Openclaw CVE-2026-33579

Attacker must be an authenticated user with pairing privileges (but not admin privileges) on OpenClaw before version 2026.3.28. The attacker exploits missing scope validation in the /pair approve command to approve device pairing requests with elevated scopes including admin access. No victim interaction required.

Real-world risk is CRITICAL despite absence of confirmed active exploitation. The CVSS 9.8 vector (AV:N/AC:L/PR:N/UI:N) indicates this is remotely exploitable with no authentication required and low attack complexity-a best-case scenario for attackers. The PR:N (no privileges required) designation conflicts with the description's mention of 'pairing privileges,' suggesting either the pairing endpoint itself is unauthenticated or the vulnerability allows complete bypass of authentication checks. SSVC dimensions show high technical impact (admin access grants full confidentiality, integrity, and availability compromise) and likely automatable exploitation given the straightforward API abuse. No EPSS score provided, but network-accessible authentication bypasses leading to admin escalation are historically high-value targets. No KEV listing or public POC identified, suggesting exploitation has not been widely observed, but the low barrier to entry and high impact make this a patch-immediately priority. The vulnerability enables lateral movement and persistence in device management infrastructures.

An attacker identifies an OpenClaw instance exposing the device pairing API to the internet. Without requiring authentication (per CVSS PR:N), they submit a device pairing request claiming admin-level scopes. They then immediately call the /pair approve endpoint, which processes the approval without validating the caller's actual permission level due to the missing scope forwarding logic. The system approves the request, granting the attacker's device full administrative access to the OpenClaw infrastructure. From this position, the attacker can manage all connected devices, exfiltrate sensitive configuration data, deploy malicious payloads to managed endpoints, or establish persistent backdoor access. No public exploit identified at time of analysis, but the straightforward API abuse pattern and network accessibility make exploitation trivial for adversaries with basic API testing capabilities.