NullHub NullClaw Auth Bypass Lets Attackers Hijack Agents and Configs

NullHub NullClaw contains a missing authentication vulnerability that exposes management interfaces to unauthenticated attackers. This flaw, tracked as CVE-2026-110616, affects versions through 2026.5.29. The issue allows remote users to access sensitive API endpoints without providing valid credentials. This exposure occurs when the service listens on non-loopback network interfaces. Attackers can exploit this gap to take control of agent instances and modify system configurations. The National Vulnerability Database rates the severity as critical with a CVSS score of 9.2.

Root cause

The underlying issue stems from how NullHub handles authentication tokens for its management API. According to the NVD record, the auth_token field defaults to null and is never assigned a valid value during operation. This logical error means that requests sent to the /api/* endpoints are not properly validated against user credentials. The vulnerability is classified under CWE-306, which covers missing authentication for a critical function. Because the token remains null, the application fails to reject unauthorized access attempts. This design flaw effectively disables authentication for any request that reaches the management interface.

Attack path

An attacker can exploit this vulnerability by sending requests to the NullHub management API without including Authorization or Origin headers. The NVD description notes that this works against non-loopback listeners, meaning any server exposed to a network or the internet is at risk. Once connected, the unauthenticated user can control agent instances directly. They can also rewrite system configurations, which includes altering provider credentials stored within the application. Additionally, the attacker gains the ability to manage cron jobs and Model Context Protocol (MCP) servers. This level of access allows for persistent control and potential credential theft.

Affected versions

• NullHub NullClaw versions prior to and including 2026.5.29 are affected by this vulnerability.
• The flaw exists in the management API implementation of these specific releases.
• Any deployment using versions up to 2026.5.29 is exposed to unauthenticated access.
• Organizations running older builds remain vulnerable.

Mitigation

• Restrict network access to NullHub management interfaces to trusted internal networks only.
• Ensure that listeners are bound to loopback addresses unless remote access is strictly required.
• Monitor logs for unauthorized access attempts to the /api/* endpoints.
• No fix has been confirmed yet.

Signals worth monitoring

Teams usually review web server access logs for requests to management endpoints that lack standard authentication headers. Look for successful responses to API calls from unexpected IP addresses or source ports. Alert on any configuration changes made without corresponding user sessions. Check for new cron job entries or modifications to provider credentials in the application logs. Investigate any traffic to non-loopback interfaces that targets the /api/* paths. Verify that all administrative actions are tied to valid, authenticated user accounts. Review firewall logs for external connections to management ports.

What to do and how to stay safe: NullHub

• Audit your NullHub deployments to confirm whether they are listening on non-loopback interfaces.
• Implement network segmentation to prevent external access to management APIs.
• Monitor application logs for unauthorized configuration changes or credential updates.
• Once the vendor provides an update, apply it immediately to mitigate the risk.

Step-by-step guide: Attack Surface Management: 6 Common Misreadings and Correct Views

General security guidance from the NetSec Observer newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

What is the CVSS score for CVE-2026-110616?

The National Vulnerability Database assigns a CVSS score of 9.2, which is rated as critical.

Which versions of NullHub are affected by this vulnerability?

NullHub NullClaw versions through 2026.5.29 are affected by the missing authentication flaw.

Can an attacker modify provider credentials using this bug?

Yes, attackers can rewrite configurations including provider credentials by accessing the management API.