Check Point Gateway Intrusions: PHP2Shell, Persistence and Wiper Activity

Check Point Research describes two 2026 waves targeting unpatched Security Gateways, with PHP2Shell access, privilege escalation, persistent backdoors and, in some first-wave cases, destructive activity.

#cybersecurity#threat-intelligence#malware#incident-response#network-security

What the investigation found

Check Point Research reports that attackers targeted organizations running unpatched Check Point Security Gateways in two waves beginning on October 30, 2026. The activity centered on Mobile Access portals. The investigation describes PHP2Shell exploitation associated with CVE-2026-109618 and CVE-2026-109619, followed in some intrusions by privilege escalation through CVE-2026-123039 in the DynamicCLI interface.

The first wave focused on Israeli organizations, and some affected environments experienced destructive activity. A second wave on November 2 reached organizations in multiple countries and sectors. Check Point Research found no apparent destructive activity in that broader wave and saw no clear sector preference.

The report concerns gateway exploitation and malware deployment. It does not describe compromised eMudhra certificates or a TLS certificate incident, so it should not be treated as evidence for that separate claim.

From portal access to persistence

After gaining access through the Mobile Access portal, attackers escalated privileges from the portal service account to root. They then placed PHP webshells in a templates directory used to build login pages. The shells enabled command execution, PHP evaluation and file uploads, giving operators a foothold beyond the initial exploit.

Check Point Research says the operators subsequently deployed TSGA, a persistent Linux backdoor, as /tmp/CPconfigdaemon. When run as root, it relocates and relaunches under the name slot_CPconfigdaemon in a boot persistence directory. TSGA provides remote access and other appliance-specific functions, including configuration theft and process injection.

The implant also carries TSProxy, a networking component that is written to /lib/libcpfwcommon.so. It can be injected into selected appliance processes, where it handles specially marked attacker connections while ordinary traffic continues. The combination of a persistent implant and concealed connections through appliance processes makes investigation of the gateway itself important, not just review of perimeter traffic.

Lateral movement and destructive activity

The attackers used the gogo scanner to discover internal systems and services, then moved from a compromised gateway to a VMware ESXi host over SSH using the vpxuser service account. On the host, they staged and ran the destroy_v6 wiper. It recursively deletes files under a target directory, with an option to overwrite files before deletion and a fast mode that skips overwriting.

The wiper also writes a politically themed DREAM_ANNEXATION.txt note into processed directories. Check Point Research reports that the first wave included destructive activity against some organizations, while it found no evidence of that behavior in the second wave.

The researchers connect TSGA and TSProxy to the PhiliKit toolset through code and configuration similarities. They also note shared configuration details between TSProxy and a PhiliKit component, including a TLS certificate and private key. That technical overlap is part of the malware analysis; the report does not identify these as eMudhra certificates.

Attribution remains uncertain

The report says PhiliKit has previously been associated with China-nexus group UNC5221, and notes infrastructure and working-hour patterns that may be consistent with Chinese-nexus activity. But the researchers do not assess attribution with high confidence. The destructive activity and its political framing do not neatly align with known motives, leaving several possibilities open: one actor may have conducted the entire campaign, an actor may have provided initial access, or the activity may have involved a false flag. The destructive stage could also have served as a smoke screen.

That uncertainty matters operationally: defenders should prioritize observed behavior and indicators over assumptions about who is responsible.

Useful open-source tools

Teams coordinating indicators across investigations can consider MISP, an active open-source platform for collecting, analyzing and sharing structured threat intelligence. OSRepos lists it at 6.7k stars.

For local investigation of gateway and system logs, Tailpipe provides SQL-based log analytics powered by DuckDB. OSRepos lists the project as active, with 586 stars. Neither tool is presented as a detector for this campaign; they can support broader intelligence and log-analysis workflows.

What to watch

For teams responsible for Check Point gateways, review patch status for the vulnerabilities identified in the report, investigate unexpected PHP session artifacts and portal template files, and look for the named implants and related lateral movement. Also examine gateway persistence and access to internal virtualization hosts. Treat attribution as unresolved while preserving evidence that can distinguish exploitation, persistence and destructive actions.

Sources

Projects in this post

Found this useful?

Share it with someone who follows open source.

Written with the help of AI from the sources listed above and OSRepos data, and reviewed before publishing. How we measure projects.