← All updates

Security / News explained

A NAT rule could cross Check Point’s protected Sync boundary

CVE-2026-120490 requires an administrator who can edit and install policy and NAT rules. On affected Scalable Platforms, that authority could reach the Sync interface and lead to superuser operations.

By Zyel · Published · 4 min read

Check Point published CVE-2026-120490 on November 12, 2026, describing a protection bypass on Scalable Platforms. An administrator who can edit and install Access Control Policy and NAT rules could redirect traffic to a gateway’s Sync interface. The CVE record says that path can enable operating-system command execution and file writes with superuser privileges. [2]

The prerequisite is substantial: the person or account already needs policy and NAT administration powers. The vendor’s CVSS 3.1 record assigns a 3.8 score and marks required privileges as high. Those conditions belong beside the potential outcome when explaining the risk. They help an owner identify the relevant authority boundary. [2]

An internal interface can become a translated destination

The CVE record identifies Scalable Platforms comprising ElasticXL, Maestro and Chassis, on R82 with Jumbo Hotfix Take 130 or below and R81.20 Take 173 or below. The vendor advisory lists R82.10 and R82.20 as unaffected. The published fixes for the two affected branches are R82 Take 132 and R81.20 Take 174. [1] [2] [3] [4]

Sync is part of the platform’s internal communications. The Maestro guide assigns 192.0.2.0/24 to internal synchronization between Orchestrators on the same site. The ElasticXL guide also documents that subnet for synchronization and calls for a dedicated Layer 2 switch or VLAN connecting the cluster’s Sync ports. [5] [6]

According to the advisory, the built-in Sync protection fails to apply when a NAT rule translates the destination to that interface. The significant change is where the traffic ends up after translation. A boundary enforced against one representation of a destination can be defeated if another stage of processing gives the traffic a different destination. [1]

Our interpretation is that policy authority and appliance authority need an explicit relationship. Someone may legitimately manage application connectivity without being intended to control the gateway’s operating system. A feature that changes packet destinations can affect that separation, so its consequences deserve attention beyond whether the application’s address looks correct.

As a hypothetical example, an organization delegates policy maintenance to an operations team while reserving appliance administration for another team. A translation rule that reaches a protected internal service could connect those two powers in an unintended way. The CVE describes that class of boundary crossing. This example supplies no malicious rule, exploit procedure or claim about an actual organization.

The mitigation includes the installed policy

The advisory recommends reviewing and removing NAT rules that translate traffic to the Sync network, then installing the changed Security Policy. It also calls for securing the Layer 2 network between Maestro Orchestrators when their external synchronization ports connect through switches. Those are distinct considerations: the policy path and the synchronization network itself. [1]

That policy-installation detail has an operational consequence. A proposed rule change and an enforced rule change represent different states of the system. In our assessment, a remediation record should make clear which policy is active on the affected platform. Merely finding an edited rule in a management session leaves the resulting gateway behavior unresolved.

Check Point’s November 12 security blog reports no known in-the-wild exploitation of this CVE. The advisory’s rulebase review guidance identifies a condition to investigate; finding a matching translation would still require context before attributing intent or concluding that an attacker used it. The public sources do not establish the state of a reader’s deployment. [7]

Review the power to change a packet’s destination

For a technical lead, the interesting question is how delegated network administration interacts with internal services. A permission that appears to manage business traffic may also alter reachability around the platform’s own control functions. That can justify reviewing approval boundaries for sensitive policy changes, even when the account already has significant privileges.

The same reasoning applies to change review. A rule can look ordinary when read only as an external address mapping. Understanding its effective destination and the service exposed there gives the reviewer a more useful picture. Translation changes the address used by the next component; it does not make that component’s security requirements disappear.

For this finding, the practical destination is the vendor’s fixed release for the installed branch, with any relevant policy changes actually enforced and the synchronization network protected as documented. The broader lesson is to assess a packet’s path through the whole policy pipeline. Internal infrastructure can become reachable through configuration powers that seem unrelated at first glance.

Sources + further reading

Source dates describe the original material. Publication above is the date this Zyel article went live.

  1. Scalable Platforms Sync protection advisory, sk1000364Check Point · Checked Nov 14, 2026
  2. Check Point-authored CVE-2026-120490 recordCVE Program / Check Point · Checked Nov 14, 2026
  3. R82 Jumbo Hotfix Take 132 release notesCheck Point · Checked Nov 14, 2026
  4. R81.20 Jumbo Hotfix Take 174 release notesCheck Point · Checked Nov 14, 2026
  5. Maestro internal IP addressesCheck Point · Checked Nov 14, 2026
  6. ElasticXL important notes, R82Check Point · Checked Nov 14, 2026
  7. Security advisory, November 12, 2026Check Point · Checked Nov 14, 2026

Reader discussion

What would you ask next?

Questions, useful corrections and thoughtful disagreement are welcome. Comments are reviewed before they appear.

No reader comments yet.

10–2,000 characters. Plain text. Leave out passwords, private project data and personal contact details. A blank display name appears as Anonymous.

Complete the CAPTCHA before sending. Verification runs on your device.