Skip to main content
All incident reports
Incident report

CISA confirms active exploitation of PHP2Shell remote code execution flaws

CISA has added two critical PHP flaws, tracked as PHP2Shell, to its catalogue of exploited vulnerabilities, confirming active attacks. The PHP project patched the flaws on 21 October, and a researcher reported successful remote code execution against honeypots on 24 October.

· First published

The short version

  • The flaws. CVE-2026-109618 and CVE-2026-109619, both rated 9.2, allow unauthenticated remote code execution in PHP. The PHP project patched them on 21 October.
  • CISA action. CISA added both flaws to its Known Exploited Vulnerabilities catalogue, confirming active exploitation. The agency has not published further details of the ongoing attacks.
  • Exploitation evidence. Perplexity's Adel Karimi confirmed remote code execution against honeypots on 24 October, running PHP 8.4.26 on Apache mod_php. Earlier attempts detected by several security firms were tagged unsuccessful, and Amazon said targeting began four days before the patches.
  • Exposure. PHP is used in roughly 70% of websites. Censys estimates more than 32.51 million PHP services still run an affected version, while VirusTotal estimates more than 5.1 million PHP hosts are exposed online.
  • Separate flaws. The PHP security team issued mitigation for CVE-2026-115560, an RCE flaw with no patch, and the PHP Foundation released Windows builds for a Brotli decoder flaw, CVE-2026-111497. Neither source links them to PHP2Shell exploitation.

Timeline

DateEvent
21 OctoberThe PHP project patches CVE-2026-109618 and CVE-2026-109619.
24 OctoberPerplexity's Adel Karimi confirms remote code execution against honeypots running PHP 8.4.26 on Apache mod_php.
26 OctoberThe PHP Foundation releases updated Windows builds for CVE-2026-111497, a Brotli decoder flaw.
26 OctoberCISA adds the two PHP2Shell flaws to its Known Exploited Vulnerabilities catalogue, according to reports published on 27 October.

What happened

CISA added CVE-2026-109618 and CVE-2026-109619 to its Known Exploited Vulnerabilities catalogue, according to reports on 27 October. The agency's listing confirms active exploitation, the sources say. Both flaws are rated 9.2 and permit unauthenticated remote code execution. CISA has not published further details of the attacks, BleepingComputer reported.

The PHP security team patched the two flaws on 21 October, one day after urging teams to install updates. It said proof-of-concept exploits published before the patches were not functional. Cloudflare, Amazon, Hadrian, watchTowr, Previdian and Patchstack detected widespread exploitation attempts before public disclosure but tagged them as unsuccessful. Amazon Threat Intelligence said attackers began targeting the flaws four days before patches were released.

Perplexity detection and response lead Adel Karimi reported successful remote code execution against honeypots on 24 October. The systems ran PHP 8.4.26 under Apache mod_php. Karimi said the evidence came from an intentionally favourable single-worker mod_php profile and did not measure exploit reliability on other worker counts, PHP builds or applications.

Who is affected

Unpatched PHP deployments are exposed. PHP is a server-side scripting language used in roughly 70% of websites. Censys estimates more than 32.51 million PHP services still run an affected version; VirusTotal estimates more than 5.1 million PHP hosts are exposed online, with about 3.8 million having at least one port open. The sources give different estimates and methods.

The Hacker News lists advisories from many software and security vendors, including WordPress, Okta, Salesforce and Siemens.

Separate flaws and responses

The PHP Foundation released updated Windows builds on 26 October to address CVE-2026-111497, an out-of-bounds write in the Brotli decoder library. The flaw affects Brotli versions 1.0.3 through 1.0.9, 1.1.0 and 1.2.0 and is patched in versions 1.0.10, 1.1.1 and 1.2.1. The foundation said PHP Windows builds can reach the vulnerable decoder through the cURL extension when automatic HTTP response decompression is enabled, which is off by default in cURL.

BleepingComputer reported that the PHP security team published mitigation for CVE-2026-115560, a new RCE flaw with no patch. The source says it is exploitable when targeting systems with an unusual combination of configuration and application code. Neither source connects either flaw to the confirmed PHP2Shell exploitation.

What is not known yet

  • Who is exploiting CVE-2026-109618 and CVE-2026-109619, and what access or data they have obtained.
  • How many systems have been successfully compromised; CISA has not published attack details.
  • Whether the confirmed honeypot exploitation used the same chain as attacks on production systems.
  • Which third-party products and appliances that embed PHP are confirmed vulnerable, beyond the vendor advisories.
  • Whether CVE-2026-115560 or CVE-2026-111497 is being exploited, and whether either is linked to the PHP2Shell activity.

What to do

If you run PHP

  1. Inventory internet-facing PHP services and record the PHP version they run.
  2. Update to PHP 8.5.12 or the corresponding fixed release for your branch, following the PHP project's advisory.
  3. Apply the PHP security team's mitigation for CVE-2026-115560; the source says no patch is available.

If you run PHP on Windows

  1. Check whether automatic HTTP response decompression is enabled in cURL.
  2. Apply the updated PHP Windows builds released on 26 October for CVE-2026-111497.

For everyone else

  1. Check vendor advisories for products that embed PHP and apply updates; WordPress, Okta, Salesforce and Siemens are among vendors that issued advisories.
  2. Use CISA's KEV catalogue to prioritise the two PHP2Shell CVEs in risk-based vulnerability management.
  3. Do not treat early reports that pre-patch exploits were non-functional as evidence that the flaws are not being exploited; CISA has confirmed active exploitation.

Update log

  • First published.

Sources

New incident reports by email

One email a week with the reports published since the last one.

Want us to look at your domain?

Give us one apex domain and we'll send back a written snapshot of what is already visible from outside. No call required.