CVE-2026-35662
Access Control Bypass in OpenClaw send Action Allows Unauthorized Messaging
Publication date: 2026-04-10
Last updated on: 2026-04-13
Assigner: VulnCheck
Description
Description
CVSS Scores
EPSS Scores
| Probability: | |
| Percentile: |
Meta Information
Affected Vendors & Products
| Vendor | Product | Version / Range |
|---|---|---|
| openclaw | openclaw | to 2026.3.22 (exc) |
Helpful Resources
Exploitability
| CWE ID | Description |
|---|---|
| CWE-862 | The product does not perform an authorization check when an actor attempts to access a resource or perform an action. |
Attack-Flow Graph
AI Powered Q&A
What immediate steps should I take to mitigate this vulnerability?
The immediate mitigation step is to upgrade OpenClaw to version 2026.3.22 or later, where the vulnerability has been fixed.
- Apply the patch that enforces strict controlScope and ownership checks on the 'send' action, preventing leaf subagents from messaging child sessions outside their authorized scope.
- Ensure that your deployment uses the fixed versions 2026.3.22, 2026.3.23, or 2026.3.23-2, which include the security fix.
If upgrading immediately is not possible, consider restricting or monitoring subagent messaging capabilities to prevent unauthorized 'send' actions until the patch can be applied.
How does this vulnerability affect compliance with common standards and regulations (like GDPR, HIPAA)?:
The vulnerability CVE-2026-35662 allows leaf subagents in OpenClaw to bypass intended access control restrictions by sending messages to child sessions beyond their authorized scope. This improper authorization enforcement could potentially lead to unauthorized access or actions within the system.
Such unauthorized access or control bypass may impact compliance with common standards and regulations like GDPR or HIPAA, which require strict access controls and protection of sensitive data. Failure to enforce proper authorization could result in unauthorized data exposure or manipulation, thereby violating these regulatory requirements.
However, the provided context and resources do not explicitly discuss the direct impact of this vulnerability on compliance with GDPR, HIPAA, or other standards.
Can you explain this vulnerability to me?
CVE-2026-35662 is a vulnerability in OpenClaw versions prior to 2026.3.22 where the software fails to enforce controlScope restrictions on the send action.
This flaw allows leaf subagents to send messages to controlled child sessions beyond their authorized scope, effectively bypassing intended access control restrictions.
The root cause is improper authorization checks during the send action, where leaf subagents could communicate with child sessions they should not have access to.
The vulnerability corresponds to CWE-285 (Improper Authorization) and CWE-862 (Missing Authorization).
How can this vulnerability impact me? :
This vulnerability can allow attackers who control leaf subagents to bypass access control restrictions and send unauthorized messages to child sessions.
Such unauthorized messaging can lead to unauthorized actions or information flow within the system, potentially compromising the integrity of session controls.
Because the attacker can communicate beyond their authorized scope, this could lead to privilege escalation or manipulation of controlled sessions.
How can this vulnerability be detected on my network or system? Can you suggest some commands?
Detection of this vulnerability involves verifying whether leaf subagents are able to send messages to child sessions beyond their authorized control scope. Since the flaw allows unauthorized messaging via the 'send' action, monitoring or testing the behavior of subagents sending messages to nested child sessions can help identify the issue.
Specifically, you can attempt to execute the `/subagents send` command from a leaf subagent with a restricted controlScope and observe if the system allows messaging to child sessions it should not control. If the command succeeds without proper authorization errors, the system is vulnerable.
There are no explicit detection commands provided in the resources, but testing the authorization enforcement on the 'send' action by simulating subagent messaging attempts outside their control scope is recommended.