Decoding the world of cybersecurity

Check Point gateway attacks reached VMware hosts

Check Point gateway intrusions reached VMware hosts and destroyed files. Investigators documented chained vulnerabilities and persistent backdoors, while responsibility for the destructive attacks remains unresolved.

Check Point gateway attacks reached VMware hosts
Summary
  • Attackers combined PHP2Shell vulnerabilities with Check Point privilege escalation flaw CVE-2026-123039 to compromise security gateways.
  • Investigators documented persistent gateway backdoors and movement into VMware ESXi, including destructive malware in some incidents.
  • CISA added the privilege escalation vulnerability to its Known Exploited Vulnerabilities catalogue on 13 November, with a federal remediation deadline of 16 November.

Attackers exploited weaknesses in Check Point security gateways to obtain privileged access and move into internal VMware infrastructure, with investigators documenting destructive malware in some organisations targeted during an initial attack wave.

Check Point published a technical investigation on 12 November describing attacks that began on 30 October against unpatched gateways exposing Mobile Access portals.

The operators combined two vulnerabilities associated with PHP2Shell with a separate local privilege escalation weakness, CVE-2026-123039, to gain extensive control over affected appliances.

In one investigated intrusion, the attackers established persistent access to a gateway, examined the internal network and reached a VMware ESXi host, where they deployed destructive malware.

Check Point identified two broad waves of activity. The first concentrated on Israeli organisations and included destructive attacks, while a second wave on 2 November extended across multiple countries and sectors without confirmed destructive activity.

Initial access led to privileged control

The initial exploitation involved CVE-2026-109618 and CVE-2026-109619, two PHP vulnerabilities affecting the exposed Mobile Access service on certain Check Point gateways.

Once the attackers obtained execution through the vulnerable portal, they exploited CVE-2026-123039 to escalate privileges within the underlying Gaia operating system.

The latter flaw affects DynamicCLI, a component that permits communication with privileged system functions. Check Point rates the vulnerability at 7.8 under CVSS version 3.1.

It is a local privilege escalation issue, not a separate unauthenticated internet-facing vulnerability. An attacker requires an initial foothold or suitable local execution capability before using it to obtain the higher privileges demonstrated in the investigation.

In the documented attack chain, PHP2Shell supplied that foothold.

The distinction has practical consequences for exposure assessment. Installing a correction for the privilege escalation issue reduces the potential impact of an attacker already operating on the device, while preventing initial compromise also requires addressing vulnerabilities affecting exposed services.

Check Point says it has observed exploitation of CVE-2026-123039. The affected product families include security gateways and management systems running specified Gaia releases.

Gateway persistence enabled internal movement

After obtaining privileged execution, the attackers installed malicious components intended to preserve access to the compromised gateway.

These included PHP webshells and a persistent backdoor identified as TSGA, together with an associated networking component called TSProxy.

The components provided remote control and mechanisms for interacting with the gateway while attempting to avoid conventional detection. Check Point also documented attempts to interfere with logging and maintain access across appliance operations.

That access is particularly significant for a device responsible for enforcing network boundaries. Security gateways commonly handle connections between external services, remote users and internal business infrastructure.

Once compromised, the device may provide visibility or connectivity unavailable to an external attacker operating solely through public-facing services.

In the investigated intrusion, the operators used the gateway to examine internal systems before moving to a VMware ESXi host through an available service account.

They subsequently deployed a Linux wiper known as destroy_v6, designed to delete files rather than encrypt them for later recovery.

The activity was accompanied by politically themed messages. Check Point found no evidence in that case of a conventional ransom demand associated with the destructive stage.

Compromise of virtualisation infrastructure can extend an incident beyond one appliance. A single ESXi host may support several virtual servers, applications and business services, meaning destructive activity at that level can affect multiple workloads.

However, Check Point has not established that every targeted gateway was successfully compromised or that destructive malware was deployed throughout the wider campaign.

Shared malware does not establish attribution

The investigation identified technical links between TSGA, TSProxy and the PhiliKit malware framework previously associated with attacks against network security appliances.

Earlier research connected PhiliKit to suspected China-linked activity, including operations involving the group tracked as UNC5221.

The tools in the Check Point investigation shared distinctive code and configuration characteristics with that earlier framework, supporting an assessment that they came from related tooling.

Nevertheless, the researchers did not attribute the current campaign confidently to a particular group or state.

The destructive operation, concentration on Israeli targets during the first wave and politically themed messages did not align neatly with all previously observed characteristics of the suspected China-linked activity.

Check Point considered several explanations, including the possibility that different operators shared infrastructure, that an initial access broker supplied another actor, or that some elements were intended to mislead investigators.

None of those scenarios has been established as the confirmed explanation. The available evidence supports a relationship between the malware families but does not establish responsibility for the destructive attacks.

Exploited flaw receives an emergency deadline

Check Point has released security fixes covering CVE-2026-123039 and other vulnerabilities affecting its Gaia-based products, with instructions to install the relevant Jumbo Hotfix for each software branch.

The company’s security guidance identifies affected versions and alternative mitigation arrangements for organisations unable to install an update immediately.

On 13 November, the US Cybersecurity and Infrastructure Security Agency added CVE-2026-123039 to its Known Exploited Vulnerabilities catalogue.

The entry carries a remediation deadline of 16 November for covered US federal systems under the applicable directive. It does not impose that same statutory deadline automatically on private businesses or organisations outside the relevant federal requirements.

The catalogue entry confirms exploitation, while the vendor’s research provides a fuller account of how the vulnerability was used after the initial gateway compromise.

Where an organisation identifies indicators of intrusion, applying a security update does not establish that persistent malware has been removed. The investigated attacks involved backdoors on the gateway and movement into internal systems, creating a need to determine whether compromise extended beyond the original appliance.

Check Point’s findings establish destructive activity in some cases associated with the 30 October wave. The broader 2 November activity has not been shown to involve the same destructive outcome, and the identity of the responsible operators remains unresolved.

×