In a rapid escalation of cyber activity, threat actors have begun aggressively probing for two critical vulnerabilities in the PHP engine, collectively dubbed "PHP2Shell," mere days after emergency patches were released. While the security community has observed widespread attempts to weaponize these flaws—most notably to propagate the Mirai botnet—there remains no confirmed evidence of successful, large-scale compromises in the wild. Despite this lack of a "smoking gun," the severity of the vulnerabilities and the speed at which attackers have pivoted to exploit them have triggered a wave of defensive, preemptive shutdowns across government, education, and healthcare sectors in Europe and beyond.
The vulnerabilities, tracked as CVE-2026-109618 and CVE-2026-109619, are critical flaws in the PHP language engine that permit unauthenticated remote code execution (RCE). With a CVSS v4.0 score of 9.2, they represent an existential risk to any server running vulnerable versions of the software. The issues were originally identified in August by researchers at Wiz, OpenAI, and Calif. Following a leak of the CVE identifiers, the PHP maintainers moved to accelerate their release schedule, shipping fixes in versions 8.5.12, 8.4.27, 8.3.36, and 8.2.35 on October 21—one day ahead of their original plan.
However, the threat environment had already begun to shift before the patches were even public. Amazon reported observing the first exploitation attempts as early as October 17, and the volume of hostile traffic surged dramatically once the official patches were distributed. In response, organizations globally opted for extreme caution; in the UK alone, 32 government services were taken offline or reported as unavailable, and four others were partially restricted, as agencies prioritized patching and system hardening over continuous uptime.
Mirai Botnet Linked to Exploitation Attempts
The race to weaponize PHP2Shell has caught the attention of major cybersecurity firms, most notably Sophos, which has documented attempts by threat actors to leverage the vulnerabilities to install "Potassium," a sophisticated, updated variant of the infamous Mirai botnet. While the initial reports suggested a successful compromise, subsequent analysis by Sophos corrected this, noting that the specific attack they intercepted was thwarted before the malware could establish a foothold on the target system.
The Potassium strain identified by researchers is a refined iteration of Mirai, characterized by enhanced functionality, including the ability to download secondary payloads and coordinate Distributed Denial of Service (DDoS) attacks. Other firms, including Arctic Wolf, have identified binary files related to Potassium hosted on the same infrastructure used for PHP2Shell payloads. Yet, like Sophos, Arctic Wolf’s telemetry has not confirmed that the PHP2Shell exploit chain was successfully used to fetch or execute these malicious binaries.
Further investigation by the security firm Hunt.io suggests that these efforts are being spearheaded by an established Mirai-style botnet that has been active since at least May 2026. This group appears to have simply added PHP2Shell to its existing arsenal of propagation methods. Sophos researchers have noted that the lack of confirmed breaches is not necessarily due to a lack of intent, but rather the inherent difficulty of weaponizing these specific bugs. Even with the assistance of automated tools and generative AI, the technical barriers to crafting a reliable, universal exploit for PHP2Shell remain high.
Attempts, but No Confirmed Breaches
The consensus among industry experts currently points toward a "loud but ineffective" campaign. Firms such as Wordfence and Hunt.io have reported persistent, high-volume exploitation attempts, yet they have observed no successful, verified compromises as a result of these efforts.
This reality stands in stark contrast to earlier, alarmist predictions. Sansec, which had initially forecasted mass exploitation within minutes of the patch release, significantly revised its outlook on October 22. The firm clarified that the vulnerabilities are highly environment-specific; they require an attacker to possess granular knowledge of the target server’s configuration and the presence of specific, non-default conditions. Publicly available exploits, according to Sansec, consistently fail against PHP installations that utilize standard production settings.
Exploit Defeats ASLR in Production Setups
The technical complexity of PHP2Shell is largely attributed to Address Space Layout Randomization (ASLR), a security feature that randomizes the memory addresses used by system processes, making it significantly harder for attackers to predict where to inject code. However, the barrier is not insurmountable for advanced threat actors. The firm VulnCheck has publicly disclosed that it successfully engineered an exploit for CVE-2026-109618 that bypasses ASLR by leveraging a memory leak.
This exploit is effective on x86-64 systems where ASLR is enabled and error display is disabled—the latter being the recommended configuration for secure production environments. While VulnCheck has restricted access to this exploit, providing it only to its Initial Access Intelligence customers, the firm emphasized that it is not currently aware of other public exploits that achieve RCE in such hardened configurations. Nonetheless, they warned that as adversaries continue their development efforts, the sophistication of these attacks will likely increase.
Additional evidence of this development work was uncovered by SOCRadar, which discovered an attacker-controlled server containing a suite of "one-click" exploitation tools. The operator had created scripts designed to fingerprint websites and test them against the specific configuration requirements of the exploit. SOCRadar monitored the operator’s progress as they probed 11 websites across various regions, including the US, UK, Canada, and China. Despite the operator’s evident technical skill, none of the probed sites met the specific criteria required for the exploit to function. Hunt.io conducted a parallel analysis of the same server, tracking 10 attack rounds targeting 123 different servers. Of the 57 servers identified as running PHP, none were successfully compromised, further underscoring the gap between the attackers’ ambitions and the reality of the software’s resilience.
Patching Lags Behind
While the technical difficulty of the exploit provides a temporary buffer, the sheer scale of unpatched software remains a primary concern for the security community. Data provided by Qualys from a sample of 12,000 customers revealed that only 199,000 PHP instances were patched within the first 12 hours of the vulnerability disclosure.
The scope of the problem is massive: an estimated 110 million PHP instances are currently running end-of-life versions (between 5.4 and 8.1), which will never receive official vendor patches. Additionally, roughly 44 million instances are running supported branches where patches are available but remain unapplied. The cybersecurity firm Censys is currently tracking this landscape in real-time, maintaining a public dashboard that monitors the number of internet-facing services still running vulnerable versions of PHP.
As of the morning of October 23, the Cybersecurity and Infrastructure Security Agency (CISA) had yet to add either of the PHP2Shell vulnerabilities to its Known Exploited Vulnerabilities (KEV) catalog. The security community remains on high alert, anticipating that as attackers refine their tooling to account for production-grade hardening, the risk of a successful breach could rise. For now, the combination of cautious system administration, the inherent technical barriers of the vulnerabilities, and the difficulty of weaponizing the flaws has kept the internet’s infrastructure largely safe from the worst-case scenario. However, with millions of servers running outdated code, the window for remediation is closing, and the threat of exploitation continues to evolve.

