only applies if your dependency tree includes openclaude (npm) < 0.5.1, fixed in 0.5.1. Check your lockfile, including transitive dependencies.

What it is

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 tooluse 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.

From the GitHub advisory

Summary 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 tooluse 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.

Details The vulnerability exists in the shouldUseSandbox() function in src/tools/BashTool/shouldUseSandbox.ts (lines 130–153):

The second guard, areUnsandboxedCommandsAllowed(), is defined in src/utils/sandbox/sandbox-adapter.ts (lines 474–477) and defaults to true:

This means that out of the box, with default configuration, the LLM can bypass the sandbox for any command simply by including "dangerouslyDisableSandbox": true in its tool call input. The project's own trust model states: "The model/agent is not a trusted principal. Assume prompt/content injection can manipulate behavior." Yet this model-controlled input directly controls a security-critical boundary.

PoC 1. Set up an environment with sandboxing enabled (default for certain deployment modes). 2. A prompt injection payload causes the LLM to return a tooluse block:

Weakness type: CWE-306, CWE-284.

Product: openclaude (npm) · Assigned by: GitHub_M · Published: 2026-06-02

Affected and fixed versions

Severity scores

• Attack vector: network (remotely reachable)
• Attack complexity: low
• Attack requirements: none
• Privileges required: none (no login)
• User interaction: none

CVSS measures how bad a flaw is if exploited, not how likely exploitation is. Read it next to the exploitation data below.

Exploitation and urgency

• CISA KEV: not listed (no confirmed in-the-wild exploitation known to CISA).
• CISA SSVC assessment: exploitation poc, automatable yes, technical impact total.
• FIRST EPSS (model of 2026-10-21): 0.72% probability of exploitation activity in the next 30 days; percentile 53.

Timeline

• 2026-05-12 — GitHub-reviewed advisory GHSA-m77w-p5jj-xmhg published (critical)
• 2026-06-02 — CVE record last updated
• 2026-06-02 — CVE record published by GitHub_M

Sources and references

Original sources first. Version and fix details should always be confirmed against the vendor or project advisory.

Vulnerability records

1. GitHub Advisory GHSA-m77w-p5jj-xmhg github.com
2. CVE.org record CVE-2026-42074 cve.org
3. NVD entry CVE-2026-42074 nvd.nist.gov

Other references from the CVE record and advisory

1. github.com github.com
2. github.com github.com
3. github.com github.com

Critical npm advisories · Today’s Security Digest