OpenClaw CVE-2026-41371

Exploitation requires valid OpenClaw credentials with write-scoped gateway access, which are typically provisioned to legitimate application integrations or service accounts for chat messaging functionality. The vulnerable chat.send endpoint must be network-accessible to the attacker, either directly or through an authenticated gateway API. No special OpenClaw configuration or non-default features are required - the vulnerability exists in the default authorization implementation of the chat.send path. Exploitation is limited to authenticated attackers who have obtained write-scope credentials through legitimate provisioning, credential compromise, or insider access; anonymous unauthenticated attackers cannot exploit this vulnerability. The impact is further constrained to session management operations (reset, rotation, archival) rather than arbitrary code execution or data exfiltration beyond transcript manipulation.

Real-world risk is HIGH despite no confirmed exploitation. CVSS 8.4 accurately reflects the severe impact: high integrity violation in both vulnerable (VI:H) and subsequent (SI:H) system components, enabling attackers to manipulate session state and transcript data across user boundaries. Attack vector is network-accessible (AV:N) with low complexity (AC:L) and no user interaction (UI:N), making exploitation straightforward for any authenticated low-privilege attacker. The scope change from write-level to admin-level operations represents a critical authorization boundary violation. However, exploitation requires valid low-privilege credentials (PR:L), limiting attack surface to authenticated insiders or compromised accounts rather than anonymous internet attackers. EPSS data not provided, but the technical characteristics (simple API call, no complexity, high impact) suggest moderate exploitation probability if credentials are obtained. No CISA KEV listing indicates targeted rather than widespread exploitation. Priority should be HIGH for organizations running OpenClaw in multi-tenant or shared environments where session integrity is critical, MEDIUM for isolated single-user deployments.

An authenticated attacker with write-scoped gateway credentials crafts a malicious chat.send API request containing admin-level session reset commands targeting another user's active session. The vulnerable OpenClaw instance fails to validate the caller's authorization scope and executes the reset operation, forcibly terminating the victim's session, rotating their session ID, and archiving their chat transcript. The attacker can repeat this process to persistently disrupt service for targeted users or manipulate session state to enable subsequent attacks such as session fixation or data exfiltration from archived transcripts.