PHP2Shell is the name given to two critical PHP vulnerabilities, CVE-2026-109618 and CVE-2026-109619. Each is rated CVSS 9.2 and can allow an unauthenticated attacker to execute code on a PHP web server. PHP published fixed releases on October 21, 2026, after reports of an exploit attempt dated four days earlier raised the urgency of the response.
4
5
24
What the PHP2Shell flaws do
The vulnerabilities involve memory-handling errors in PHP’s request processing. PHP describes CVE-2026-109618 as an out-of-memory corruption issue during request cleanup, and CVE-2026-109619 as stale request-owned state being reused by a later request handled by the same worker. Both can lead to remote code execution.
4
10
The risk applies to PHP installations serving web requests, including deployments using PHP-FPM or Apache with mod_php, as well as other listed web-serving configurations.
4
5 The PHP security team has said the risk is highest for mod_php.
7
Fixed PHP versions
PHP released these patched versions on October 21, 2026:
4
- PHP 8.5.12
- PHP 8.4.27
- PHP 8.3.36
- PHP 8.2.35
Administrators should update to the fixed release for their PHP branch. Because the flaws affect web-request-serving installations, review every deployment that processes web traffic rather than checking only the application’s framework or CMS.
4
5
Reports said premature disclosure brought the fixes forward from their originally scheduled October 22 release. The available reporting does not establish precisely what information leaked, who disclosed it, or how it circulated.
3
20
What is known about attack activity
Amazon Threat Intelligence reported an exploit attempt against its MadPot honeypot, dated October 17—four days before the public disclosure and patch release. Amazon said the attempt came from an IP address associated with a commercial anonymization network, but that it remained unclear how the actor obtained information about the vulnerabilities in advance.
24
Cloudflare reported that exploit attempts rose into the millions after disclosure. These are counts of attempted activity, not a tally of confirmed compromised websites.
18
19 Likewise, an attempt recorded by a honeypot is evidence of targeting, not proof that an attacker successfully compromised a production server. The available reporting does not establish how many production systems, if any, were breached.
24
46
What administrators should do
- Install the fixed PHP release for each affected branch in use.
4
- Review activity from October 17 onward. Amazon’s early honeypot detection makes patch day an incomplete starting point for retrospective checks.
24
46
- Treat attempted exploitation and confirmed compromise as different findings. Attempt counts can show that attackers are probing a flaw, but they do not establish successful code execution on a particular server.
18
24
The clearest response is to patch affected PHP installations and investigate relevant activity dating back to October 17. High exploit-attempt volumes underscore the urgency, but they should not be mistaken for evidence that every targeted server was compromised.