Openclaw CVE-2026-35649

This vulnerability presents moderate but meaningful real-world risk based on multiple converging signals. The CVSS score of 6.5 with AV:N (network), AC:L (low complexity), and PR:N (no authentication required) indicates unauthenticated remote attackers can exploit this flaw. However, the attack does not result in confidentiality loss (C:L) or integrity loss (I:L) - it specifically re-enables previously revoked access rather than creating new compromise vectors. The lack of publicly disclosed active exploitation (no KEV status provided) and absence of public proof-of-concept code means this is not yet weaponized, though the relatively straightforward nature of the logic flaw and low attack complexity suggest PoC development would be feasible. From an SSVC perspective, the technical impact is meaningful (permission restoration), but the automatable nature and lack of special conditions make this exploitable by a broad attacker population. Organizations managing access control policies in OpenClaw should prioritize patching, particularly in environments where denying access to specific resources has been explicitly configured.

An attacker with network access to an OpenClaw instance could trigger the settings reconciliation process after an administrator has explicitly configured an empty allowlist to deny all access to a sensitive resource. The attacker exploits the reconciliation logic by interacting with the configuration system in a manner that causes the empty allowlist to be treated as unset, causing OpenClaw to restore the previous access permissions. This could occur repeatedly, allowing the attacker to continuously re-enable access to resources that were intended to be restricted without needing authentication credentials or special system privileges. No public proof-of-concept has been identified at the time of analysis, but the straightforward nature of the logic flaw suggests exploitation would be relatively simple once understood.