KEY TAKEAWAYS:
- PHP2Shell is the name given to two critical, unauthenticated remote code execution flaws in PHP, tracked as CVE-2026-109618 and CVE-2026-109619, each scoring 9.2 out of 10 on CVSS and affecting PHP versions 5.4 to 8.5.
- Coordinated disclosure was planned for 22 October 2026, but details leaked online on 20 October, and fixes were released in PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35; PHP 8.1 and older no longer receive official fixes.
- Exploitation in the wild is confirmed: CISA added both flaws to its Known Exploited Vulnerabilities catalogue, and a security firm reported root access on 24 firewalls in the United States, Japan, Taiwan and South Korea using an exploit chained with a Linux kernel flaw.
- Beyond those cases, the report says mass compromise of real-world organisations has not been publicly confirmed, and that attack attempts and precautionary shutdowns should not be read as proof of successful breaches.
- Around a week after disclosure, Censys estimated 30–32.5 million PHP services still ran vulnerable versions.
- MitKat rates Europe and APAC as Elevated and the Americas as Moderate, and stresses that patching does not undo credential theft that happened before the update.
Governments and companies around the world spent late October taking web services offline or racing to patch PHP. This summary draws on MitKat’s special report dated 03 November 2026. The situation is still developing, so check the current vendor and national-agency advisories before acting, and treat the report’s figures and assessments as attributed to MitKat and its cited sources.
What is PHP2Shell and why is it serious?
PHP2Shell is the name given to a pair of critical flaws in PHP, the programming language that, according to the report, sits behind about 70 percent of web servers and powers WordPress, Drupal, Joomla and Magento. Tracked as CVE-2026-109618 and CVE-2026-109619, both score 9.2 out of 10 for severity. They let an attacker send a specially crafted request over the internet and run commands on a vulnerable server without a password, and may also expose server memory that holds passwords and access keys. The report offers an analogy: a website is the car body and PHP is the engine, so repainting the car, meaning updating the site or its plugins, does not fix an engine fault. The PHP runtime itself has to be updated. Any website server running PHP versions 5.4 to 8.5 that has not been updated is exposed, and the report advises that firewall rules alone are not a complete fix, a caution it attributes to the US National Security Agency.
How did the PHP2Shell flaws come to light?
According to the report, researchers using advanced AI models reviewed PHP’s source code in early August 2026 and found memory-handling faults that had gone unnoticed for more than a decade. They reported them privately to the PHP project, and coordinated disclosure was scheduled for 22 October 2026. Details leaked online on 20 October, ahead of that date, and fixes followed in PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35. PHP 8.1 and older no longer receive official fixes, so organisations on those branches need to move to a supported branch or apply a verified vendor or long-term-support backport. The report also notes that a second team independently rediscovered the issue with a basic prompt, and draws the wider point that AI is shortening the time between discovery and exploitation for both defenders and attackers.
What is confirmed, and what is not?
The report is careful to separate verified facts from worries. Confirmed: attack attempts rose from thousands to millions within hours of disclosure; the US Cybersecurity and Infrastructure Security Agency (CISA) added both flaws to its Known Exploited Vulnerabilities catalogue by 26 October; a researcher recorded 35 successful command executions on three decoy systems running PHP 8.4.26, while cautioning that this did not show reliable exploitation on normal production systems; and between 23 and 24 October a security firm reported root access on 24 firewalls in the United States, Japan, Taiwan and South Korea, including three systems in a US government-affiliated network, using a custom exploit chained with a Linux kernel flaw (CVE-2026-46300). A firewall vendor reportedly confirmed active exploitation of its management interfaces, with fixes now available. Not confirmed: mass compromise of real-world organisations beyond those cases, and a suspected DPRK-linked staging campaign. Several early claims, including a claimed European government-agency intrusion, were later withdrawn. Delft University’s Threat Intelligence Lab had logged more than 73,000 exploitation requests from 668 source IPs as of 29 October, but these are attempts seen by sensors, not breaches.
Which regions and sectors are most exposed?
On targeting rates, the report says Europe recorded the highest share of organisations targeted since disclosure, at 14 percent against a global average of 10 percent, with Asia-Pacific and the Americas at 6 percent each; these count attempts, not successful break-ins. MitKat rates Europe and APAC as Elevated and the Americas as Moderate, and cautions that lower reported activity does not mean lower vulnerability. The most visible impact has been precautionary shutdowns: Dutch national government portals, 32 UK central-government digital services, Germany’s federal IT service provider and Australia’s cyber agency website went offline while patches were applied, and CERT-In in India rated the risk critical, with no Indian compromises officially reported as of writing. By sector, MitKat rates government and public services, e-commerce and digital media, and technology and cloud infrastructure as Elevated, and healthcare, education and research, and financial services as Medium.
What are the business and individual consequences?
For organisations, the report lists undetected data exposure, financial loss through stolen session credentials and API keys, repeat compromise where only application code was patched and not the PHP runtime, regulatory exposure under reporting and data-protection regimes such as CERT-In guidelines, India’s DPDP regime and GDPR, and reputational harm from outages or disclosures. It also flags that insurance coverage for voluntary precautionary outages may be unclear. For individuals, the main risk is indirect: where records or session tokens were accessed, people face a higher chance of targeted phishing, account takeover and credential-stuffing, and they have little visibility into whether their data was exposed. Updating PHP closes the flaw but does not revoke credentials that were already taken, so the report treats credential rotation and log review as part of the same fix, not optional follow-up.
What should organisations do now?
The report’s immediate actions are to apply vendor-approved PHP patches without delay and verify them with follow-up scans, and where patches are not yet available to apply the recommended mitigations and restrict unnecessary services. Organisations should isolate confirmed or suspected compromised systems, preserve forensic evidence and restore services only after validation. They should monitor web server logs and network traffic for suspicious requests and unauthorised command execution, audit third-party PHP libraries, frameworks and plugins, review and rotate administrator, service and API credentials, invalidate active sessions, and enforce multi-factor authentication and least privilege. Servers that were internet-facing during the exposure window should be assessed for earlier access, especially where logging is limited. On a separate point, PHP issued an informational advisory on 23 October about a distinct issue, CVE-2026-115560, and recommends setting session.upload_progress.enabled=0 in relevant configurations until its next regular release. Longer term, the report recommends assessing hosting providers and third-party products that bundle PHP, replacing unsupported systems, preparing stakeholder communication and regulatory notification templates in advance, and testing continuity and incident-response plans.
Frequently asked questions
Which PHP versions fix PHP2Shell?
The report lists PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35 as the fixed versions. PHP 8.1 and older no longer receive official fixes, so those systems need to be upgraded to a supported branch or covered by a verified vendor or long-term-support backport.
Does updating my website or plugins fix it?
No. The flaw is in PHP itself, so the PHP runtime on the server has to be updated. Security and network products that have PHP built in must wait for their supplier’s fix and should be isolated or closely watched until then.
Does patching mean my data is safe?
Not necessarily. Patching closes the weakness but does not reverse anything taken beforehand, so organisations exposed before patching should review logs and rotate credentials and keys.
Are attack attempts the same as breaches?
No. The report stresses that exploitation attempts and precautionary outages are not evidence of successful compromise, and that the scale of breaches and losses remains uncertain.
Download the full report now
MitKat’s Special Report, PHP2Shell: Global Cyber Security Threat Outlook, carries the full incident timeline, the country-by-country government response and impact tables for APAC, EMEA and the Americas, the sector exposure analysis, and the complete recommendations. Download the full report now to brief your security and continuity teams.
Filed under: Threat Intelligence Special Reports