ThreatBeaconXSubscribe
CriticalThreat Intelligence · 8 min read

PHP2Shell Exploitation Confirmed in the Wild as Attackers Probe Global Infrastructure

Threat activity surrounding PHP2Shell has escalated following the disclosure of two critical PHP vulnerabilities, CVE-2026-109618 and CVE-2026-109619. Security researchers have recorded successful exploitation in honeypot environments, with attackers executing shell commands and attempting to retrieve additional payloads. Separately, CYFIRMA reported a suspected North Korea-linked campaign scanning vulnerable PHP infrastructure across multiple countries. The developments highlight the growing risk to internet-facing PHP systems and the need for organizations to apply security updates, verify mitigations, and monitor suspicious server activity.

Written by Jakka Lakshmi Koumudi·Published Oct 27, 2026


1. Description

PHP2Shell refers to exploitation activity associated with two critical vulnerabilities in the PHP runtime that can enable unauthenticated remote code execution under affected conditions.

The vulnerabilities, tracked as CVE-2026-109618 and CVE-2026-109619, affect PHP's memory-management and request-handling behavior. Their impact is particularly concerning because the weaknesses exist within the PHP runtime rather than a specific content management system, plugin, or web application.

In an October 25, 2026, report, The Hacker News described multiple developments surrounding PHP2Shell, including successful exploitation observed against honeypot servers, scanning of vulnerable infrastructure, and attempts to retrieve additional payloads.

The report also discussed a suspected North Korea-linked campaign targeting selected vulnerable hosts across government, financial, logistics, aviation, manufacturing, and online transaction sectors. CYFIRMA attributed this activity to a Korean-speaking operator associated with the Lazarus Group ecosystem with moderate confidence.

Importantly, Unit 42 subsequently retracted its earlier assessment that an attack against a European government agency involved exploitation of PHP2Shell. It stated that the initial access method could not be confirmed as involving the specific vulnerabilities. This distinction is essential when assessing the scope of confirmed exploitation.

2. Technical Description

PHP2Shell involves two vulnerabilities in the PHP runtime that can potentially be chained to achieve remote code execution.

  • CVE-2026-109618: Associated with an out-of-memory corruption vulnerability that can enable unauthenticated remote code execution.

  • CVE-2026-109619: Associated with stale internal request-state handling that can also enable unauthenticated remote code execution.

Both vulnerabilities were reported with a CVSS score of 9.2. Exploitation depends on factors such as the PHP version, server configuration, request handling, and the ability to satisfy the exploit's technical prerequisites.

Research described by The Hacker News indicates that exploitation can involve specially crafted HTTP requests and PHP session-upload-progress behavior. Certain exploit chains also attempt to obtain memory information to overcome Address Space Layout Randomization (ASLR), a protection that makes it more difficult to predict where code and data reside in memory.

Because the weaknesses are located in the PHP runtime, changing application code alone is not sufficient to remediate affected installations. Organizations must update the underlying PHP deployment to a fixed version.

3. Attack Overview

The reported activity can be summarized in the following stages:

  1. Target discovery: Threat actors identify or scan internet-facing systems running potentially vulnerable PHP deployments.

  2. Exploit probing: Attackers send crafted HTTP requests to test whether the target exhibits vulnerable behavior.

  3. Exploit execution: Under suitable conditions, an attacker may trigger memory corruption or abuse request-state handling to obtain code execution.

  4. Command execution: Successful exploitation can allow commands to execute with the privileges of the affected PHP process.

  5. Payload retrieval: Observed honeypot activity included commands using tools such as curl, wget, and the PHP CLI to contact remote infrastructure.

  6. Potential follow-on activity: Depending on the attacker's objectives, access could be used for reconnaissance, credential collection, persistence, or delivery of additional malware.

These stages describe the general technique and reported observations. Not every stage was confirmed in every investigated activity, and the presence of an exploit probe does not establish successful compromise.

4. Technical Analysis

4.1 Confirmed exploitation in honeypot environments

The Hacker News reported that researcher Adel Karimi observed what appeared to be successful in-the-wild exploitation of PHP2Shell against honeypot servers.

The captured shell commands attempted to contact a remote server at 162.35.99[.]94 using curl, wget, or PHP's file_get_contents() function. The requests included a target-specific path, suggesting an attempt to identify or distinguish individual honeypot systems.

The reporting did not establish that a second-stage payload had been successfully downloaded. Therefore, the evidence supports successful command execution in the honeypot environment, but not a confirmed subsequent malware infection.

4.2 Suspected North Korea-linked scanning campaign

CYFIRMA reported scanning activity involving 21 compromised nodes and a selected set of 216 unpatched or partially mitigated hosts across multiple countries.

The reported targets included government organizations, financial institutions, logistics providers, aviation-related entities, manufacturers, and online transaction platforms.

According to the report, the scanning was deliberately paced, with infrastructure rotated or abandoned when blocking or rate-limiting was encountered. This approach can reduce the effectiveness of detection systems that rely exclusively on high request volumes or fixed thresholds.

CYFIRMA assessed the activity as potentially linked to the Lazarus Group ecosystem with moderate confidence. This attribution remains an intelligence assessment, not definitive confirmation of the operators' identity.

The scanning activity should also be distinguished from confirmed exploitation: identifying a vulnerable host does not prove that code execution was achieved on that host.

4.3 AI-assisted post-exploitation reporting and attribution correction

The same report initially described an attack against a European government entity in which attackers reportedly accessed password hashes and multi-factor authentication secrets, conducted Active Directory reconnaissance, and attempted privilege escalation.

However, Unit 42 later clarified that it could not confirm that the initial access was associated with CVE-2026-109618 or CVE-2026-109619.

Consequently, these post-compromise observations should not be presented as confirmed PHP2Shell exploitation. They demonstrate the potential consequences of a compromised environment, but the specific initial access technique remains unverified.

4.4 Detection challenges

PHP2Shell presents several challenges for defenders:

  • Exploitation may occur within the PHP runtime before application-level logging captures meaningful activity.

  • Memory exhaustion and worker-process crashes can resemble ordinary application instability.

  • Initial scanning may be distributed across multiple IP addresses and deliberately paced.

  • Generic HTTP status codes alone may not reliably distinguish malicious requests from legitimate traffic.

  • Successful code execution can provide access using the privileges already assigned to the web application.

Defenders should therefore correlate web access logs, PHP error logs, web-server error logs, system events, and endpoint telemetry rather than relying on a single alert.

5. Indicators of Compromise (IOCs)

The following indicators were reported in the source material. They are a focused selection, not a complete list of every indicator associated with PHP2Shell.

Vulnerability identifiers

  • CVE-2026-109618

  • CVE-2026-109619

Reported remote infrastructure

  • 162.35.99[.]94 — remote host contacted by commands observed during honeypot exploitation.

Reported command patterns

  • Use of curl, wget, or PHP CLI to request a target-specific path from the remote host.

  • HTTP POST requests involving PHP session-upload-progress behavior.

  • Sequences of PHP memory-exhaustion errors, Apache errors, and worker-process crashes that may help identify exploitation attempts.

Important validation note: The remote IP address is associated with reported exploitation activity, but an IP address alone is not proof of malicious activity in every context. Validate it against current threat intelligence and local telemetry before blocking it.

6. Potential Impact

Successful exploitation of PHP2Shell could expose affected systems to significant security risks:

  • Remote code execution: Attackers may execute commands with the privileges of the vulnerable PHP process.

  • Sensitive data exposure: Application-accessible credentials, API keys, database connections, and internal service information may be exposed.

  • Server compromise: Attackers may use access to conduct reconnaissance, deploy additional tools, or establish persistence.

  • Lateral movement: A compromised web server may provide a foothold for attempting access to internal systems, depending on network segmentation and available privileges.

  • Service disruption: Exploitation attempts or resulting worker crashes may affect application availability.

  • Supply-chain and business exposure: Organizations operating shared web platforms or internet-facing services may face operational and downstream risks if affected systems are compromised.

These are potential consequences of successful exploitation, not evidence that every targeted system experienced these outcomes.

7. Recommendations

Immediate Actions

  • Patch affected PHP deployments: Upgrade to a fixed release appropriate for the installed branch. The October 21, 2026, emergency releases listed by PHP include 8.5.12, 8.4.27, 8.3.36, and 8.2.35.

  • Inventory exposed systems: Identify internet-facing PHP installations, including legacy applications, containers, virtual machines, and systems managed by third parties.

  • Investigate suspicious requests: Search for unusual POST requests, PHP session-upload-progress markers, memory-exhaustion errors, and abnormal web-server worker crashes.

  • Review outbound connections: Investigate unexpected connections to remote hosts associated with exploitation attempts, especially when correlated with suspicious server-side commands.

  • Check for post-exploitation activity: Review process creation, unexpected scripts or files, new accounts, suspicious scheduled tasks, unusual database access, and attempts to access secrets.

  • Escalate suspected compromises: Preserve logs and forensic evidence. If code execution is confirmed, treat the server as potentially compromised rather than assuming that patching alone restores trust.

Preventive Measures

  • Maintain supported PHP versions: Establish a lifecycle process for upgrades and retire unsupported PHP branches wherever possible.

  • Harden web-server configurations: Apply vendor-recommended PHP and web-server settings, and minimize unnecessary runtime features.

  • Use layered monitoring: Correlate web-server logs, PHP error logs, network telemetry, endpoint detection, and cloud audit records.

  • Apply least privilege: Run web applications with restricted operating-system permissions and limit access to credentials, internal services, and sensitive files.

  • Segment internet-facing services: Prevent web servers from having unrestricted access to internal administrative systems and databases.

  • Protect secrets: Use managed secret storage where possible, rotate credentials exposed during a confirmed compromise, and audit access to API keys and database credentials.

  • Validate remediation: Confirm that systems are running fixed versions and perform follow-up checks for suspicious activity that may have occurred before patching.

A web application firewall may provide supplementary protection, but it should not be treated as a replacement for updating the PHP runtime.

8. Conclusion

PHP2Shell illustrates how vulnerabilities in a widely deployed server-side runtime can create exposure across many different applications and industries. Recent reporting has moved beyond generic scanning observations, with researchers recording successful command execution in honeypot environments and identifying additional attempts to probe vulnerable infrastructure.

At the same time, defenders must distinguish confirmed exploitation from suspected campaigns and unverified attribution. The reported North Korea-linked scanning activity warrants attention, but it does not establish that every selected target was compromised. Likewise, Unit 42's correction means that the European government incident should not be described as confirmed PHP2Shell exploitation.

The immediate priority for organizations is to patch affected PHP installations, validate remediation, and investigate evidence of suspicious server-side execution. Rapid patching combined with layered monitoring and effective incident response remains the strongest defense against this class of vulnerability.

MITRE ATT&CK Mapping

T1190 – Exploit Public-Facing Application