Openclaw CVE-2026-27524

The CVSS 3.1 score of 3.1 with vector CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:N indicates low overall severity, but this scoring reflects several mitigating factors that are contextually important. Attack Vector is Network (AV:N), which is unfavorable for defenders, however Access Complexity is High (AC:H), significantly reducing real-world exploitability. The requirement for prior authentication (PR:L) acts as a substantial gating factor, limiting the attacker pool to authorized users. The integrity impact is limited (I:L) with no confidentiality or availability impact, suggesting the attack enables subtle manipulation rather than wholesale system compromise. The absence of User Interaction (UI:N) means exploitation can be automated once access is obtained. While no EPSS score or KEV status is provided in the available intelligence, the combination of authentication requirement, high complexity, and limited impact suggests this is not an actively exploited vulnerability in the wild. The availability of a vendor patch and the relatively constrained attack surface (only /debug endpoint with prior auth) make this a lower priority compared to similar CVSS scores without authentication requirements, though it warrants timely patching for defense-in-depth.

An insider with authenticated access to the OpenClaw application discovers the /debug set endpoint and crafts a request injecting __proto__ keys to modify the prototype chain of security-related objects. By poisoning the prototype, they manipulate the command gate restriction logic that normally prevents execution of privileged commands, allowing them to bypass authorization checks and execute restricted operations. This attack is possible because the /debug endpoint does not validate or block prototype-reserved keys during override processing, trusting that authenticated users would not attempt such manipulation.