CVE-2026-47213: Boxlite Timeout Bypass Vulnerability – SIGALRM Signal Handling Flaw
Boxlite is a service that runs untrusted code safely by executing it inside lightweight virtual machines (called Boxes) with OCI containers. To prevent code from consuming resources indefinitely, administrators can set timeouts. When the timeout expires, Boxlite should forcefully terminate the process. However, in versions 0.8.2 and earlier, Boxlite uses a catchable signal (SIGALRM) instead of an uncatchable one (SIGKILL). This means malicious code can ignore the shutdown command and keep running, exhausting the virtual machine's resources and potentially disrupting the service for all users.
Source data · NVD / CISA · public domain
- CVSS
- 3.1 · 6.5 MEDIUM · CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
- Weaknesses (CWE)
- CWE-404
- Affected products
- 0 configuration(s)
- Published / Modified
- 2026-06-10 / 2026-06-17
NVD description (verbatim)
Boxlite is a sandbox service that allows users to create lightweight virtual machines (Boxes) and launch OCI containers within them to run untrusted code. In versions 0.8.2 and prior, Boxlite allows users to configure a timeout for services running inside the virtual machine. When the timeout is triggered, Boxlite sends a signal to kill the process. However, instead of using the uncatchable SIGKILL signal, Boxlite uses the catchable SIGALRM signal. Malicious code running inside the sandbox can exploit this vulnerability to continue running after the timeout is triggered, leading to resource exhaustion within the virtual machine and affecting the availability of the Boxlite service. This issue has been patched via commit 28159fc.
3 reference(s) · View on NVD →
SEC.co analysis · AI-assisted, reviewed against source
Technical summary
The vulnerability stems from improper use of signal handling for process termination in Boxlite's timeout mechanism. When a configured timeout is reached, the service attempts to terminate the target process using SIGALRM, a signal that can be caught and ignored by application code. An attacker executing code within the sandbox can register a signal handler for SIGALRM, allowing the process to continue execution despite the timeout. This bypasses the intended resource containment strategy, enabling denial-of-service through resource exhaustion. The issue is classified under CWE-404 (Improper Resource Validation). A fix is available via commit 28159fc, which presumably corrects the termination mechanism to use SIGKILL or an equivalent uncatchable signal.
Business impact
Organizations relying on Boxlite to safely execute untrusted code face availability risk. A malicious actor or compromised application running inside a Box can circumvent timeout controls, consuming CPU, memory, and I/O resources. This starves the host virtual machine and can degrade or crash the Boxlite service, affecting all users and workloads dependent on it. In multi-tenant environments, one user's malicious code could impact others' jobs. The financial and operational impact depends on how critical Boxlite is to the organization's code execution and testing workflows.
Affected systems
Boxlite versions 0.8.2 and earlier are affected. Organizations should inventory their Boxlite deployments and check version numbers to identify exposure. Verify against the vendor's release notes and repository to confirm which versions include the patched commit 28159fc and are therefore safe to deploy.
Exploitability
Exploitation requires network access and authenticated credentials (the CVSS vector indicates PR:L, meaning Low Privilege is needed), plus the ability to submit and execute code within a Boxlite container. The attack is straightforward: code running inside the sandbox simply registers a handler for SIGALRM and ignores the signal when received. No exotic techniques or public exploit code is required. The barrier to exploitation is low for any user with permission to launch containers, making this a practical risk in multi-tenant or untrusted-code execution scenarios.
Remediation
Update Boxlite to a version that includes the fix from commit 28159fc or later. Verify the patched version with the vendor's release notes. Until patching is complete, consider restricting who can launch containers and monitor for suspicious resource consumption patterns (processes that ignore timeout signals or consume resources continuously). Implement OS-level process limits and memory cgroups as a compensating control to prevent a single runaway process from exhausting the entire system.
Patch guidance
Apply updates from the vendor that include commit 28159fc. Test the patched version in a non-production environment to ensure compatibility with your existing Boxlite configurations and containerized workloads. Pay particular attention to any timeout-dependent workflows, as the corrected signal-handling mechanism may change how graceful shutdown occurs (though the new behavior should be more reliable). Once validated, roll out the patch systematically across your deployment, prioritizing instances that run untrusted or user-supplied code.
Detection guidance
Monitor Boxlite service logs for processes that continue running after timeout events are recorded. Track resource usage (CPU, memory) for containers that should have terminated but remain active. On the host system, examine process accounting or cgroup metrics to identify Box virtual machines with unexpectedly high resource consumption or many zombie processes. Alert on repeated timeout events followed by continued process activity for the same container or user. Review containerized application logs for code patterns that register SIGALRM handlers, as this is a strong indicator of exploit attempts.
Why prioritize this
Although CVSS is MEDIUM (6.5), this vulnerability warrants relatively swift remediation because it directly undermines a core security property of the Boxlite service—resource isolation and timeout enforcement. Denial-of-service vectors are often underweighted by CVSS, yet in a sandboxing context, availability failure can block legitimate work and create cascading impact. Prioritize if you use Boxlite for multi-tenant code execution or accept code from untrusted sources. Lower priority if Boxlite is used only internally with trusted code and strong network segmentation.
Risk score, explained
The CVSS 6.5 score reflects High impact on Availability (A:H), Low authentication requirement (PR:L), Network reachability (AV:N), and no impact on Confidentiality or Integrity. The score captures the denial-of-service severity but does not account for the specific trust model of a sandbox—where the attacker is already inside the boundary and the goal is containment. In that context, the practical risk may feel higher to teams managing critical sandboxed workloads. Conversely, if Boxlite is lightly used or isolated, risk is lower. Apply your organization's risk appetite and workload criticality to contextualize this base score.
Frequently asked questions
Can an attacker outside the Boxlite service exploit this directly?
No. The attacker must already have credentials and permission to launch a container within a Box (Privilege: Low). The vulnerability allows code running inside the sandbox to misbehave; it does not grant external code direct access. However, in a multi-tenant environment, one user's code can disrupt others' work.
What is the difference between SIGALRM and SIGKILL?
SIGALRM is a catchable signal—application code can register a handler and ignore or delay the shutdown. SIGKILL is uncatchable; the kernel terminates the process immediately. For timeout enforcement in a sandbox, SIGKILL is the correct choice because it prevents malicious code from circumventing the control.
Do I need to restart services after patching Boxlite?
Very likely. Check the vendor's release notes and deployment guide. Most service updates require a restart or rolling update to apply the fix. Test in a staging environment first to validate the restart procedure and confirm that containerized workloads resume correctly.
What if we cannot patch immediately?
Implement compensating controls: restrict Box container launches to trusted users, enforce strict OS-level resource limits (cgroups, ulimits), monitor resource usage and timeout events closely, and consider temporarily disabling user-submitted code execution until the patch is applied. Network segmentation and authentication logging will help detect suspicious activity.
This analysis is based on the CVE description and public information available as of the publication date. Specific patch version numbers, vendor advisory links, and deployment procedures should be verified directly with Boxlite's official documentation and release notes. No exploit code is provided or endorsed. Organizations should conduct their own risk assessment, testing, and validation before applying patches in production environments. The CVSS score is a baseline; your organization's actual risk may differ based on workload criticality, network architecture, and threat landscape. Source: NVD (public-domain), retrieved 2026-07-20. Analysis generated by SEC.co (claude-haiku-4-5).
Weaknesses (CWE)
Related vulnerabilities
- CVE-2026-10113MEDIUMOpen5GS NF-Profile Parser Denial of Service Vulnerability
- CVE-2026-10115MEDIUMOpen5GS NF Profile Parser DoS Vulnerability
- CVE-2026-10116MEDIUMOpen5GS UE Authentication Denial-of-Service Vulnerability
- CVE-2026-10117MEDIUMOpen5GS nghttp2-server Denial of Service Vulnerability
- CVE-2026-10156MEDIUMOpen5GS Resource Exhaustion Vulnerability in nf-instances Endpoint
- CVE-2026-10190MEDIUMTenda W12 Web Interface Denial-of-Service Vulnerability
- CVE-2026-10224MEDIUMNousResearch hermes-agent Webhook Resource Exhaustion Vulnerability
- CVE-2026-10650MEDIUMLibwebsockets SSH Handler Resource Exhaustion Vulnerability