A sophisticated China-linked threat actor has been identified conducting a targeted campaign against internet-facing firewalls, leveraging a lethal combination of newly disclosed PHP vulnerabilities and a critical Linux kernel flaw. Cybersecurity firm Dream recently released an in-depth analysis detailing how these attackers are chaining the "PHP2Shell" vulnerability set with a kernel privilege escalation bug known as "Fragnesia" to gain root-level access to network infrastructure and maintain persistent, long-term access via webshells.
The operation, which gained significant traction in late October, demonstrates a high level of operational discipline and strategic planning. By utilizing automated reconnaissance and precise filtering techniques, the threat actor managed to identify and compromise a specific vendor’s firewall appliances, ultimately gaining a foothold in sensitive environments, including those affiliated with the United States government.
The Anatomy of the Attack Chain
The exploit chain centers on two vulnerabilities in the PHP engine, tracked as CVE-2026-109618 and CVE-2026-109619. When exploited individually, these flaws present significant risks; however, when combined, they provide a powerful mechanism for remote code execution. To escalate their privileges to the highest possible level—root—the attackers integrated a third vulnerability: CVE-2026-46300, a flaw dubbed "Fragnesia." This Linux kernel vulnerability acts as the final key, allowing the attackers to bypass standard security restrictions and execute commands with unrestricted system-level permissions.
The attackers’ methodology was highly methodical. Between October 23 and October 24, they utilized the Chinese internet asset search engine FOFA to scan for internet-facing firewalls manufactured by a specific vendor. Their initial reconnaissance identified approximately 19,000 devices globally that were exposed to the public internet.
Rather than launching a broad, indiscriminate attack that might alert network defenders, the threat actors employed a rigorous filtering process. They narrowed their scope to firewalls located in the United States, Japan, Taiwan, and South Korea—nations that represent primary strategic and intelligence priorities for Chinese state-aligned cyber operations. Further refining their target list, they filtered out devices running older, incompatible software, leaving them with a high-value pool of 240 potential targets that were guaranteed to be vulnerable to their specific exploit chain.
Persistent Access and Escalation
Once the target list was finalized, the attackers deployed an automated tool to execute the PHP2Shell-Fragnesia chain against the selected firewalls. The result was the successful deployment of a PHP-based root webshell on 24 distinct devices. This webshell provided the actors with a permanent "backdoor," allowing them to interact with the compromised systems long after the initial exploit had been executed.
While 22 of the compromised firewalls showed no immediate signs of post-exploitation activity, indicating that the attackers were likely "seeding" their presence for future use, the remaining two servers—which were part of a US government-affiliated environment—saw aggressive follow-up maneuvers.
The attackers’ tenacity was particularly evident during the compromise of these government-linked assets. One of the targeted firewalls crashed shortly after the initial exploitation on October 23. Undeterred, the attackers monitored the device for two hours using 240 automated probes, waiting for the system to reboot and come back online. When the device failed to recover, they shifted their focus to a second firewall that had timed out during the initial batch. On October 24, the attackers successfully exploited this second device, establishing a persistent foothold.
Deep Internal Penetration
Having established a beachhead on the government-affiliated network, the threat actors did not simply stop at the firewall. They utilized the compromised gateway to launch a reverse SSH tunnel, creating a secure bridge between their external infrastructure and the target’s internal network. This tunnel was used to pivot directly to an internal firewall management server.
By applying the same PHP2Shell-Fragnesia exploit chain to this management server, the attackers gained administrative control over the entire network ecosystem. This level of access allowed them to harvest configurations, system credentials, and security settings for every single firewall managed by that organization.
The scope of the subsequent activity highlights the gravity of the breach. The attackers accessed and decrypted LDAP service account passwords, which allowed them to enumerate active users and groups within the domain. With this information in hand, they launched password-spraying attacks to broaden their reach and attempted to create unauthorized Active Directory backdoor accounts. By creating these "shadow" accounts, the threat actors ensured that even if the primary vulnerability was patched or the webshells were removed, they would retain the ability to move through the network at will.
Strategic Implications and Attribution
While Dream has not formally attributed this campaign to a specific state-sponsored group, the nature of the targets—coupled with the geographic focus on the Pacific theater—points toward a well-resourced, nation-state-aligned actor. The deliberate selection of Japan, South Korea, Taiwan, and the United States aligns with the established intelligence requirements of Chinese state-aligned cyber operations.
"Since these nations are the four primary strategic adversaries and intelligence targets for Chinese state-aligned cyber operations in the Pacific theater, this indicates potential nation-state adversary efforts," Dream noted in their post-incident analysis. The use of advanced kernel-level exploits and the patient, multi-stage approach to lateral movement within a sensitive environment suggest a level of sophistication typically reserved for state-backed entities rather than common cybercriminal syndicates.
This campaign serves as a stark reminder of the risks associated with internet-facing management interfaces. By exploiting vulnerabilities that bridge the gap between web-facing PHP applications and the underlying Linux operating system, attackers are effectively turning the very tools designed to protect organizations—firewalls—into the primary entry points for network-wide compromise.
As security teams globally grapple with the implications of the PHP2Shell and Fragnesia flaws, the incident underscores the urgent need for robust patch management and, perhaps more importantly, the segregation of management interfaces from the public internet. As this investigation continues to evolve, organizations are encouraged to audit their firewall configurations and monitor for unauthorized lateral movement, particularly for accounts associated with domain administration and LDAP services.
The persistence displayed by these actors—waiting hours for a crashed system to recover and methodically pivoting through internal management servers—highlights a threat that is not only capable of technical exploitation but also exhibits the patience and strategic intent characteristic of modern, high-stakes cyber espionage. For many, this event is a warning that even well-defended perimeter devices can be rendered ineffective when faced with a coordinated, multi-vulnerability exploit chain.

