CVE-2026-53866: OpenClaw Shell Inline-Command Allowlist Bypass (CVSS 8.1)
OpenClaw, a tool used by authenticated operators, has a flaw in how it processes shell commands. An attacker with valid credentials could bypass the system's approval controls and execute commands that should have been blocked. The issue stems from a parsing gap where certain command formats skip the required allowlist check, letting unauthorized shell commands run without triggering approval prompts.
Source data · NVD / CISA · public domain
- CVSS
- 3.1 · 8.1 HIGH · CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N
- Weaknesses (CWE)
- CWE-862
- Affected products
- 9 configuration(s)
- Published / Modified
- 2026-06-16 / 2026-06-18
NVD description (verbatim)
OpenClaw before 2026.5.12 contains an allowlist bypass vulnerability in shell inline-command parsing that allows authenticated operators to execute unapproved commands. A command request using shell inline-command forms could route through a parser case missing the expected allowlist decision, enabling shell content execution without intended approval prompts.
2 reference(s) · View on NVD →
SEC.co analysis · AI-assisted, reviewed against source
Technical summary
CVE-2026-53866 is an authorization bypass vulnerability in OpenClaw versions before 2026.5.12. The vulnerability exists in the shell inline-command parsing logic, where specific command request patterns fail to route through the allowlist decision point. This parser case omission permits authenticated users to execute shell commands that would normally be subject to approval workflows, violating the intended access control model. The vulnerability is classified under CWE-862 (Missing Authorization).
Business impact
This vulnerability enables credential-holding insiders or compromised operator accounts to execute unauthorized shell commands on OpenClaw systems without triggering approval workflows. Depending on OpenClaw's operational role, this could lead to data exfiltration, system manipulation, lateral movement, or sabotage—all while circumventing governance checkpoints designed to prevent such actions. Organizations relying on OpenClaw's approval mechanisms for compliance or operational safety face a direct bypass of those controls.
Affected systems
OpenClaw versions prior to 2026.5.12 are affected. All installations running earlier versions are vulnerable if they rely on shell inline-command functionality and approval workflows. Environments where OpenClaw is used by multiple operators or in multi-tenant scenarios face heightened risk, as the allowlist bypass could be exploited by any authenticated user.
Exploitability
The vulnerability requires valid operator credentials to exploit—it is not unauthenticated. However, once authenticated, exploitation is straightforward: an attacker simply crafts a shell inline-command request using the specific parser case that lacks allowlist enforcement. No complex interaction, timing, or user intervention is needed. The attack succeeds reliably in vulnerable versions.
Remediation
Upgrade OpenClaw to version 2026.5.12 or later to obtain the patch that corrects the parser case and reinstates allowlist enforcement for all shell inline-command forms. Organizations unable to upgrade immediately should restrict operator credentials and monitor shell command execution logs for suspicious patterns that bypass approval workflows.
Patch guidance
Apply OpenClaw update 2026.5.12 or later. Verify the patch version against the vendor advisory to confirm it addresses the shell inline-command parsing bypass. Test the patch in a staging environment to ensure approval workflows function as expected before production deployment. Consider rolling out patches in phases if OpenClaw is mission-critical.
Detection guidance
Monitor OpenClaw logs for shell commands executed without corresponding approval records or for command requests that skip the approval decision point. Look for suspicious shell inline-command usage patterns, particularly from operator accounts with lower privilege levels. If available, enable detailed parser-level logging to catch requests that might traverse the vulnerable code path. Correlate shell command execution with approval workflow logs to identify gaps.
Why prioritize this
This is a HIGH-severity issue (CVSS 8.1) because it enables authenticated attackers to bypass intended security controls, achieve high-impact outcomes (data confidentiality and integrity breaches), and operate reliably across all network conditions. While it requires valid credentials, the ease of exploitation and the direct violation of governance controls make it a critical priority for operators who depend on approval workflows for compliance or security posture. Organizations should treat this as urgent.
Risk score, explained
The CVSS 3.1 score of 8.1 reflects a network-accessible vulnerability requiring a low attack complexity and low privilege level (authenticated user). The impact assessment shows high confidence and integrity impact with no availability impact. The score correctly penalizes the bypass of authorization controls while acknowledging the authentication prerequisite. No CVSS temporal adjustments are included here; however, the fact that this is not yet in the KEV catalog should not lower its priority—internal approval workflow bypasses are often exploited by insiders before broad public attention.
Frequently asked questions
Can this vulnerability be exploited by unauthenticated users?
No. The vulnerability requires valid OpenClaw operator credentials. However, this does not lower the risk if operator credentials are shared, weak, or compromised.
Does the patch break existing approval workflows?
No. Version 2026.5.12 repairs the parser case to ensure all shell inline-command requests are routed through the allowlist decision. Legitimate commands that passed approval before will continue to work.
How can we detect if someone has exploited this in our environment?
Review OpenClaw logs for shell commands that lack matching approval records, and check for operator activity that skips approval decision logs. If detailed parser logging is available, enable it to catch requests traversing the vulnerable code path.
Is there a temporary workaround if we cannot patch immediately?
Consider restricting shell inline-command permissions to a minimal set of trusted operators until you can patch. Monitor those operators closely and require explicit approval documentation for any shell command execution.
This analysis is based on the vendor advisory and public vulnerability data as of the publication date. Patch version numbers and remediation steps should be verified against the official OpenClaw security advisory. Exploitation risk may vary depending on your specific OpenClaw deployment, operator credential controls, and network isolation posture. Test any patches in a staging environment before production rollout. This page does not constitute professional security advice; consult your security team and OpenClaw support for deployment-specific guidance. Source: NVD (public-domain), retrieved 2026-07-24. Analysis generated by SEC.co (claude-haiku-4-5).
Related vulnerabilities
- CVE-2026-32905HIGHOpenClaw Device-Pair Authorization Bypass (CVSS 8.3)
- CVE-2026-35630HIGHOpenClaw Authorization Bypass in QQBot Approval Buttons
- CVE-2026-53816HIGHOpenClaw Event Provenance Validation Bypass (CVSS 7.2)
- CVE-2026-53821HIGHOpenClaw WebSocket Authorization Bypass (CVSS 8.8)
- CVE-2026-53815MEDIUMOpenClaw Authorization Bypass in Channel Message Read
- CVE-2026-53818MEDIUMOpenClaw MCP Loopback Authorization Bypass (CVSS 6.6)
- CVE-2026-53820MEDIUMOpenClaw MCP Denylist Bypass Allows Privilege Escalation
- CVE-2026-53844MEDIUMOpenClaw Session Visibility Bypass in Shared Memory Search