php2shell: Detecting and Investigating CVE-2026-109618 and CVE-2026-109619

php2shell: Detecting and Investigating CVE-2026-109618 and CVE-2026-109619

Ajmal Kohgadai
Ajmal Kohgadai
October 27, 2026

Updated October 27, 2026. CISA added php2shell, the PHP flaws tracked as CVE-2026-109618 and CVE-2026-109619, to its Known Exploited Vulnerabilities catalog on October 26. Federal civilian agencies have until October 29 to patch. The listing followed public confirmation of code execution on honeypots on October 24. This update adds what that confirmed exploitation looked like and the first published indicators. It also covers the earlier start date Amazon reported, a third PHP flaw (CVE-2026-115560), and the vendor appliances that ship PHP, including Citrix NetScaler.

PHP shipped emergency fixes for php2shell on October 21. Exploitation attempts started earlier: Amazon Threat Intelligence traced the earliest attempt it has seen to October 17, four days before disclosure. Cloudflare reported attacks against the flaws before the patches shipped, and watchTowr reported attempts in the wild from its honeypot network shortly after. Prophet Security has since observed activity consistent with php2shell exploitation, including command execution through a PHP worker. A host stays exposed until it is updated and its PHP workers have restarted.

The question security leaders ask after a disclosure like this one is whether anyone got in before the patch, and how soon the team can answer that. Answering it takes an accurate list of where PHP runs and a way to check each of those hosts for signs that someone exploited either flaw. This post covers both, using the directed approach to proactive threat hunting that starts from a named threat and the assets it can reach.

What php2shell is and which PHP builds are affected

php2shell is the name researchers gave CVE-2026-109618 and CVE-2026-109619, two critical flaws in the PHP interpreter that can let an unauthenticated web request run code as the web worker account. The stale-state flaw can also disclose live worker memory, which may hold passwords, tokens, and keys, without executing code. PHP fixed both on October 21, 2026, in versions 8.5.12, 8.4.27, 8.3.36, and 8.2.35.

Details leaked on October 20, and the PHP security team scheduled its release for 14:00 UTC on October 21, a day ahead of the original embargo. The team rated both flaws CVSS v4.0 9.2. Its security advisory says they affect all PHP installations serving web requests, including PHP-FPM, Apache with mod_php, php-cgi, FrankenPHP, and LiteSpeed. The team considers the risk highest for mod_php, where a demonstrated exploit triggers from an unauthenticated request before any application code runs. CISA published an alert the same day calling for immediate patching.

The two CVEs reach code execution through different memory-safety failures:

  • CVE-2026-109618, "RCE via OOM corruption." Out-of-memory handling can interrupt PHP while request data is only partly initialized, and request cleanup can then treat the invalid data as valid. The advisory lists multipart request parsing among the triggers, including requests without a usable Content-Length, parts with unbounded header counts, and deeply nested field names.
  • CVE-2026-109619, "RCE via stale global state." State owned by one request can survive after its memory is freed, and a later request handled by the same worker can use it. The confirmed paths run through session upload progress handling and, on PHP 8.5, fatal error backtraces. Session upload progress is enabled by default and is reachable even if the application never calls session_start().

Which application sits on top of PHP makes little difference. Wiz, one of the research teams credited with reporting the flaws, demonstrated unauthenticated code execution against current WordPress, Drupal, and Joomla container images. Wiz also reported that exploitation does not depend on a flaw in the application's code. By its count, every PHP version released in the last decade is affected by at least one of the two flaws. Pre-auth code execution on a web server commonly ends with a web shell for persistent access, the pattern our web shell detection guide covers in depth.

PHP 8.1 and older no longer get fixes from the PHP project. The PHP team published minimal source backports for 7.4, 8.0, and 8.1 for anyone building from source, and some Linux distributions and LTS vendors ship backported packages.

A third flaw, CVE-2026-115560, also runs through session upload progress handling, and the October 21 releases do not fix it. The PHP security team disclosed it on October 23 in an informational advisory, which says it can only occur under an unusual combination of configuration and application code. All three of these conditions have to hold:

  • The server configuration sets session.serialize_handler to php_binary or php_serialize.
  • The application switches the handler to php with ini_set() before calling session_start().
  • Session upload progress is still enabled, which is the default.

PHP plans to fix CVE-2026-115560 in its next regular release and recommends session.upload_progress.enabled=0 until then. That is the same setting listed below as an interim measure for CVE-2026-109619, and PHP says hosts that already applied it need no further action.

On Windows, PHP published rebuilt packages with a -1 suffix, such as 8.4.27-1, on October 26. The rebuilds update the bundled Brotli library for CVE-2026-111497, a separate decompression flaw also reported as HTTP2Shell or Broccoli2Shell. The PHP runtime versions inside them are unchanged, so their php2shell status is the same as the releases they rebuild.

Which of your hosts are exposed to php2shell

Scoping starts with an inventory. PHP usually runs in more places than the application owners' list suggests, including CMS and e-commerce sites, internal admin consoles, container base images, and vendor appliances that bundle their own interpreter. Prophet AI's PHP runtime inventory hunt reports each install's version, whether its branch still receives security fixes, and whether the host is reachable from the internet. The same list shortens the response to the next PHP vulnerability, because the installs, versions, and exposure are already known.

Work the list in order of exposure. Internet-facing hosts come first, and mod_php hosts come ahead of PHP-FPM, given the PHP team's risk assessment. End-of-life branches go into an upgrade plan. The exploitation case Prophet Security observed, described below, was on an internet-facing PHP 7.1 build, older than any branch the PHP team published backports for.

A version string can mislead in both directions. Distribution packages often carry the fix under the original upstream version number, so a host that reports an 8.2 build below 8.2.35 may already be patched. Check the vendor's package changelog before escalating it. An updated package on disk also leaves a long-running PHP-FPM or Apache process on the old code until that process restarts. Compare each worker's process start time with the package install time.

Vendor appliances that ship PHP get fixed when the vendor releases an update, since customers cannot upgrade the bundled interpreter themselves. Citrix and Palo Alto Networks both say the exposure on their products is confined to the management interface. Status from the larger appliance vendors as of October 27:

  • Citrix NetScaler ADC and NetScaler Gateway. The management GUI and NITRO API are affected on the NSIP, on SNIPs with management access enabled, and on the Cluster IP, in all supported versions (CTX697263). Fixed builds 14.1-73.49, 13.1-64.32, 14.1-73.49 FIPS, and 13.1-37.286 FIPS/NDcPP were released on October 27. Until a fixed build is installed, Citrix recommends limiting management access to trusted administrative hosts. NetScaler Console and SVM are not affected.
  • Palo Alto Networks PAN-OS and Panorama. The management web interface and XML API are affected, along with any dataplane interface that has a management profile attached. The GlobalProtect portal and gateway, Clientless VPN, and Captive Portal are not affected. Hotfixes are shipping by branch through October 30, and Palo Alto Networks says it is not aware of exploitation against its products (advisory).
  • Cisco. Expressway Series and Expressway Select Series are affected, with the fix planned for 15.5.21 on a date not yet set and workarounds documented in the bug. Ultra-Reliable Wireless Backhaul is also affected, and Duo Access Gateway is fixed in 3.2.0. Cisco says it is not aware of malicious use (advisory).
  • Fortinet. FortiOS is listed as not affected. FortiDDoS, FortiSOAR, and FortiSOAR PaaS are still under investigation (FG-IR-26-186).

Put any appliance management interface that is reachable from the internet at the top of the exposure list. The October 27 NetScaler builds also include the fixes for the September zero-days covered in our Citrix NetScaler zero-day post, so one upgrade closes both.

Two configuration details affect exposure to CVE-2026-109619 on specific SAPIs. On mod_php, the advisory identifies php_value and php_flag settings in .htaccess or directory configuration as one route to the stale state, because Apache applies them before request startup. On PHP-FPM, the advisory treats setups where the web server passes PHP_VALUE or PHP_ADMIN_VALUE to the pool as exposed. Both are quick searches across web server configuration.

If a host cannot be patched today, the PHP team lists interim settings that reduce exposure to specific attack paths, and it warns that none of them makes an installation safe:

  • session.upload_progress.enabled=0, set in php.ini or the pool configuration rather than through ini_set() or .user.ini, followed by a restart of every worker.
  • fatal_error_backtraces=0 on PHP 8.5, also followed by a restart of every worker.
  • memory_limit=-1, enforced through php_admin_value so applications cannot override it, with a container or cgroup memory limit as the availability boundary.
  • post_max_size = 8M for applications that take no uploads, or 32M where uploads are required, with web server or proxy limits on multipart part count, header count, and nesting depth.
  • pm.max_requests=1 on PHP-FPM, if performance allows, so each worker is recycled after one request and cannot carry stale state into the next.

For applications that never accept uploads, the CVE-2026-109618 advisory also recommends blocking multipart/form-data requests at the web server or proxy.

What php2shell exploitation looked like on an IIS server

Prophet Security observed activity consistent with php2shell exploitation on an internet-facing IIS server running an end-of-life PHP 7.1 build through FastCGI. A PHP worker crashed, and Windows Error Reporting (WerFault.exe) captured the failure. About 90 minutes later, php-cgi.exe started spawning cmd.exe under the IIS anonymous identity (IUSR). That is command execution through the PHP worker, with no web shell or dropped payload observed.

The post-exploitation activity looked automated. Across four PHP workers, Prophet Security observed 73 command executions in about 135 seconds, roughly 1.8 seconds apart, with no meaningful pauses or branching. The same seven discovery commands repeated in sequence:

  • whoami
  • hostname
  • ipconfig
  • dir <web root>
  • systeminfo
  • net user
  • whoami /priv

All of the activity focused on local host and identity discovery. No domain enumeration, download tooling, or attempts to establish persistence were observed. In this case, the activity appears limited to automated host reconnaissance following command execution.

What confirmed php2shell exploitation looked like on honeypots

On October 24, Perplexity researcher Adel Karimi recorded 35 shells across three honeypots running PHP 8.4.26 under Apache mod_php, and published the evidence. His report describes the traffic as consistent with CVE-2026-109618. Each attack used a chain of four multipart POST requests, sometimes retried:

  • A request that exhausted PHP's memory limit before any application code ran.
  • A request whose response leaked worker memory, including memory addresses, through an oversized PHPSESSID cookie of about 819 KB decoded.
  • A second memory-exhaustion request.
  • An execution request adjusted to the leaked addresses, which started /bin/sh as a child of the Apache worker, running as the web user (uid 33).

The commands that followed identified the user and host with id, hostname, and uname -m, and tried to write proof files to temporary and web-root paths. The attackers then tried to fetch a second stage with wget, curl, or PHP's file_get_contents, and none of the downloads completed. Karimi counted a shell only when a worker-parented /bin/sh appeared. A PHP out-of-memory error, an oversized session cookie, or a worker crash on its own did not count as a successful exploit. He also notes that each honeypot ran a single Apache worker, a setup favorable to the attacker. The result does not measure how reliably the exploit works against other worker counts, PHP builds, or applications.

What php2shell exploitation looks like in logs and telemetry

The checks below look for exploitation behavior, because the few indicators published so far, listed at the end of this section, are mostly shared hosting and anonymization infrastructure. Each check follows from how the flaws work and comes with what normal looks like, because PHP workers produce some of these events in ordinary operation. A scanner that announces itself in the User-Agent, or a worker that simply crashes, tells you a host was probed or hit, not that an exploit succeeded. The process-lineage and web-root checks below are what catch a successful, quiet exploit. Search from October 17, the date of the earliest exploitation attempt Amazon has reported, through the time each host was patched and restarted.

Attempt · probed or hit, success unknownCompromise · exploit succeededNot every attempt succeeds1Version probeGET /phpinfo.php or/index.php that fingerprintthe PHP build, often froma self-identifying scanner.Web, proxy and WAF logs.Confirm it's a sanctionedscan before escalating.2Exploit attemptA multipart POST to anendpoint that never takesuploads, or a worker crashor memory-limit error.PHP-FPM / Apache logs, WAF.The host was hit;success isn't yet known.3Code executionA PHP worker spawns ashell, interpreter ordownload tool, or a new.php file in the web root.EDR process and file telemetry.Likely compromised.Investigate to confirm.4PersistenceNew cron jobs, SSH keysor CMS admin accounts,and outbound connectionsfrom the worker.EDR and network telemetry.Contain, preserve logs,rotate secrets.

Version-probing scans. The most common php2shell signal is reconnaissance: requests to paths like /phpinfo.php or /index.php that fingerprint a server's PHP version, often from tools that identify themselves in the User-Agent. Most is untargeted internet scanning, and some is a security team's or vendor's authorized exposure check, so confirm whether a scan was sanctioned before treating it as an incident. Read it as a prompt to verify the target's patch level, not as evidence of compromise.

Multipart requests outside normal upload traffic. In web server, proxy, or WAF logs, look for multipart/form-data POSTs to endpoints that never take uploads, and for multipart requests sent with chunked transfer encoding instead of a declared length. Legitimate uploads usually go to a short, known set of endpoints and declare their size. Review requests with oversized or deeply nested form fields the same way.

Memory-limit errors and HTTP 500s after those requests. Wiz recommends watching for unusual multipart requests followed by memory-limit errors or repeated 500 responses. PHP logs out-of-memory fatal errors, when error logging is on, with one of the strings below. The second appears when PHP's own allocator fails with no finite PHP memory limit in force. If a cgroup limit or the kernel out-of-memory killer stops the worker first, PHP may log neither string, so pair this with the worker-crash check below. Occasional memory errors from known report or import jobs are normal, while a cluster that follows multipart requests from one external source is a lead.

  • Allowed memory size of
  • Out of memory (allocated

ReliaQuest recommends searching for POST requests to PHP resources that return HTTP 500 and are followed by those memory errors. In its lab testing, the exploit's memory errors ended with in Unknown on line 0, which means they fired before any script ran. An out-of-memory error from a report job names a script file and line instead. ReliaQuest also lists zend_mm_heap corrupted as a related error. It advises against filtering on large allocation sizes, because the final failed allocation in its lab was small.

Worker crashes and churn. Wiz also lists unexplained PHP worker churn as a signal. The CVE-2026-109619 advisory says an attacker's requests have to reach the same worker, and that running several workers lowers the success rate per request. Expect repeated attempts from one source as a result. Search PHP-FPM and Apache error logs for these strings, then line up the crash times against the access log to find the requests that preceded them:

  • exited on signal 11
  • exit signal Segmentation fault
  • exit signal Abort (6)
  • AH00051 and AH00052, the Apache error codes on those exit-signal messages

Where the logs record process IDs, ReliaQuest suggests confirming that the crashed worker is the one that served the suspicious request.

On Windows servers running PHP under IIS, look for Windows Error Reporting (WerFault.exe) capturing a php-cgi.exe crash. In the IIS case above, the first command followed the crash by about 90 minutes, so extend the search well past the minutes right after a crash.

Execution timeouts. The stale-state flaw can also be reached by forcing a request to time out, since PHP treats a timeout as the kind of fatal interruption that leaves stale state behind. Watch PHP and PHP-FPM logs for Maximum execution time of entries, and line them up with the worker crashes and repeated requests from one source above.

Session upload progress activity. The confirmed CVE-2026-109619 paths run through session upload progress handling. Where a WAF or request logging captures POST fields, look for unusual session upload activity, including requests carrying the PHP_SESSION_UPLOAD_PROGRESS field. Amazon recommends searching form-upload POSTs for that string. Defused Cyber reported probes that set the field to p2s-rider with an empty file upload, which Defused reads as a check for exposure rather than an exploit attempt.

Oversized or malformed session cookies. Wiz also lists unexpectedly large or malformed PHPSESSID cookies in responses. Default web server and proxy logs do not record response headers, so this check works only where a WAF or load balancer captures Set-Cookie values. In the confirmed honeypot exploitation, the memory-disclosure step returned a PHPSESSID cookie of about 819 KB decoded.

PHP workers spawning shells. Code execution runs as the web worker account, so whatever an attacker runs next usually starts as a child of a PHP worker process. In EDR or process telemetry, review shells, interpreters, and download tools started by these parents:

  • php-fpm, php-cgi, lsphp, frankenphp, httpd, and apache2 on Linux
  • php-cgi.exe on Windows servers running PHP under IIS, with w3wp.exe (the IIS worker) as its parent

Children such as sh, bash, curl, wget, python, perl, cmd.exe, powershell.exe, and certutil.exe need review, though some are expected. PHP's mail() function calls sendmail through /bin/sh on Linux, and some applications call external image or PDF converters through a shell. Baseline the parent-child pairs each host produced before October 17 and investigate the new ones. In the IIS case above, the new pair was php-cgi.exe spawning cmd.exe under IUSR. Our post on the SharePoint zero-day applied the same process lineage check to w3wp.exe. Our EDR alert investigation guide covers tracing a suspicious process back through its process tree.

What the worker account changed. Three more checks cover the period after a possible compromise:

  • New or modified files in the web root and other writable, web-served paths, where a web shell would be written.
  • Outbound connections from PHP worker processes to destinations the host had not contacted before October 17. PHP applications routinely call payment, email, and update services, so compare against each host's own history.
  • New persistence owned by the web user, such as user cron jobs, SSH keys, and CMS administrator accounts.

Published indicators. Amazon describes its addresses as shared anonymization infrastructure and advises investigating a match rather than blocking it. The honeypot sources are registered to hosting providers. Use these to find hosts worth a closer look:

  • Amazon Threat Intelligence: 103.216.220.92 (October 17), plus 31.223.184.166, 37.19.200.5, 134.209.159.64, 103.114.163.188, and 95.217.98.37 (October 21).
  • Confirmed shells on Karimi's honeypots: exploit sources 15.204.59.70 and 162.35.99.94, with download attempts to 162.35.99.94:8080 and 15.204.59.70:8081. Exploit-shaped requests with no confirmed shell came from 104.28.208.84 and 144.48.4.242.
  • Exposure probes against Palo Alto GlobalProtect honeypots, reported by Defused Cyber according to The Hacker News: 204.76.203.7, sending the p2s-rider marker.

If a check returns a hit. Treat the host as compromised until an investigation rules it out. Preserve the web server, PHP, and PHP-FPM logs before log rotation removes them. Then rotate every credential the worker account could read, including database passwords in application configuration files. Rebuild from a patched image instead of cleaning the host in place. Our guide to automated incident response covers the containment workflow that follows a confirmed compromise.

How Prophet AI Threat Hunter covers php2shell

Prophet AI Threat Hunter runs the php2shell hunt automatically across customer environments, against the telemetry each customer has connected, on the scope and schedule each team sets. Vulnerable and end-of-life PHP installs come back flagged for patching. Any suspicious finding goes to Prophet AI SOC Analyst for a full investigation, which returns a determination with the queries and evidence behind it.

When indicators are published, Prophet AI Threat Hunter adds them to its emerging-threat hunts. Validated findings become candidate detections that go through the peer review described in our post on detection engineering in an AI-driven SOC. Our overview of AI threat hunting explains how an investigation's findings turn into hunts across the rest of the environment.

To run the php2shell checks across your own environment without writing the queries by hand, see how Prophet AI Threat Hunter handles it.

What is php2shell?

php2shell is the name for CVE-2026-109618 and CVE-2026-109619, two critical flaws in the PHP interpreter that can let an unauthenticated web request run code as the web worker account. PHP fixed both on October 21, 2026, in PHP 8.5.12, 8.4.27, 8.3.36, and 8.2.35, and CISA added both to its Known Exploited Vulnerabilities catalog on October 26. Every PHP install serving web requests is in scope, whatever application it runs.

Is php2shell being exploited in the wild?

Yes, php2shell is being exploited. CISA added both CVEs to its Known Exploited Vulnerabilities catalog on October 26, 2026, with an October 29 deadline for federal civilian agencies. Amazon Threat Intelligence traced the earliest attempt it saw to October 17, before disclosure, and a researcher recorded 35 shells on mod_php honeypots on October 24. Prophet Security has observed activity consistent with exploitation, including command execution through a PHP worker.

Which PHP versions fix php2shell?

PHP 8.5.12, 8.4.27, 8.3.36, and 8.2.35 fix both php2shell vulnerabilities, and the Windows rebuilds with a -1 suffix contain the same fixes. PHP 8.1 and older get no official releases, although the PHP team published source backports for 7.4, 8.0, and 8.1. Restart every PHP-FPM and Apache worker after updating. None of these releases fixes CVE-2026-115560, a separate session upload progress flaw.

What is CVE-2026-115560, and is it part of php2shell?

CVE-2026-115560 is a separate PHP remote code execution flaw, disclosed on October 23, 2026, that the php2shell releases do not fix. PHP says it requires an unusual setup: an application that switches session.serialize_handler to php with ini_set() before calling session_start(), while session upload progress is enabled. PHP will fix it in its next regular release and recommends session.upload_progress.enabled=0 until then.

Is Citrix NetScaler affected by php2shell?

Yes, NetScaler ADC and NetScaler Gateway are affected through the PHP-based management GUI and NITRO API on the NSIP, management-enabled SNIPs, and the Cluster IP, in all supported versions. Citrix released fixed builds 14.1-73.49 and 13.1-64.32 on October 27. Until one is installed, restrict management access to trusted administrative hosts. The PAN-OS, Panorama, and Cisco Expressway management interfaces are also affected.

How do you check a server for php2shell exploitation?

Search from October 17 for a PHP worker process, such as php-fpm, php-cgi, httpd, or apache2, starting a shell, interpreter, or download tool, and for new files in the web root. Supporting signals include multipart POSTs to endpoints that never take uploads, PHP memory errors ending in Unknown on line 0, worker crashes, oversized PHPSESSID cookies, and form posts carrying PHP_SESSION_UPLOAD_PROGRESS.

Table of contents
Add as Google Preferred Sources

Insights

Definitive Guide to Detection Engineering

Where detection engineering falls short, how to close those gaps, and what it looks like in an AI-enabled SOC.

Download Guide
Ajmal Kohgadai

Ajmal Kohgadai

As the Director of Product Marketing at Prophet Security, Ajmal drives marketing and growth strategies and helps security professionals see how AI is transforming security operations.