A Server-Side Request Forgery vulnerability in OpenClaw allows unauthenticated attackers to bypass security filters and perform unauthorized requests against internal network resources.

This vulnerability is a Server-Side Request Forgery (CWE-918) occurring in the SSRF protection mechanism. By utilizing specifically crafted IPv4-mapped IPv6 literals, an unauthenticated attacker can bypass filters intended to block access to loopback, private network, or link-local metadata services.

The ability to perform SSRF attacks against internal infrastructure poses a significant risk to organizational security. Attackers can leverage this flaw to interact with internal services that are not exposed to the public internet, potentially leading to unauthorized data access or the exposure of sensitive cloud metadata. Given the CVSS score of 7.5, this high-severity vulnerability requires immediate attention to prevent lateral movement within the network.

Immediate Action: Update the OpenClaw package to version 2026.2.14 or later immediately to apply the necessary input validation fixes.

Proactive Monitoring: Review application and network logs for anomalous outbound requests originating from the OpenClaw service, specifically looking for attempts to reach internal loopback or private IP ranges.

Compensating Controls: If immediate patching is not feasible, implement a Web Application Firewall (WAF) or egress filtering rules to restrict the application from initiating connections to internal or reserved IP address spaces.

Public Exploit Available: Unknown

This vulnerability represents a critical security oversight in the handling of address literals within the OpenClaw AI assistant. Organizations relying on this software should prioritize upgrading to version 2026.2.14 as the primary method of risk mitigation. Failure to patch allows for potential unauthorized access to internal network assets, which could be leveraged for further exploitation of the internal environment.