CVE-2026-35673: OpenClaw SSRF Policy Bypass in Debug and Export Routes

Technical summary

CVE-2026-35673 is an improper authorization vulnerability (CWE-863) affecting OpenClaw's debug and export routes. The flaw allows an authenticated user to bypass SSRF policies by manipulating the browser tab state; specifically, by reusing tabs that were previously blocked or restricted. The vulnerability permits circumvention of private-network access controls, enabling export or inspection of protected content. The CVSS 3.1 score of 6.5 reflects a network-exploitable attack requiring authentication and user interaction, with high confidentiality impact but limited integrity and no availability impact. The attack surface is contingent on route access and the presence of blocked tabs in the browser session.

Business impact

This vulnerability primarily threatens organizations that rely on OpenClaw for secure content inspection or export workflows, particularly in environments handling sensitive internal data. Insider threats or compromised user accounts with access to debug/export routes become significantly more dangerous, as the SSRF bypass allows them to extract or inspect data from private network resources that should remain isolated. For teams using OpenClaw in security-sensitive or compliance-heavy roles, the ability to bypass SSRF policies represents a meaningful data exfiltration risk. However, the requirement for prior authentication and user interaction limits the blast radius compared to unauthenticated or fully automatic exploits.

Affected systems

OpenClaw versions prior to 2026.4.29 are affected. Organizations running OpenClaw should verify their installed version and upgrade to 2026.4.29 or later. The vulnerability affects the browser debug and export functionality specifically, so only deployments actively using these features are exposed to the risk.

Exploitability

Exploitation requires an authenticated user (insider or compromised account) with access to OpenClaw's debug or export routes. The attacker must also interact with the browser interface to reuse blocked tabs, meaning fully passive or remote exploitation is not feasible. The conditional nature of this attack—needing both valid credentials and the presence of previously-blocked tabs—moderates exploitability. However, in operational environments, blocked tabs may persist in user sessions, making this a realistic attack path for an insider with appropriate route permissions.

Remediation

Upgrade to OpenClaw 2026.4.29 or later. Verify the upgrade against the vendor's official release notes or advisory to confirm that SSRF policy enforcement has been patched. Additionally, audit access controls to the debug and export routes, ensuring that only authorized personnel retain these permissions. Consider implementing session management controls to prevent reuse of sensitive browser states.

Patch guidance

Deploy OpenClaw version 2026.4.29 or later across all affected instances. Review your deployment's patching procedures and test the update in a non-production environment first to confirm compatibility with your configuration. If you are running a version between the initial release and 2026.4.29, prioritize upgrading. Consult OpenClaw's official release notes and security advisory for detailed upgrade steps and any breaking changes.

Detection guidance

Monitor access to OpenClaw's debug and export routes, particularly by users with limited typical access patterns. Log and review instances where these routes export or access private-network resources, especially if those resources would normally be blocked by SSRF policies. Inspect browser session logs or debug telemetry for anomalous tab reuse or state manipulation. Network-level monitoring may also reveal unexpected outbound connections from OpenClaw instances to private network ranges that should not be reachable.

Why prioritize this

Although the CVSS score is moderate (6.5), this vulnerability warrants prompt attention because it directly undermines security policy enforcement in a debugging/export context. Organizations that intentionally restrict access to certain network resources—via SSRF policies—rely on those controls for compliance and data protection. A policy bypass erodes that protection and creates an insider-threat amplification vector. Prioritize this vulnerability in environments where OpenClaw handles sensitive data or private network inspection.

Risk score, explained

The CVSS 3.1 score of 6.5 reflects: (1) Network accessibility (AV:N) and low attack complexity (AC:H) reflecting the need for authentication and user interaction; (2) Privilege requirement at the user level (PR:L), since the attacker must already have authenticated access to debug/export routes; (3) Requirement for user interaction (UI:R) to reuse tabs; (4) Scope change (S:C) because the SSRF bypass affects resources outside the original trust boundary; (5) High confidentiality impact (C:H) from unauthorized data access, and limited integrity impact (I:L) from potential data export manipulation. Availability is not impacted. The 'MEDIUM' severity reflects a meaningful but conditional threat.

Frequently asked questions