Check Point unpatched: TSGA backdoor and ESXi wiping
Attackers use the PHP2Shell flaws and a privilege escalation in Gaia OS to enter Check Point gateways, leave the TSGA backdoor and delete ESXi files.
4 min read

In this article
A two-phase attack against the network's entry point
Check Point has detailed a campaign that began on October 30, 2026 and that targeted Israeli organizations, although the vendor points to broader tracking later on. The threat actor's modus operandi, still unidentified, is a lesson in patience and in exploiting the weakest link: first it enters through the flaws known as PHP2Shell in security gateways that were not up to date, and then it escalates privileges with a local vulnerability in Gaia OS that had not been documented until now.
The result is a backdoor dubbed TSGA and, in some cases, the deletion of ESXi files. In other words, the attacker is not content with looking: it installs itself, stays and, if it suits it, destroys. For an SME, a Check Point gateway is not just another gadget; it is the border between its network and the rest of the world. If it falls, everything behind it falls.
Why this affects us even if we are small
The usual temptation in an SME is to think that these headlines are a matter for large corporations or intelligence services. That is not the case. Attackers do not always choose by size, but by ease. An unpatched gateway, with an exposed management interface and no segmentation, is an opportunity too convenient to ignore. And the pattern repeats itself: first the known flaw is sought, it is exploited, and then any local weakness is used to consolidate access.
The Gaia OS detail is especially uncomfortable because we are talking about a local privilege escalation vulnerability that was not known. That means it is not enough to have the gateway "reasonably up to date": real patching discipline is needed, with defined windows and subsequent verification. And, above all, it is necessary to assume that the perimeter can be broken and that the response cannot depend only on a device at the edge of the network.
A compromised gateway is not a perimeter incident: it is a foothold inside your infrastructure.
What we recommend from ForgeNEX
The first thing is to review the fleet of Check Point gateways and confirm the status of the updates. If there are pending devices, that is the absolute priority of the week. We are not talking about routine maintenance: we are talking about closing an entry route that is already being used in real attacks.
- Inventory and patching. Locate all gateways, check the version and apply the updates published by the vendor. Document what has been done and when.
- Reduce the surface. The management interface should not be accessible from anywhere. Restrict access by origin, use VPN or controlled jumps and review privileged accounts.
- Watch for what is not normal. New processes, unexpected outbound connections, configuration changes and access outside business hours. The TSGA backdoor does not install itself: it leaves a trace if you know where to look.
- Protect ESXi. If your virtualization runs on VMware, review who can delete files and how backups are protected. An attack that destroys virtual disks can only be countered with isolated and tested backups.
- Response plan. Be clear about who you call, what you isolate first and how you communicate. Improvising in the middle of an incident is the most expensive way to lose time.
This case fits a trend we have already discussed on the blog: attackers exploit the known and the unknown in the same operation. We saw it with UnrealIRCd and with SCKit and ScreenConnect. The lesson is the same: patching is not a formality, it is the difference between a scare and an incident with real damage.
The deeper analysis
The combination of an initial access flaw with an unknown local escalation reminds us that security is not one layer, it is several. An SME may not have its own SOC, but it can have inventory, patching, restricted access and backups that work. That does not make it invulnerable, but it takes it off the list of easy targets.
If you manage infrastructure for clients, this is the time to review contracts and procedures: who applies the gateway patches, how often, how is it verified? And if you are the IT manager of an SME, do not wait for the vendor to publish a more alarming advisory. The campaign is already underway and the deletion of ESXi files shows that the attacker has no intention of being discreet.
Source: The Hacker News. Analysis and adaptation: ForgeNEX.
Keep reading
- UnrealIRCd: exploited flaw allows command injection on the server15 November 2026
- Deepfakes and crypto: the North Korean fraud that deceives via video call13 November 2026
- SCKit exploits ScreenConnect: patch now or expose your remote support13 November 2026
- Critical Apache HTTP/2: unauthenticated remote code execution12 November 2026
- Qlik Sense, new target for reverse shells and exploits11 November 2026