<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
<title>everythingsfine.io</title>
<link>https://everythingsfine.io/</link>
<atom:link href="https://everythingsfine.io/feed.xml" rel="self" type="application/rss+xml"/>
<description>Working fixes for the security fires everyone is pretending not to smell. Written by engineers who have to deploy them, with code you can paste on Monday.</description>
<language>en</language>
<item>
<title>NGINX: 10 CVEs, 5 Critical. Install 1.30.7 or 1.31.9</title>
<link>https://everythingsfine.io/fires/nginx-1-30-7-1-31-9-nov-2026/</link>
<guid isPermaLink="true">https://everythingsfine.io/fires/nginx-1-30-7-1-31-9-nov-2026/</guid>
<pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate>
<category>Fix published</category>
<description>&lt;p&gt;A script that finds which of the ten NGINX issues of November 18, 2026 apply to your configuration, the fixed versions, and the tested interim step for each issue.&lt;/p&gt;&lt;h2&gt;Summary&lt;/h2&gt;
&lt;p&gt;On November 18, 2026, F5 published an out-of-band security notification, F5 article K000164058. It has ten articles for NGINX Open Source and NGINX Plus.&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;By the CVSS v4.0 scores that F5 gives, 5 are Critical, 1 is High and 4 are Medium.&lt;/li&gt;
&lt;li&gt;For the 5 Critical issues, F5 says that an attacker can stop or corrupt an NGINX worker process. F5 also says that the attacker can run code if Address Space Layout Randomization (ASLR) is disabled or the attacker can bypass it.&lt;/li&gt;
&lt;li&gt;The High issue can put a worker in an endless loop. The worker then stops serving.&lt;/li&gt;
&lt;li&gt;Of the Medium issues, two stop a worker, one lets a client skip the client certificate check in &lt;code&gt;stream&lt;/code&gt;, and one can send a &lt;code&gt;stream&lt;/code&gt; connection to the wrong backend.&lt;/li&gt;
&lt;li&gt;Each issue needs a specific module or configuration. A configuration that does not have it is not affected by that issue.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;NGINX released 1.30.7 (stable) and 1.31.9 (mainline) on November 18, 2026. They correct all the issues that apply to their branch.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Install NGINX Open Source 1.30.7 or 1.31.9, or the NGINX Plus release in each F5 article. This is the only complete correction.&lt;/strong&gt; Until your change window, the check script in this page tells you which issues apply to your configuration. For six issues, F5 gives a mitigation. For the other four, the only interim step is to turn off the protocol that the issue needs. We tested the script, each F5 mitigation and each interim step on NGINX Open Source 1.30.6, 1.30.7, 1.31.8 and 1.31.9. We did not test NGINX Plus.&lt;/p&gt;
&lt;p&gt;The same notification has twelve BIG-IP and BIG-IQ issues. This page does not include them: see F5 article K000164058.&lt;/p&gt;
&lt;h2&gt;Does this apply to you?&lt;/h2&gt;
&lt;p&gt;You have more exposure if one or more of these statements is true:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;You use NGINX Open Source 1.30.6 or older, or 1.31.0 to 1.31.8.&lt;/li&gt;
&lt;li&gt;You use NGINX Plus R34 to R36, 37.0.x before 37.0.8.1, or 37.1.x or 37.2.x before 37.3.0.1.&lt;/li&gt;
&lt;li&gt;Your configuration has the condition for one of the issues. See &amp;quot;Conditions&amp;quot;.&lt;/li&gt;
&lt;li&gt;Clients that you do not trust can connect to the listener that the condition applies to.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;a href=&quot;https://everythingsfine.io/fires/nginx-1-30-7-1-31-9-nov-2026/&quot;&gt;Read the full runbook: the ten issues, the check script, the F5 mitigations, the interim steps and what to monitor&lt;/a&gt;&lt;/p&gt;</description>
</item>
<item>
<title>OpenJDK Java4Shell: turn off Base64 intrinsics until you patch</title>
<link>https://everythingsfine.io/fires/openjdk-17-to-27-java4shell-base64/</link>
<guid isPermaLink="true">https://everythingsfine.io/fires/openjdk-17-to-27-java4shell-base64/</guid>
<pubDate>Tue, 17 Nov 2026 00:00:00 GMT</pubDate>
<category>Fix published</category>
<description>&lt;p&gt;How to find the JVMs that CVE-2026-114938 exposes, a Linux check script, the JVM option that removes the vulnerable code, and a lab-tested BIG-IP iRule, until you install a fixed JDK.&lt;/p&gt;&lt;h2&gt;Summary&lt;/h2&gt;
&lt;p&gt;On November 17, 2026, Oracle and the OpenJDK Vulnerability Group published fixes for CVE-2026-114938. Calif found the vulnerability and named it Java4Shell.&lt;/p&gt;
&lt;p&gt;The vulnerability is in the HotSpot JVM. The JVM has special code that decodes Base64 with AVX-512 instructions. Calif says a short, malformed Base64 input makes this code write 62 bytes after the end of the output buffer. Java then refuses the input, but the write is already done.&lt;/p&gt;
&lt;p&gt;Oracle gives the vulnerability a CVSS 3.1 score of 8.1 and says it is difficult to exploit. Oracle also says an attacker needs no account and can use more than one network protocol. Calif gives a CVSS 4.0 score of 9.2. Calif says it got remote code execution without a logon on default installations of Keycloak, Jira, Confluence, Artifactory, Jenkins and PingFederate. Calif has published a test for the memory corruption. Calif has not published the attacks on those products.&lt;/p&gt;
&lt;p&gt;We did the memory corruption test in our lab on November 18, 2026. These are our results:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;The corruption is real.&lt;/strong&gt; On an AMD Zen 5 CPU, Temurin 21.0.13+9 and 25.0.5+8 wrote 62 bytes after the end of the buffer for the Calif input. Temurin 17.0.21+9 did the same for an input of 128 characters. An emulated Intel CPU gave the same results.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The fixed builds are clean.&lt;/strong&gt; Amazon Corretto 17.0.21.1.2, 21.0.13.1.2 and 25.0.5.1.2 did not corrupt memory.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The interim option works.&lt;/strong&gt; With &lt;code&gt;-XX:-UseBASE64Intrinsics&lt;/code&gt;, the affected builds did not corrupt memory.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The iRule works.&lt;/strong&gt; In block mode on a BIG-IP, it refused each input that corrupted an affected JVM.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;Install the fixed JDK from your vendor in each affected JVM.&lt;/strong&gt; Until you can do that, the option &lt;code&gt;-XX:-UseBASE64Intrinsics&lt;/code&gt; stops the JVM from using the vulnerable code. This page gives you a script that finds the exposed JVMs on a Linux host, and the steps to set the option. It also gives you a BIG-IP iRule that finds and refuses the malformed input in HTTP requests.&lt;/p&gt;
&lt;h2&gt;Does this apply to you?&lt;/h2&gt;
&lt;p&gt;A JVM is exposed if all of these statements are true:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;The JVM is HotSpot from OpenJDK, or from a product built from OpenJDK. Calif lists Oracle Java SE, Oracle GraalVM for JDK, Amazon Corretto, Eclipse Temurin, Microsoft Build of OpenJDK, Azul Zulu, BellSoft Liberica, SAP SapMachine and Red Hat OpenJDK. We tested Eclipse Temurin.&lt;/li&gt;
&lt;li&gt;The Java version is on the 17, 21, 25 or 27 line and does not have the fix. The OpenJDK advisory gives 17.0.21.0.3, 21.0.13.0.3, 25.0.5.0.3, 27.0.1.0.3 and earlier. Calif says the vulnerable code starts at 17.0.6. In our lab, Temurin 17.0.5+8 did not have the vulnerable code. We did not test 17.0.6 or the 27 line.&lt;/li&gt;
&lt;li&gt;The CPU is x86-64 and has the flags &lt;code&gt;avx512f&lt;/code&gt;, &lt;code&gt;avx512bw&lt;/code&gt;, &lt;code&gt;avx512vbmi&lt;/code&gt; and &lt;code&gt;bmi2&lt;/code&gt;. Calif names Intel Xeon from Ice Lake and AMD from Zen 4. We tested AMD Zen 5. We tested Intel Ice Lake and Sapphire Rapids only on an emulator.&lt;/li&gt;
&lt;li&gt;The application decodes Base64 data that an attacker can send. Calif says that logon functions and file uploads do this.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Many products include a JDK of their own. The Java version of the operating system does not tell you the version that such a product uses. Examine each running JVM.&lt;/p&gt;
&lt;p&gt;You are not affected in these conditions:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;A CPU without AVX-512 VBMI.&lt;/strong&gt; In our lab, Temurin 21.0.13+9 on an Intel Kaby Lake CPU did not corrupt memory.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;ARM64.&lt;/strong&gt; Calif says ARM64 is not affected. We did not test ARM64.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Java 8 and Java 11, standard builds.&lt;/strong&gt; The OpenJDK advisory does not list the 8 and 11 lines for this CVE. Oracle lists one Java 8 build as affected: &lt;code&gt;8u511b33-perf&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;BIG-IP, BIG-IQ, F5 Insight and F5OS.&lt;/strong&gt; F5 article K000164290 says these products are not vulnerable. On November 18, 2026, the article showed that F5 was still examining F5 Distributed Cloud and NGINX One Console.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Java lines that have no more updates, such as 18 to 20, 22 to 24 and 26, get no fix. Our estimate is that they have the vulnerable code. We did not test them. Move them to a line that has the fix.&lt;/p&gt;
&lt;p&gt;&lt;a href=&quot;https://everythingsfine.io/fires/openjdk-17-to-27-java4shell-base64/&quot;&gt;Read the full runbook&lt;/a&gt;&lt;/p&gt;</description>
</item>
<item>
<title>BIG-IP: 12 fixes in the F5 out-of-band notice of November 18</title>
<link>https://everythingsfine.io/fires/bigip-nov-2026-engineering-hotfix/</link>
<guid isPermaLink="true">https://everythingsfine.io/fires/bigip-nov-2026-engineering-hotfix/</guid>
<pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate>
<category>Fix published</category>
<description>&lt;p&gt;F5 announced 12 BIG-IP security issues on November 18, 2026. A script that compares your configuration with each one, and the hotfix for each version.&lt;/p&gt;&lt;h2&gt;Summary&lt;/h2&gt;
&lt;p&gt;On November 18, 2026, F5 published an out-of-band security notification, F5 article K000164058. It has 12 articles for BIG-IP.&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;By the CVSS v3.1 scores that F5 gives, 6 are Critical and 6 are High.&lt;/li&gt;
&lt;li&gt;F5 says that 10 of them let a remote attacker stop the Traffic Management Microkernel (TMM). Traffic stops while TMM starts again. For the 6 Critical issues, F5 says that the attacker may also be able to run code.&lt;/li&gt;
&lt;li&gt;9 of the 12 are corrected only in the engineering hotfixes that F5 released on the same day. F5 corrected the other 3 in earlier releases.&lt;/li&gt;
&lt;li&gt;Each issue needs a specific module, iRule command or setting. A system that does not have it is not affected by that issue.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;We tested one of the issues, CVE-2026-125272, in a lab before F5 published. The result agrees with the F5 article. We will add an interim iRule for it after the request is public.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Install the engineering hotfix for your version. This is the only complete correction.&lt;/strong&gt; Until your change window, the check script in this page shows you which issues apply to your configuration.&lt;/p&gt;
&lt;h2&gt;Does this apply to you?&lt;/h2&gt;
&lt;p&gt;F5 lists these BIG-IP versions as vulnerable: 17.1.0 to 17.1.3, 17.5.0 to 17.5.1, and 21.1.0. F5 examines only versions that it supports. For an older version, ask F5 Support.&lt;/p&gt;
&lt;p&gt;An issue applies to a system with one of those versions if its condition is true. The conditions are from the F5 articles.&lt;/p&gt;
&lt;div class=&quot;table-wrap&quot;&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;F5 article&lt;/th&gt;
&lt;th&gt;CVE&lt;/th&gt;
&lt;th&gt;Condition&lt;/th&gt;
&lt;th&gt;CVSS v3.1&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;K000163937&lt;/td&gt;
&lt;td&gt;CVE-2026-123637&lt;/td&gt;
&lt;td&gt;APM: a per-request policy with the Client Information agent and Endpoint Security agents&lt;/td&gt;
&lt;td&gt;Critical 9.8&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163939&lt;/td&gt;
&lt;td&gt;CVE-2026-120285&lt;/td&gt;
&lt;td&gt;An iRule with the &lt;code&gt;PSC::ip_address&lt;/code&gt; command, and an HTTP profile&lt;/td&gt;
&lt;td&gt;Critical 9.8&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163940&lt;/td&gt;
&lt;td&gt;CVE-2026-120292&lt;/td&gt;
&lt;td&gt;An iRule with the &lt;code&gt;MR::message route&lt;/code&gt; command, and an HTTP profile&lt;/td&gt;
&lt;td&gt;Critical 9.8&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163941&lt;/td&gt;
&lt;td&gt;CVE-2026-116953&lt;/td&gt;
&lt;td&gt;An iRule that uses the &lt;code&gt;GENERICMESSAGE::route&lt;/code&gt; command on client traffic&lt;/td&gt;
&lt;td&gt;Critical 9.8&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163945&lt;/td&gt;
&lt;td&gt;CVE-2026-120722&lt;/td&gt;
&lt;td&gt;An iRule that uses the &lt;code&gt;DNS::scrape&lt;/code&gt; command on client traffic&lt;/td&gt;
&lt;td&gt;Critical 9.8&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163946&lt;/td&gt;
&lt;td&gt;CVE-2026-121805&lt;/td&gt;
&lt;td&gt;An iRule that uses the &lt;code&gt;PEM::subscriber&lt;/code&gt; command on client traffic&lt;/td&gt;
&lt;td&gt;Critical 9.8&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163936&lt;/td&gt;
&lt;td&gt;CVE-2026-125274&lt;/td&gt;
&lt;td&gt;All configurations: an iControl SOAP user with the Guest role or higher&lt;/td&gt;
&lt;td&gt;High 8.8&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163929&lt;/td&gt;
&lt;td&gt;CVE-2026-120718&lt;/td&gt;
&lt;td&gt;All configurations: iControl REST tokens&lt;/td&gt;
&lt;td&gt;High 8.1&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163938&lt;/td&gt;
&lt;td&gt;CVE-2026-125272&lt;/td&gt;
&lt;td&gt;APM: a per-request policy with a form-based Logon Page&lt;/td&gt;
&lt;td&gt;High 7.5&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163942&lt;/td&gt;
&lt;td&gt;CVE-2026-126822&lt;/td&gt;
&lt;td&gt;APM: an access profile that is a SAML service provider&lt;/td&gt;
&lt;td&gt;High 7.5&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163943&lt;/td&gt;
&lt;td&gt;CVE-2026-126336&lt;/td&gt;
&lt;td&gt;APM: a SAML service provider with encrypted assertions&lt;/td&gt;
&lt;td&gt;High 7.5&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;K000163944&lt;/td&gt;
&lt;td&gt;CVE-2026-121812&lt;/td&gt;
&lt;td&gt;APM: a per-request policy, and the &lt;code&gt;tmm.access.policytrace&lt;/code&gt; key enabled&lt;/td&gt;
&lt;td&gt;High 7.5&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;
&lt;p&gt;CVE-2026-120718 is High by CVSS v3.1 and Critical, 9.2, by CVSS v4.0.&lt;/p&gt;
&lt;p&gt;You are not affected by an issue if its condition is false. A BIG-IP system that has none of these modules, commands or settings still has the two iControl issues, unless its version already has those corrections. See &amp;quot;Find the fix for your version&amp;quot; below.&lt;/p&gt;
&lt;p&gt;This page is for BIG-IP. Some of the articles also apply to BIG-IQ Centralized Management, BIG-IP SSL Orchestrator and BIG-IP Next for Kubernetes. The notification also has 10 articles for NGINX. This page does not include them. Read F5 article K000164058 for those products.&lt;/p&gt;
&lt;p&gt;&lt;a href=&quot;https://everythingsfine.io/fires/bigip-nov-2026-engineering-hotfix/&quot;&gt;Read the full runbook: the 12 issues, the check script, the hotfix for each version, and what to monitor&lt;/a&gt;&lt;/p&gt;</description>
</item>
<item>
<title>node2shell: Node.js Brotli RCE, patch and block br bodies</title>
<link>https://everythingsfine.io/fires/nodejs-brotli-cve-2026-111497/</link>
<guid isPermaLink="true">https://everythingsfine.io/fires/nodejs-brotli-cve-2026-111497/</guid>
<pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate>
<category>Fix published</category>
<description>&lt;p&gt;Node.js 22, 24 and 26 decode Brotli with a memory bug that CISA lists as exploited. The fixed versions, a version check script, and a BIG-IP iRule for br request bodies.&lt;/p&gt;&lt;h2&gt;Summary&lt;/h2&gt;
&lt;p&gt;On October 21, 2026, the Node.js project released 22.23.4, 24.22.1 and 26.12.1. The releases correct CVE-2026-111497, which researchers call Node2Shell. Other names for it are Broccoli2Shell and HTTP2Shell.&lt;/p&gt;
&lt;p&gt;The Node.js advisory says this: an invalid block switch in a Brotli stream causes the decoder to write past the end of an eight-byte internal buffer. An attacker who can send Brotli data that an application decompresses with &lt;code&gt;node:zlib&lt;/code&gt; can corrupt process memory. The result can be a crash, disclosure of process memory, or remote code execution. No logon is necessary. The bug is in the Google Brotli library, so other software that decodes Brotli is also affected.&lt;/p&gt;
&lt;p&gt;On October 27, 2026, CISA added CVE-2026-111497 to its Known Exploited Vulnerabilities catalog. Node.js rates the issue high. The CVE record gives it CVSS 4.0 9.2, critical.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Upgrade Node.js to 22.23.4, 24.22.1, 26.12.1 or later, then restart each Node.js process.&lt;/strong&gt; Until you can do this, this page helps you find the vulnerable runtimes and stop the most direct attack path at the BIG-IP.&lt;/p&gt;
&lt;h2&gt;Does this apply to you?&lt;/h2&gt;
&lt;p&gt;You have more exposure if one or more of these statements is true:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;You run Node.js 22.x before 22.23.4, 24.x before 24.22.1, or 26.x before 26.12.1.&lt;/li&gt;
&lt;li&gt;You run an end-of-life Node.js line, for example 20.x, 23.x or 25.x. These lines get no fix.&lt;/li&gt;
&lt;li&gt;Your application uses Express 5 or body-parser 2. By default, their JSON, raw, text and form parsers decompress a request body that has the header &lt;code&gt;Content-Encoding: br&lt;/code&gt;. Any client can send such a request.&lt;/li&gt;
&lt;li&gt;Your code calls &lt;code&gt;zlib.brotliDecompress()&lt;/code&gt;, &lt;code&gt;zlib.brotliDecompressSync()&lt;/code&gt; or &lt;code&gt;zlib.createBrotliDecompress()&lt;/code&gt; on data from a user, an upload, a queue or a WebSocket.&lt;/li&gt;
&lt;li&gt;Your application calls &lt;code&gt;fetch()&lt;/code&gt; to servers that you do not control. The built-in &lt;code&gt;fetch()&lt;/code&gt; decompresses Brotli responses automatically, so a malicious or compromised server can attack the client.&lt;/li&gt;
&lt;li&gt;Your Node.js comes from a Linux distribution that links it to the system Brotli library. On those systems, the OS &lt;code&gt;brotli&lt;/code&gt; package must also be updated.&lt;/li&gt;
&lt;li&gt;You run other software that decodes Brotli: the Python &lt;code&gt;brotli&lt;/code&gt; and &lt;code&gt;brotlicffi&lt;/code&gt; packages, Envoy, Bun, or .NET. Their vendors have their own fixes.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;According to Akamai, a bare Node.js HTTP server does not decode Brotli request bodies. The application decides. A server that never decompresses Brotli data from outside and never calls untrusted servers has less exposure. It is still affected, so upgrade it.&lt;/p&gt;
&lt;p&gt;F5 says in article K000163902 that BIG-IP (all modules) and BIG-IQ are not vulnerable. F5 WAF for NGINX is affected. Use F5&#39;s own article for NGINX products.&lt;/p&gt;
&lt;p&gt;&lt;a href=&quot;https://everythingsfine.io/fires/nodejs-brotli-cve-2026-111497/&quot;&gt;Read the full runbook: the version check, the iRule, the Express setting, and what to monitor&lt;/a&gt;&lt;/p&gt;</description>
</item>
<item>
<title>php2shell: update PHP and BIG-IP, then close the gaps</title>
<link>https://everythingsfine.io/fires/php-8-php2shell/</link>
<guid isPermaLink="true">https://everythingsfine.io/fires/php-8-php2shell/</guid>
<pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate>
<category>Fix published</category>
<description>&lt;p&gt;Two critical PHP flaws, exploited from day one. The fixed releases, PHP settings, Advanced WAF signatures, an iRule for PHP behind BIG-IP, and the fix for BIG-IP itself.&lt;/p&gt;&lt;h2&gt;Summary&lt;/h2&gt;
&lt;p&gt;On October 21, 2026, the PHP project released PHP 8.2.35, 8.3.36, 8.4.27 and 8.5.12. These releases correct two critical vulnerabilities, CVE-2026-109618 and CVE-2026-109619. Together they are named &amp;quot;php2shell&amp;quot;. Each has a CVSS 4.0 score of 9.2. An attacker who has not logged in can run code on the web server with the privileges of the PHP worker.&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;CVE-2026-109618.&lt;/strong&gt; PHP can stop in the middle of a request when it runs out of memory. Request cleanup then uses memory that is not ready. The PHP project confirmed remote code execution on Apache with mod_php, with the default limits, before any application code runs.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;CVE-2026-109619.&lt;/strong&gt; A persistent PHP worker can keep data from a request after PHP releases that memory. A later request to the same worker can use the old data. The confirmed paths go through session upload progress and, on PHP 8.5, fatal error backtraces.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;F5 Labs saw the first scans five hours before the releases. CISA added both CVEs to its Known Exploited Vulnerabilities catalog on October 26, 2026.&lt;/p&gt;
&lt;p&gt;BIG-IP also uses PHP. F5 article K000163821 lists the BIG-IP Configuration utility on all modules, and the APM access profile, as affected. F5 released fixed BIG-IP versions on October 23, 2026.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Update PHP on every server, and upgrade BIG-IP. The PHP project says that no setting is a complete mitigation.&lt;/strong&gt; The steps on this page decrease your risk until your change window.&lt;/p&gt;
&lt;h2&gt;Does this apply to you?&lt;/h2&gt;
&lt;p&gt;You are at risk if one or more of these statements is true:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;A web server runs PHP 8.2, 8.3, 8.4 or 8.5 earlier than 8.2.35, 8.3.36, 8.4.27 or 8.5.12.&lt;/li&gt;
&lt;li&gt;A web server runs PHP 8.1 or earlier, and your vendor has not supplied a package that names these CVEs. The PHP project does not release fixes for these branches.&lt;/li&gt;
&lt;li&gt;You run PHP as an Apache module (mod_php). The PHP project confirmed remote code execution for both CVEs on mod_php. PHP-FPM, FrankenPHP and LiteSpeed (LSAPI) are also at risk.&lt;/li&gt;
&lt;li&gt;An appliance or product that you operate uses PHP for its web interface. Many network and storage vendors have published their own advisories.&lt;/li&gt;
&lt;li&gt;You operate BIG-IP 17.1.0 to 17.1.3, 17.5.0 to 17.5.1, or 21.1.0. If APM or SSLO is provisioned, the attack does not need a logon. On other systems, the attacker must log on to the Configuration utility.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Not affected, as F5 states in K000163821: BIG-IQ, F5OS, NGINX products, F5 Distributed Cloud and BIG-IP Next.&lt;/p&gt;
&lt;p&gt;Linux distributions often backport fixes and keep the old version number. A server that shows an old number can be fixed. Read the package changelog to confirm. The check script below does this for RPM and Debian packages.&lt;/p&gt;
&lt;p&gt;&lt;a href=&quot;https://everythingsfine.io/fires/php-8-php2shell/&quot;&gt;Read the full runbook: the version checks, PHP settings, Advanced WAF signatures, the iRule, the BIG-IP steps, and what to monitor&lt;/a&gt;&lt;/p&gt;</description>
</item>
<item>
<title>eMudhra G3: find the clients that still trust it</title>
<link>https://everythingsfine.io/fires/emudhra-g3/</link>
<guid isPermaLink="true">https://everythingsfine.io/fires/emudhra-g3/</guid>
<pubDate>Sat, 17 Oct 2026 00:00:00 GMT</pubDate>
<category>Interim fix</category>
<description>&lt;p&gt;Browsers block the eight weak eMudhra intermediates. Your other clients may not. Detect-mode iRules for a BIG-IP egress gateway that list them, with no decryption.&lt;/p&gt;&lt;h2&gt;Summary&lt;/h2&gt;
&lt;p&gt;On October 17, 2026, a researcher reported that eight eMudhra intermediate CAs have weak private keys. Six are TLS CAs and two are code-signing CAs. The keys were made with too little randomness, so anyone can recover them. With a recovered key, an attacker can make a certificate for any host name.&lt;/p&gt;
&lt;p&gt;Google and Mozilla confirmed the problem. Chrome and Firefox block the six TLS intermediates. On October 27, Chrome removed emSign ECC Root CA - G3 from its root store.&lt;/p&gt;
&lt;p&gt;Browsers got the fix. Many other clients did not. Java runtimes, network appliances, printers, containers and apps with their own CA bundle can still trust a certificate made with a recovered key.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Update each client trust store so it distrusts the eight weak intermediates.&lt;/strong&gt; Until you know which clients need it, the iRules in this page show you. They run on a BIG-IP egress gateway, they do not decrypt traffic, and they start in detect mode.&lt;/p&gt;
&lt;h2&gt;Does this apply to you?&lt;/h2&gt;
&lt;p&gt;You have more exposure if one or more of these statements is true:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;You have clients that do not use a browser trust store: Java with its own &lt;code&gt;cacerts&lt;/code&gt;, appliances, printers, IoT and OT devices, or containers with an old &lt;code&gt;ca-certificates&lt;/code&gt; package.&lt;/li&gt;
&lt;li&gt;You have apps that pin a CA or ship their own CA bundle.&lt;/li&gt;
&lt;li&gt;Your clients trust Sectigo AAA Certificate Services and do not check revocation. Sectigo cross-signed the weak intermediates in May 2025. Sectigo revoked the cross-signs on October 19, 2026, but a client that does not check revocation still accepts them.&lt;/li&gt;
&lt;li&gt;Your clients still trust emSign ECC Root CA - G3.&lt;/li&gt;
&lt;li&gt;Your distrust list names the weak intermediates by certificate hash only. On November 17, 2026, we found a third certificate for one weak key. See &amp;quot;Data group: the CA list&amp;quot;.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Current Chrome and Firefox are not affected: they block the six TLS intermediates. The iRules need a BIG-IP 17.5 or later system that all outbound TLS from your clients goes through. If your egress gateway already decrypts outbound TLS, you do not need the iRules. See &amp;quot;Before you use the iRules&amp;quot;.&lt;/p&gt;
&lt;p&gt;&lt;a href=&quot;https://everythingsfine.io/fires/emudhra-g3/&quot;&gt;Read the full runbook: the data group, three iRules, install and removal steps, and what to monitor&lt;/a&gt;&lt;/p&gt;</description>
</item>
</channel>
</rss>
