<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
	>

<channel>
	<title>Cybersecurity archivos | Fali Fuentes</title>
	<atom:link href="https://falifuentes.com/tag/cybersecurity/feed/" rel="self" type="application/rss+xml" />
	<link>https://falifuentes.com/tag/cybersecurity/</link>
	<description>Blog de Fali Fuentes (Málaga) &#124; Ciberseguridad, IA y Tecnología: Protege tu vida digital, domina tendencias tech y descubre análisis expertos.   ¡Actualizaciones diarias!</description>
	<lastBuildDate>Thu, 19 Nov 2026 19:12:01 +0000</lastBuildDate>
	<language>es</language>
	<sy:updatePeriod>
	hourly	</sy:updatePeriod>
	<sy:updateFrequency>
	1	</sy:updateFrequency>
	<generator>https://wordpress.org/?v=7.0</generator>

<image>
	<url>https://falifuentes.com/wp-content/uploads/2024/08/Favi_sec_p.png</url>
	<title>Cybersecurity archivos | Fali Fuentes</title>
	<link>https://falifuentes.com/tag/cybersecurity/</link>
	<width>32</width>
	<height>32</height>
</image> 
	<item>
		<title>NGINX2Root: 5 Critical RCEs in HTTP/2, HTTP/3, Resolver, and Mail Proxy — Patches Available</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-19-pm-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-19-pm-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Thu, 19 Nov 2026 19:12:01 +0000</pubDate>
				<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[English]]></category>
		<category><![CDATA[Threat hunting]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-19-pm-en/</guid>

					<description><![CDATA[<p>F5 releases patches for five RCEs in NGINX; CVE-2026-123007 affects HTTP/2 since 2015 and enables root escape chained with unpatched kernel SCTP bug.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-19-pm-en/">NGINX2Root: 5 Critical RCEs in HTTP/2, HTTP/3, Resolver, and Mail Proxy — Patches Available</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-19-pm-es/" hreflang="es">Leer este artículo en español</a></p>
<p>On November 18, F5 released patches for five remote code execution vulnerabilities in NGINX grouped under the name NGINX2Root. The most critical, CVE-2026-123007, affects the HTTP/2 module since version 1.9.5 from 2015 and only requires HTTP/2 over TLS enabled — the default configuration in virtually any modern HTTPS deployment. A working public exploit already achieves RCE as the NGINX worker user, and researchers have demonstrated chaining it with an unpatched flaw in the Linux kernel&#8217;s SCTP subsystem to escape the container and gain root on the host. If you administer web servers, Kubernetes ingress controllers, or base images with NGINX, this news forces you to act today.</p>
<h2>CVE-2026-123007 enables RCE with just HTTP/2 over TLS, the default configuration in virtually all modern HTTPS</h2>
<p>The flaw is a buffer overflow in HTTP/2 flow control window handling (ngx_http_v2_module). All it takes is for the server to accept encrypted HTTP/2 connections — which the vast majority of HTTPS sites do by default — for an unauthenticated attacker to corrupt worker memory. The exploit published by Rapid7 on GitHub (sfewer-r7/CVE-2026-123007) works against NGINX 1.31.8 in a container with HTTP/2/TLS, requires roughly 128 GB of HTTP/2 traffic, and bypasses ASLR by leaking heap addresses from server responses. The code is specific to glibc on x86-64, but it proves the vulnerability is exploitable under real-world conditions. <a href='https://my.f5.com/manage/s/article/K000163926'>The F5 advisory K000163926</a> confirms CVSS 4.0 9.2 and CVSS 3.1 8.1.</p>
<h2>Chaining the NGINX bug with a kernel SCTP use-after-free escapes to root on the host</h2>
<p>Researchers from Theori (Xint) have shown that after compromising the NGINX worker, a use-after-free in the Linux kernel&#8217;s SCTP implementation (present since version 4.15) can be leveraged to escalate privileges to root on the host. The kernel flaw has no assigned CVE and no public patch as of November 18; Xint is withholding technical details. This breaks the container boundary: if the container can open SCTP sockets (whether because it has CAP_SYS_MODULE or because the module is already loaded on the host), the jump to root is real. <a href='https://www.upwind.io/blog/nginx-cve-2026-123007-and-cve-2026-123633-four-web-facing-rces-one-patch-and-where-kubernetes-is-exposed'>Upwind&#8217;s analysis</a> details the full chain and notes that NGINX doesn&#8217;t use SCTP legitimately, so any SCTP socket opened from a worker process is an indicator of compromise.</p>
<h2>Ubuntu, Debian, RHEL, and OpenResty have no backported fix as of November 18, and ingress-nginx community is archived</h2>
<p>Here&#8217;s the operational problem: many organizations believe they&#8217;re patched because nginx -v shows a recent version, but distributions apply fixes without changing the version string. As of November 18, Ubuntu 22.04/24.04/26.04, Debian 12, RHEL 8/9, and OpenResty 1.31.8.1 have no fixed packages. Alpine 3.24 and edge do include 1.30.7-r0. Additionally, the ingress-nginx community controller has been archived since March 2026 (v1.15.1) and will not receive updates. The official Docker images were updated the same day with tags 1.30.7, 1.31.9, mainline, stable, and latest, but you must pull again and rebuild. <a href='https://www.tenable.com/blog/nginx2root-cve-2026-123007-frequently-asked-questions-about-five-critical-nginx'>Tenable&#8217;s FAQ</a> and <a href='https://cycode.com/blog/nginx2root-cve-2026-123007'>Cycode&#8217;s article</a> explain how to verify the actual package with dpkg -l, rpm -q, or apk info.</p>
<h2>Reloading the configuration isn&#8217;t enough: you must restart the binary or perform a hot binary upgrade</h2>
<p>An nginx -s reload doesn&#8217;t replace the in-memory binary; old workers keep serving existing connections with the vulnerable code. The only safe way is to restart the service (systemctl restart nginx or equivalent) or perform a hot binary upgrade. As a temporary measure, if you can&#8217;t patch today, disable HTTP/2 in exposed server blocks (http2 off and remove http2 from the listen directive) and reload — though old workers will remain vulnerable until they die. Verify with openssl s_client -connect host:443 -alpn h2 that h2 is no longer negotiated. This doesn&#8217;t protect against the HTTP/3 (CVE-2026-121810) or resolver (CVE-2026-123634) flaws. Also block the SCTP kernel module (echo «install sctp /bin/false» &gt; /etc/modprobe.d/disable-sctp.conf) and reboot the host or unload the module if it&#8217;s not in use. Start threat hunting from November 18 onward, looking for workers spawning binaries, repeated crashes, anomalous connection patterns, and SCTP sockets.</p>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;NGINX2Root: 5 Critical RCEs in HTTP/2, HTTP/3, Resolver, and Mail Proxy — Patches Available&quot;,
  &quot;summary&quot;: &quot;F5 releases patches for five RCEs in NGINX; CVE-2026-123007 affects HTTP/2 since 2015 and enables root escape chained with unpatched kernel SCTP bug.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-19T20:01:16.058+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-18&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-19-pm-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-19-pm-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;English&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://my.f5.com/manage/s/article/K000163926&quot;,
    &quot;https://www.upwind.io/blog/nginx-cve-2026-123007-and-cve-2026-123633-four-web-facing-rces-one-patch-and-where-kubernetes-is-exposed&quot;,
    &quot;https://www.tenable.com/blog/nginx2root-cve-2026-123007-frequently-asked-questions-about-five-critical-nginx&quot;,
    &quot;https://cycode.com/blog/nginx2root-cve-2026-123007&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-19-pm-en/">NGINX2Root: 5 Critical RCEs in HTTP/2, HTTP/3, Resolver, and Mail Proxy — Patches Available</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
		<item>
		<title>CISA adds critical RCE in Puppet Enterprise (CVE-2026-115731) to KEV</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-19-am-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-19-am-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Thu, 19 Nov 2026 05:11:43 +0000</pubDate>
				<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[English]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-19-am-en/</guid>

					<description><![CDATA[<p>CISA adds CVE-2026-115731 to its exploited vulnerabilities catalog; affects Puppet Enterprise and Core, allows RCE as puppet user, and requires patching before December 1 for federal agencies.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-19-am-en/">CISA adds critical RCE in Puppet Enterprise (CVE-2026-115731) to KEV</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-19-am-es/" hreflang="es">Leer este artículo en español</a></p>
<p>CISA added CVE-2026-115731 to its KEV catalog on November 17, a critical deserialization flaw in Puppet Enterprise and Puppet Core that is already being actively exploited and allows remote code execution as the Puppet Server service user; the deadline for U.S. federal agencies expires on December 1 and serves as a reference for <em>due diligence</em> for audits under ENS or ISO 27001 in Spain and Latam.</p>
<h2>Unsafe deserialization in report processing allows RCE as the puppet user</h2>
<p>The vector lies in the component that processes the JSON reports sent by agents. The server uses <code>FromDataConverter</code> to reconstruct types from a field <code>__ptype</code> controlled by the attacker; a malicious type definition reaches a code path that evaluates arbitrary Ruby code and executes with the privileges of the Puppet Server service user. The CVSS 4.0 score is 9.4 (AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H), confirming critical severity, low complexity, and scope change.</p>
<h2>Certificate auto-signing turns the flaw into RCE without prior authentication</h2>
<p>To exploit the vulnerability you need an agent certificate signed by the Puppet CA. If the CA is configured with auto-signing without strong verification (<code>autosign.conf</code> permissive), any actor with network access to port 8140 can obtain a valid certificate and chain the exploitation without prior credentials. The advisory from <a href='https://ccb.belgium.be/advisories/warning-deserialization-untrusted-data-puppet-server-can-be-exploited-execute-arbitrary'>CCB Belgium</a> rates it as &#8216;actively exploited in the wild, public PoC, patch immediately.&#8217;</p>
<h2>Patches available since November 3; the same update covers the authorization bypass</h2>
<p>Perforce released the fixes on November 3. The fixed versions are:</p>
<ul>
<li>Puppet Enterprise 2026.0.0 → <strong>2026.0.1</strong></li>
<li>Puppet Enterprise 2025.0.0 → <strong>2025.12.1</strong></li>
<li>Puppet Enterprise 2023.8.0 → <strong>2023.8.13</strong></li>
<li>Puppet Core 9.0.0 → <strong>9.1.3</strong></li>
<li>Puppet Core 8.x → <strong>8.22.2</strong></li>
</ul>
<p>These same releases also fix <strong>CVE-2026-115732</strong> (authorization bypass via certificate name, CVSS 9.0). Verify in the <em>release notes</em> from the vendor that both CVEs are listed as resolved.</p>
<h2>Federal December 1 deadline and reference for <em>due diligence</em> in ENS and ISO 27001</h2>
<p>The BOD 26-04 directive requires U.S. federal civilian agencies to remediate before <strong>December 1, 2026</strong> (14 days after inclusion in KEV). Although not binding outside that scope, CISA urges <em>all organizations</em> to prioritize remediation, and frameworks such as NIST SP 800-53 Rev. 5 (control SI-2) or PCI DSS 4.0 (req. 6.3.3) use it as a reference for <em>timely patching</em>. In Spain and Latam it serves as an objective argument in ENS, ISO 27001 audits, or contracts with vulnerability management clauses.</p>
<h2>What to do today: update, audit certificates, and look for prior compromise</h2>
<ol>
<li><strong>Update now</strong> to the patched versions and <strong>restart Puppet Server</strong>. Test in staging before production.</li>
<li><strong>Disable weak auto-signing</strong>: remove or harden <code>autosign.conf</code> requiring PSK, OIDC, or manual approval. This does not replace the patch, because a legitimately signed certificate can still exploit the bug.</li>
<li><strong>Restrict network access</strong> to trusted management networks; never expose port 8140 to the internet.</li>
<li><strong>Audit and revoke certificates</strong>: list with <code>puppetserver ca list --all</code> and revoke unrecognized ones with <code>puppetserver ca revoke --certname &lt;nombre&gt;</code>. You need a reliable inventory of managed nodes.</li>
<li><strong>Look for prior compromise (forensic)</strong>: review logs of the endpoint <code>/puppet/v3/report</code> (anomalous PUTs), child processes of the user <code>puppet</code>, CA keys in <code>/etc/puppetlabs/puppet/ssl/ca/</code> and new service accounts. If you detect compromise, rotate <strong>all</strong> CA keys and credentials distributed by Puppet.</li>
<li><strong>If you use Puppet Open Source (community)</strong>: there is no official patch; plan migration to <strong>OpenVox</strong> or <strong>Puppet Core</strong> supported, evaluating module compatibility and <em>workflows</em>.</li>
</ol>
<blockquote><p>Note on the contradiction in sources: <em>SecurityAffairs</em> cites Perforce as saying that “there are no reports of active exploitation,” while the alert from <a href='https://www.cisa.gov/news-events/alerts/2026/11/17/cisa-adds-one-known-exploited-vulnerability-catalog'>CISA</a>, the Belgian CCB, Neodyme AG, and Reddit witnesses confirm exploitation in the wild since November 4. Technical evidence and KEV inclusion prevail.</p></blockquote>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;CISA adds critical RCE in Puppet Enterprise (CVE-2026-115731) to KEV&quot;,
  &quot;summary&quot;: &quot;CISA adds CVE-2026-115731 to its exploited vulnerabilities catalog; affects Puppet Enterprise and Core, allows RCE as puppet user, and requires patching before December 1 for federal agencies.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-19T06:01:16.057+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-17&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-19-am-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-19-am-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;English&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://www.cisa.gov/news-events/alerts/2026/11/17/cisa-adds-one-known-exploited-vulnerability-catalog&quot;,
    &quot;https://ccb.belgium.be/advisories/warning-deserialization-untrusted-data-puppet-server-can-be-exploited-execute-arbitrary&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-19-am-en/">CISA adds critical RCE in Puppet Enterprise (CVE-2026-115731) to KEV</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
		<item>
		<title>CISA adds Check Point Gaia LPE (CVE-2026-123039) to KEV, chained with PHP2Shell</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-16-pm-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-16-pm-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Mon, 16 Nov 2026 19:13:22 +0000</pubDate>
				<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[English]]></category>
		<category><![CDATA[Malware]]></category>
		<category><![CDATA[Firewall]]></category>
		<category><![CDATA[malware]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-16-pm-en/</guid>

					<description><![CDATA[<p>CISA added CVE-2026-123039 to its KEV catalog on November 13; it is exploited in chain with PHP2Shell to gain root on Check Point gateways and, in some cases, destroy ESXi hypervisors.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-16-pm-en/">CISA adds Check Point Gaia LPE (CVE-2026-123039) to KEV, chained with PHP2Shell</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-16-pm-es/" hreflang="es">Leer este artículo en español</a></p>
<p>CISA added CVE-2026-123039 to its KEV catalog on November 13, a local privilege escalation in the Check Point Gaia DynamicCLI service that is being chained with PHP2Shell to gain root control of perimeter gateways and, in documented cases, destroy ESXi hypervisors.</p>
<h2>The vulnerability and why it is critical</h2>
<p>The process <code>dynamic_server</code>, which runs as root, listens on the local socket <code>/tmp/dynamic_socket</code> created with 0777 permissions and without client authentication. Any authenticated local user can send arbitrary commands that execute as root. Check Point rates it CVSS 3.1 7.8 High (<code>AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H</code>).</p>
<h2>Observed exploitation chain</h2>
<p>Check Point Research observed activity between October 30 and November 4: first PHP2Shell (CVE-2026-109618 and CVE-2026-109619, CVSS 9.2) on the internet-exposed Mobile Access portal, then the LPE to escalate to root, and finally persistence with PHP webshells and the TSGA/TSProxy backdoor. In a limited number of cases they deployed the wiper <code>destroy_v6</code> against VMware ESXi hosts. Only a few dozen unpatched systems in a single region were compromised according to the <a href='https://blog.checkpoint.com/security/security-advisory-active-exploitation-of-php2shell-and-local-privilege-escalation'>Check Point advisory</a>.</p>
<h2>Affected versions and mandatory patch</h2>
<p>Affects Security Gateways, Management/Log Servers, Multi-Domain Security Management, and Maestro Orchestrator on Gaia R82.20, R82.10, R82, R81.20, R81.10, and R81. Spark Firewalls and Smart-1 Cloud are not affected. The fix arrives in the November Jumbo Hotfix with minimum Takes: R82.20 Take 14+, R82.10 Take 52+, R82 Take 132+, R81.20 Take 174+. R81.10 and R81 are end-of-support and receive no patch; the only safe path is to migrate to a supported version. The same Jumbo also fixes PHP2Shell (sk1000302) and CVE-2026-107484, as detailed in the <a href='https://www.rescana.com/post/checkpoint-gaia-cve-2026-123039-kev-patch-now'>Rescana analysis</a>.</p>
<h2>Temporary mitigation if you cannot patch yet</h2>
<p>Stop <code>dynamic_server</code> and prevent its restart with the exact commands:</p>
<pre><code>dbset process:dynamic_server
dbset :save</code></pre>
<p>This disables the Dynamic CLI documented in sk144112; Check Point advises following the SK1000362 steps exactly. It does not replace the patch.</p>
<h2>What to check in your logs and IOCs</h2>
<p>Look in <code>/var/log/dynamicCli.log</code> commands that don&#8217;t use the expected Clish wrapper; on patched systems you&#8217;ll see «Rejected connection from unauthorized uid» entries indicating blocked attempts. The IOCs published by Check Point include the IPs <code>130.94.19.204</code> and <code>130.94.19.168</code>, PHP session files containing the string <code>upload_progress_pwn</code>, processes disguised as <code>CPconfigdaemon</code>, scanning tools and wiper in <code>/tmp</code> or <code>/temp</code>, the note <em>Dream Annexation</em> with no ransom demand and unexplained gaps in traffic logs after the logging daemon crashed. Absence of matches does not rule out compromise.</p>
<h2>Actions for providers and MSPs</h2>
<p>If a third party operates your gateways, demand evidence of the installed Take, installation date, and confirmation that the PHP2Shell fix (sk1000302) is applied on gateways with exposed Mobile Access. Also verify that the Mobile Access portal is not accessible from the internet unless strictly necessary. The <a href='https://www.cisa.gov/news-events/alerts/2026/11/13/cisa-adds-two-known-exploited-vulnerabilities-catalog'>CISA alert</a> sets a November 16 deadline for U.S. federal agencies, raising operational priority to a 72-hour SLA even if you&#8217;re not subject to BOD 26-04.</p>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;CISA adds Check Point Gaia LPE (CVE-2026-123039) to KEV, chained with PHP2Shell&quot;,
  &quot;summary&quot;: &quot;CISA added CVE-2026-123039 to its KEV catalog on November 13; it is exploited in chain with PHP2Shell to gain root on Check Point gateways and, in some cases, destroy ESXi hypervisors.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-16T20:01:16.061+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-13&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-16-pm-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-16-pm-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;English&quot;,
    &quot;Malware&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://www.cisa.gov/news-events/alerts/2026/11/13/cisa-adds-two-known-exploited-vulnerabilities-catalog&quot;,
    &quot;https://blog.checkpoint.com/security/security-advisory-active-exploitation-of-php2shell-and-local-privilege-escalation&quot;,
    &quot;https://www.rescana.com/post/checkpoint-gaia-cve-2026-123039-kev-patch-now&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-16-pm-en/">CISA adds Check Point Gaia LPE (CVE-2026-123039) to KEV, chained with PHP2Shell</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
		<item>
		<title>Google launches Gemini 4 Hydrogen: fast model for agents at $1.50/M tokens</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-13-pm-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-13-pm-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Fri, 13 Nov 2026 19:11:04 +0000</pubDate>
				<category><![CDATA[AI]]></category>
		<category><![CDATA[Artificial Intelligence]]></category>
		<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[English]]></category>
		<category><![CDATA[artificial intelligence]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-13-pm-en/</guid>

					<description><![CDATA[<p>Google completes the Gemini 4 family with Hydrogen, a fast model for agents and daily code at $1.50/M tokens, unified 1M window and configurable thinking; warns of more refusals in multimodal.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-13-pm-en/">Google launches Gemini 4 Hydrogen: fast model for agents at $1.50/M tokens</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-13-pm-es/" hreflang="es">Leer este artículo en español</a></p>
<p>Google has released <a href='https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-hydrogen'>Gemini 4 Hydrogen</a> on November 12, just 48 hours after <a href='https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-carbon-now-available'>Gemini 4 Carbon</a>. Hydrogen is the «fast model for everyday execution», adapted from Carbon, and closes the Gemini 4 family: Argon (general balance, October), Carbon (deep reasoning and frontier cybersecurity), Hydrogen (agents and daily code). It affects you if you develop agent flows, migrate from 3.8 Flash, or evaluate inference costs at scale.</p>
<h2>Unified 1M-token window and long decode change the economics of long-running agents</h2>
<p>Hydrogen shares a 1,048,576-token window across input, thinking, and output. The default output is 65,536 tokens, but with <code>long decode</code> in background reaches 1M output tokens. Previously you had 1M input and 64K output separate; now the same budget covers large artifacts (code, reports, 3D visualizations) in a single conversation. Note: each turn charges the full input tokens.</p>
<h2>Configurable thinking with high default: adjust cost and latency</h2>
<p>Four levels: <code>low</code>, <code>medium</code>, <code>high</code> (default) and <code>xhigh</code> (new). In 3.8 Flash the default was <code>medium</code>. Raising to <code>xhigh</code> raises the AA Index from 51 to 54, but consumes more tokens and latency. For simple tasks, <code>low</code> or <code>medium</code> reduce cost. Configure it explicitly in the API call.</p>
<h2>Benchmarks: wins in its class but the regression in multimodal refusals forces monitoring</h2>
<p>Google reports Terminal-Bench 4.1 at 57% versus 18.2% for 3.8 Flash, reduces fabricated claims &gt;5× and is 2.5× faster in setting <code>high</code>. Leads its class in HLE-Diamond (58.8%), Vals Tax Agent Bench (52.3%, #1 global), Vals Legal Research Bench (55.8%, tie #1) and AutomationBench-AA. Agent safety: overreach -81%, under-service -79%, misaligned actions -12×, honesty 96.7%, CoT readability 99.9% without deceptive edits in &gt;200k turns, resistance to indirect injection &gt;2.3× (Gray Swan IPI).</p>
<p>The model card acknowledges a real regression: unjustified refusals rise 5.57 percentage points vs 3.8 Flash, «especially in multimodal prompts». Image-to-text improves 7 pp in safety. If your app uses images, PDFs, or audio intensively, check production logs for false positives.</p>
<h2>Orchestrator in Antigravity combines Carbon and Hydrogen to save quota and time</h2>
<p>The Orchestrator dynamically alternates Carbon (coordinator/auditor) and Hydrogen (workers/verifiers). In Terminal-Bench 4.0 it achieves success rates close to Carbon while consuming 24% less quota and 19% less time. The native multi-model strategy (/swarm) allows maintaining quality on mixed tasks without paying for Carbon for everything.</p>
<h2>What to do now: migration, Fairwind, and scientific program</h2>
<ol>
<li><strong>Evaluate 3.8 Flash → Hydrogen migration</strong> with your real workloads (latency, cost/1k accepted tasks). Partners report 2–2.8× speedup and -12% to -23% tokens, but Google does not detail test conditions.</li>
<li><strong>Use Orchestrator/swarm in Antigravity</strong> for mixed workloads: Carbon for architecture and deep debugging, Hydrogen for writing, tests, and verification.</li>
<li><strong>Configure thinking level explicitly</strong>; do not assume the default <code>high</code> if your tasks are simple.</li>
<li><strong>Monitor unjustified refusals in multimodal</strong>; the 5.57 pp increase is significant in production.</li>
<li><strong>Leverage the unified 1M window + long decode</strong> for agents that generate large artifacts, accounting that each turn reloads the input.</li>
<li><strong>Enroll in the Fairwind Program</strong> if you audit open source code or critical infrastructure; Carbon demonstrates 100% recall and 95% precision in scanning and triage, with &gt;25% relative improvement vs Argon in detection of real logic flaws.</li>
<li><strong>Researchers: apply to the scientific program</strong> (10k spots, 6 months free Argon+Carbon in Antigravity) for formal verification, structural prediction, enzymatic assays, etc.</li>
</ol>
<p>Hydrogen&#8217;s flat pricing ($1.50/M input, $7.50/M output) doubles current 3.8 Flash, but 3.8 Flash will rise to that level on January 1, 2027. Knowledge cutoff: March 2026.</p>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;Google launches Gemini 4 Hydrogen: fast model for agents at $1.50/M tokens&quot;,
  &quot;summary&quot;: &quot;Google completes the Gemini 4 family with Hydrogen, a fast model for agents and daily code at $1.50/M tokens, unified 1M window and configurable thinking; warns of more refusals in multimodal.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-13T20:01:16.036+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-12&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-13-pm-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-13-pm-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;AI&quot;,
    &quot;Artificial Intelligence&quot;,
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;English&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-carbon-now-available&quot;,
    &quot;https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-hydrogen&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-13-pm-en/">Google launches Gemini 4 Hydrogen: fast model for agents at $1.50/M tokens</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
		<item>
		<title>Apache HTTP Server 2.4.71 Fixes Critical RCE in mod_http2 (CVE-2026-119618)</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-12-pm-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-12-pm-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Thu, 12 Nov 2026 19:20:30 +0000</pubDate>
				<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[English]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-12-pm-en/</guid>

					<description><![CDATA[<p>Apache releases version 2.4.71 fixing a race condition in mod_http2 that allows unauthenticated remote code execution on servers with HTTP/2 enabled.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-12-pm-en/">Apache HTTP Server 2.4.71 Fixes Critical RCE in mod_http2 (CVE-2026-119618)</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-12-pm-es/" hreflang="es">Leer este artículo en español</a></p>
<p>On November 11, the Apache Software Foundation released Apache HTTP Server 2.4.71, which fixes three vulnerabilities; the most severe is CVE-2026-119618, rated as <em>critical</em> by Apache: a race condition in the function <code>assign_headers()</code> of the module <code>mod_http2</code> (CWE-362) that corrupts the heap and allows unauthenticated remote code execution <a href='https://www.rescana.com/post/apache-httpd-mod-http2-cve-2026-119618-patch-now'>according to the Apache advisory</a>. It affects versions 2.4.55 through 2.4.70 inclusive and the standalone module <code>mod_h2</code> 2.0.8–2.0.42, whose repository has been archived since June 2026 and will not receive an official patch <a href='https://www.rescana.com/post/apache-httpd-mod-http2-cve-2026-119618-patch-now'>confirmed by Rescana</a>.</p>
<h2>The race condition in assign_headers allows unauthenticated RCE</h2>
<p>The flaw occurs when concurrent HTTP/2 requests access shared resources without proper synchronization. Only servers with HTTP/2 enabled —the <code>Protocols h2</code> or <code>h2c</code>directive— reach the vulnerable code; HTTP/1.1-only configurations are not exposed <a href='https://socprime.com/blog/cve-2026-119618-critical-apache-http-2-flaw-enables-remote-code-execution'>as clarified by SOCPrime</a>. The report came through two independent channels: Anthropic&#8217;s OSS scanner (October 26, identifier ANT-2026-MAPKT6MZ) and Anthony Weems with Simon Scannell of Google Cloud Vulnerability Research (November 3). Apache landed the fix in commit 824beab9b9424d6cfb29595248fdbf6c3dddeab1 and released 2.4.71 eleven days later. The same release also addresses CVE-2026-106023 (null dereference in <code>mod_md</code>, severity <em>important</em>) and CVE-2026-109163 (stack overwrite in <code>mod_vhost_alias</code>, severity <em>low</em>).</p>
<h2>Standalone mod_h2 left without official patch, forcing migration</h2>
<p>The module <code>mod_h2</code> maintained on GitHub (used in custom builds and containers) has no fixed version published. The only supported path is to migrate to the <code>mod_http2</code> integrated in httpd 2.4.71 or manually apply the backported patch in your fork. If you maintain base images or containers that embed <code>mod_h2</code>, audit them now: there will be no official backported patch.</p>
<h2>Distros and vendors move at different paces</h2>
<p>Debian unstable (<code>sid</code>) already includes <code>apache2 2.4.71-1</code>; <code>bookworm</code>, <code>trixie</code>, <code>forky</code> and Ubuntu still had no package as of November 11. Red Hat released updates for RHEL 9 (<code>mod_http2-2.0.26-7.el9_9.1</code>) and RHEL 10 (<code>2.0.29-5.el10_3.1</code>), but rates the severity as <em>Important</em> arguing that its per-stream allocator raises exploitation complexity; RHEL 8 and earlier do not contain the vulnerable code <a href='https://www.rescana.com/post/apache-httpd-mod-http2-cve-2026-119618-patch-now'>Rescana details</a>. Verify your provider&#8217;s schedule before assuming coverage.</p>
<h2>What to do today: update, disable HTTP/2, or mitigate while the package arrives</h2>
<p>The priority is to update to 2.4.71 or your OS security package as soon as it’s available, starting with internet‑facing front‑ends that have HTTP/2 enabled. Validate third‑party module compatibility and configuration in a test environment before promoting to production. If you can’t update right away, disable HTTP/2 by removing <code>h2</code> and <code>h2c</code> from the <code>Protocols</code> (global and in <strong>all</strong> VirtualHosts) and force <code>Protocols http/1.1</code>; it’s a defensive hardening, not a fix, and you lose multiplexing and push <a href='https://socprime.com/blog/cve-2026-119618-critical-apache-http-2-flaw-enables-remote-code-execution'>recommends SOCPrime</a>. Take inventory with <code>httpd -M | grep http2</code> and review <code>Protocols</code> in <code>httpd.conf</code> and all includes. Monitor logs after patching: worker crashes with <code>SIGSEGV</code> in <code>assign_headers</code> or unexpected restarts may indicate failed exploitation attempts <a href='https://www.rescana.com/post/apache-httpd-mod-http2-cve-2026-119618-patch-now'>notes Apache</a>. Finally, ask your SaaS, PaaS, or CDN providers that use Apache as a front‑end whether they have already applied 2.4.71 or an equivalent backported patch.</p>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;Apache HTTP Server 2.4.71 Fixes Critical RCE in mod_http2 (CVE-2026-119618)&quot;,
  &quot;summary&quot;: &quot;Apache releases version 2.4.71 fixing a race condition in mod_http2 that allows unauthenticated remote code execution on servers with HTTP/2 enabled.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-12T20:01:16.040+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-11&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-12-pm-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-12-pm-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;English&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://www.rescana.com/post/apache-httpd-mod-http2-cve-2026-119618-patch-now&quot;,
    &quot;https://socprime.com/blog/cve-2026-119618-critical-apache-http-2-flaw-enables-remote-code-execution&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-12-pm-en/">Apache HTTP Server 2.4.71 Fixes Critical RCE in mod_http2 (CVE-2026-119618)</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
		<item>
		<title>OpenSSH &#8216;SSHole&#8217; (CVE-2026-123158): Authentication Bypass in Alpine Linux and musl libc</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-10-am-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-10-am-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Tue, 10 Nov 2026 05:26:16 +0000</pubDate>
				<category><![CDATA[Cloud Security]]></category>
		<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[Email]]></category>
		<category><![CDATA[English]]></category>
		<category><![CDATA[Supply Chain]]></category>
		<category><![CDATA[Firewall]]></category>
		<category><![CDATA[Password]]></category>
		<category><![CDATA[Threat hunting]]></category>
		<category><![CDATA[VPN]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-10-am-en/</guid>

					<description><![CDATA[<p>Technical analysis of the authentication bypass in OpenSSH on musl without PAM: conditions, mechanism, cloud image impact, and immediate mitigations.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-10-am-en/">OpenSSH &#8216;SSHole&#8217; (CVE-2026-123158): Authentication Bypass in Alpine Linux and musl libc</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-10-am-es/" hreflang="es">Leer este artículo en español</a></p>
<p>The CVE-2026-123158 vulnerability, dubbed &#8216;SSHole,&#8217; allows a remote attacker to authenticate without a password on OpenSSH servers compiled against musl libc without PAM. With a CVSS 8.1 score, it primarily affects Alpine Linux in its default configuration, where PasswordAuthentication yes and the absence of PAM create the attack surface. The flaw was fixed in OpenSSH 10.6p2 and in Alpine packages published on November 5, 2026; although disclosure occurred five days ago, the patching window remains critical because the PoC is public, scanning of port 22/TCP with oversized credentials is automated, and the path to root in cloud images is documented and reproducible.</p>
<h2>Exploitation Conditions and Scope in Alpine Linux</h2>
<p>Exploitation requires three simultaneous conditions: OpenSSH prior to 10.6p2 linked to musl libc without PAM, PasswordAuthentication yes (the default in Alpine), and at least one local account with a SHA-256 or SHA-512 hash ($5$/$6$) in /etc/shadow. If an account with a password field of * (no password, typically for key-only access) exists, the attacker can send an oversized password (257–1024 bytes) and achieve authentication. Alpine confirmed that all branches 3.16–3.24 and edge are vulnerable.</p>
<p>The official advisory from <a href='https://www.alpinelinux.org/posts/2026-11-05-openssh-password-authentication-bypass'>Alpine Linux</a> details that the openssh-server package is compiled without PAM by default, exposing the vast majority of installations. The &#8216;tiny-cloud&#8217; cloud images (3.18.11–3.24.3) include the alpine user with a * field and passwordless doas, which turns an initial bypass into full root compromise once any $6$ hash is set on the system. This privilege chain is documented in the Alpine advisory and in the analysis by <a href='https://safe.security/resources/blog/sshole-high-severity-openssh-vulnerability-turns-disabled-passwords-into-unauthorized-access'>Safe Security</a>.</p>
<p>No exploitation in real-world campaigns has been confirmed as of November 6, only scanning and proof-of-concept activity in a honeypot. The first observed exploitation occurred in less than nine hours after publication, from an IP geolocated in Japan, executing reconnaissance commands id, uname, and hostname. The speed of public PoC emergence drastically reduces the safe patching window.</p>
<h2>Technical Mechanism: Chain of Three Failures in musl libc</h2>
<p>The bug is not memory corruption but a logical flaw in password validation introduced by the CVE-2016-6210 patch. In musl libc, the crypt() function returns an empty string when the input exceeds 256 bytes, and OpenSSH interprets that result as a match for the * field of passwordless accounts. Reuse of the salt from an account with a valid hash completes the chain of three failures: 256-byte limit, erroneous comparison, and absence of PAM.</p>
<p>Safe Security explains that the attacker needs to know a username with a * field and that at least one $5$/$6$ hash exists on the system. The oversized password forces crypt() to fail silently, producing the empty string that OpenSSH equates with the disabled-account marker. This behavior only occurs in builds without PAM, because PAM handles verification separately and is unaffected.</p>
<p>The upstream patch in OpenSSH 10.6p2 fixes password validation for accounts with a * field. Alpine backported the fix to all branches, including EOL 3.16–3.20, demonstrating good supply-chain practice. The fix requires restarting sshd after updating the package; a mere apk upgrade does not replace the running daemon.</p>
<h2>Impact on Cloud Images and Privilege Escalation to Root</h2>
<p>Alpine&#8217;s official tiny-cloud images are a critical vector: they boot with the alpine user (* field) and passwordless doas. Setting any console password creates a $6$ hash, instantly activating the vulnerability for that user. A remote attacker then gains credentialless SSH access and, via doas -n id, obtains root in a single step. Safe Security published a full demonstration of this path.</p>
<p>This dynamic breaks the assumed security of &#8216;key-only&#8217; accounts: the * field no longer guarantees immunity to password authentication if it coexists with any SHA-crypt hash. Blocking root with PermitRootLogin prohibit-password does not mitigate the risk if another user with passwordless escalation exists. Defense must address both the patch and local privilege configuration.</p>
<p>In containerized environments, the vulnerable base image persists even if the host is patched. The recommendation is to rebuild images on corrected Alpine packages (e.g., 10.3_p1-r2 in 3.24, 9.0_p1-r6 in 3.16) and redeploy containers. The attack surface includes appliances, routers, and any device exposing SSH with Alpine&#8217;s default configuration.</p>
<h2>Detection, exploitation speed, and patching window</h2>
<p>Standard Alpine logs (/var/log/messages) record Accepted password for on passwordless accounts, but not the length of the password sent. The key indicator is a successful password login on an account configured for key-only access. Correlating with user creation, authorized SSH keys, and subsequent doas/sudo commands enables effective threat hunting.</p>
<p>Exploitation in a honeypot within nine hours demonstrates that mass scanning of port 22/TCP with oversized credentials is automated. The critical patching window demands prioritizing updates for systems exposed to the internet or management networks. The absence of confirmed campaign exploitation does not reduce urgency; the public PoC and exploit simplicity (one connection, no retries) guarantee rapid adoption by threat actors.</p>
<p>For reference, the same week saw disclosure of CVE-2026-116606/CVE-2026-111560 in Barracuda ESG (pre-auth authorization bypass + command injection), different in vector (email appliance) but equal in patching urgency. Both cases illustrate the convergence of authentication flaws in critical infrastructure and the need for automated, tested patching processes.</p>
<h2>Immediate mitigations and hardening recommendations</h2>
<p>The priority action is to update to OpenSSH 10.6p2 or the corresponding fixed Alpine package for each branch (e.g., 10.3_p1-r2 in 3.24, 9.0_p1-r6 in 3.16) and restart sshd with rc-service sshd restart. In containers, rebuild the image and redeploy. Verify with sshd -T | grep -i passwordauthentication that the value is no after applying PasswordAuthentication no as a temporary measure.</p>
<p>If immediate patching is not possible, disable password authentication (PasswordAuthentication no in sshd_config or a drop-in at /etc/ssh/sshd_config.d/00-no-passwords.conf), validate configuration with sshd -t, and restart. Requires working SSH key access; keep the session open until validating the new login. Does not protect if UsePAM yes and PAM allows passwords through another path.</p>
<p>Audit accounts with * field in /etc/shadow and escalation privileges (doas, passwordless sudo). Remove or lock unnecessary accounts. Restrict SSH exposure to management networks (VPN, bastion, cloud firewall). Disable sshd in containers that do not require it. These measures are defense in depth; they do not replace the patch but reduce attack surface while the update is deployed.</p>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;OpenSSH 'SSHole' (CVE-2026-123158): Authentication Bypass in Alpine Linux and musl libc&quot;,
  &quot;summary&quot;: &quot;Technical analysis of the authentication bypass in OpenSSH on musl without PAM: conditions, mechanism, cloud image impact, and immediate mitigations.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-10T06:01:35.051+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-05&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-10-am-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-10-am-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;Cloud Security&quot;,
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;Email&quot;,
    &quot;English&quot;,
    &quot;Supply Chain&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://www.alpinelinux.org/posts/2026-11-05-openssh-password-authentication-bypass&quot;,
    &quot;https://safe.security/resources/blog/sshole-high-severity-openssh-vulnerability-turns-disabled-passwords-into-unauthorized-access&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-10-am-en/">OpenSSH &#8216;SSHole&#8217; (CVE-2026-123158): Authentication Bypass in Alpine Linux and musl libc</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
		<item>
		<title>Authorization flaws in Grafana 12.4 and path bypass in Armeria: patches available</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-09-pm-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-09-pm-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Tue, 10 Nov 2026 00:41:44 +0000</pubDate>
				<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[English]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-09-pm-en/</guid>

					<description><![CDATA[<p>Technical analysis of CVE-2026-33379 in Grafana and CVE-2026-121823 in Armeria: root cause, real impact, and mitigations with operational limits for observability and service mesh environments.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-09-pm-en/">Authorization flaws in Grafana 12.4 and path bypass in Armeria: patches available</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-09-pm-es/" hreflang="es">Leer este artículo en español</a></p>
<p>The near-simultaneous publication of two security advisories on November 6, 2026 exposes authorization flaws in Grafana 12.4 and a route-based control bypass in Armeria prior to 1.43.0. Both components are ubiquitous in observability stacks and Java/Kotlin service meshes, so the exposure window, though brief, demands immediate patch deployment. This analysis breaks down root cause, real impact, and mitigations with their limits.</p>
<h2>CVE-2026-33379: authorization bypass in Grafana snapshots</h2>
<p>The flaw CVE-2026-33379, classified as CWE-862 (Missing Authorization), affects exclusively Grafana OSS and Enterprise 12.4.0 through 12.4.3; earlier branches and version 13 are out of scope. An authenticated user with the minimal Viewer role or a permissionless service account can create, publish externally, and delete dashboard snapshots, breaking the confidentiality of internal metrics and the integrity of observability history.</p>
<p>The CVSS 3.1 vector is AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N, yielding a base score of 8.5 (HIGH). Exploitation requires only valid low-privilege credentials and network access to the Grafana HTTP API; no victim interaction is needed. The absence of authorization checks on the /api/snapshots endpoints allows exfiltration of sensitive data embedded in dashboards, such as datasource tokens or queries with secrets.</p>
<p>Grafana Labs released the patch in version 12.4.4 on the same day as disclosure. The upgrade is mandatory and requires a service restart. As a compensating measure, network access can be restricted to management CIDRs and external snapshot functionality disabled via snapshot_remove_expired = true and an empty external_snapshot_url in grafana.ini, though this removes a legitimate panel-sharing capability. See details in <a href='https://ubuntu.com/security/CVE-2026-33379'>Ubuntu Security Tracker</a>.</p>
<h2>Technical details and attack vector in Grafana</h2>
<p>The root of the problem lies in the lack of permission validation before executing snapshot creation, publication, and deletion operations. The POST /api/snapshots endpoint accepts any authenticated user without verifying roles, and the external parameter allows publishing the snapshot to an Internet-accessible URL. Additionally, DELETE /api/snapshots/:key does not verify snapshot ownership, allowing an attacker to delete other users&#8217; evidence.</p>
<p>In multi-tenant environments or with dashboards embedding datasource credentials, a public snapshot exposes API keys, database passwords, and cloud tokens. The high confidentiality score (C:H) reflects this risk. There is no public record of active exploitation as of November 9, 2026, but the exploit&#8217;s triviality suggests that detection via audit logs (POST/DELETE on /api/snapshots) is the only early signal.</p>
<p>If snapshots have been published, rotation of exposed credentials is imperative: datasource keys, API keys, and embedded passwords. The operation carries high operational cost and compromise must be assumed if logs are inconclusive. Temporary disabling of snapshots (external_snapshot_name = «») reduces the attack surface while the patch is applied, but affects legitimate collaboration workflows between SRE and development teams.</p>
<h2>CVE-2026-121823: inconsistent path normalization in Armeria</h2>
<p>In Armeria, the flaw CVE-2026-121823 stems from inconsistent normalization of route segments containing a dot (.). Versions prior to 1.43.0 allow requests such as GET /api/../admin to evade path-scoped decorators used for authorization. The CVSS 4.0 vector is AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N, with a base score of 8.7 (HIGH), and requires no authentication.</p>
<p>The patch in version 1.43.0, released on November 5, rejects with 400 Bad Request any request that includes a &#8216;.&#8217; (dot) route segment. This closes the bypass in decorators such as PathMapping, Route, and Decorator that rely on pathPrefix or exactPath to enforce access policies. The official advisory (GHSA-xqhm-4ph4-423v) also groups CVE-2026-121819 (symlink in armeria-xds), CVE-2026-121824 (Thrift DoS), CVE-2026-121820 (SAML), and CVE-2026-121821 (CORS). Consult the <a href='https://armeria.dev/release-notes/1.43.0'>release notes</a>.</p>
<p>Exploitation allows an unauthenticated attacker to access protected resources: admin APIs, internal metrics endpoints, sidecar or gateway management panels. In service mesh architectures where Armeria acts as control plane or ingress, the bypass is the first step for lateral movement, SSRF, or privilege escalation. The absence of known public exploitation does not reduce urgency: the framework&#8217;s prevalence in Spring Boot and Kotlin environments makes it a priority target.</p>
<h2>Impact on service meshes and Armeria-based gateways</h2>
<p>Services that use path-scoped decorators to enforce path-prefix-based authorization are vulnerable if they do not validate normalization before delegating to the decorator. Common patterns such as /internal/**, /actuator/**, or /admin/** remain exposed. The fix requires recompiling and redeploying with Armeria 1.43.0, which entails verifying compatibility with Thrift 0.25, Netty 4.2.19, Jetty 12.0.40, and Spring Boot 4.1.2 per the release notes.</p>
<p>As a quick validation, one can send a request with a &#8216;.&#8217; segment in the path (e.g., GET /api/./../admin) and confirm that the server responds with 400 Bad Request. This test only covers CVE-2026-121823; it does not detect the other CVEs in the same release. Auditing your own code to identify all path-scoped decorators is necessary, because generic scanners cannot distinguish custom authorization logic.</p>
<p>If immediate updating is not possible, partial mitigations can be applied: exclude the brotli4j:native module and use the Java fallback for CVE-2026-111497, limit the size of SAML HTTP-Redirect messages to 1 MiB via SamlConfigurationBuilder.maxIncomingMessageSize(1048576), and review any-origin CORS policies that might leak headers to undeclared routes. These measures are temporary and do not replace the full patch.</p>
<h2>Priority mitigations and operational limits</h2>
<p>For Grafana, the primary action is to upgrade to 12.4.4 via docker pull grafana/grafana:12.4.4 or distribution packages (RHEL, Ubuntu, Debian backports). A restart is mandatory, and the changelog should be reviewed for breaking changes in plugins. As defense in depth, restrict API access to management networks and block egress to the Internet from the pod to reduce the impact of externally published snapshots.</p>
<p>In Armeria, upgrading to 1.43.0+ requires modifying Maven/Gradle dependencies (implementation &#8216;com.linecorp.armeria:armeria:1.43.0&#8217;) and the Spring Boot starter (armeria-spring-boot-starter:1.43.0). Redeployment is unavoidable. All PathMapping, Route, and Decorator instances that use pathPrefix/exactPath for authorization must be audited, because the fix only blocks &#8216;.&#8217; segments; other normalization bypasses could persist if the routing logic is flawed.</p>
<p>Both vulnerabilities share a short exposure window (patches available since November 5-6) but a broad attack surface. The operational priority is to deploy patches within the next 24-48 hours, validate with regression tests that snapshot functionality and path-based authorization work as expected, and rotate credentials if there are signs of compromise. Continuous monitoring of GitHub Security Advisories for Grafana and Armeria completes the response cycle.</p>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;Authorization flaws in Grafana 12.4 and path bypass in Armeria: patches available&quot;,
  &quot;summary&quot;: &quot;Technical analysis of CVE-2026-33379 in Grafana and CVE-2026-121823 in Armeria: root cause, real impact, and mitigations with operational limits for observability and service mesh environments.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-09T20:01:35.048+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-06&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-09-pm-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-09-pm-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;English&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://ubuntu.com/security/CVE-2026-33379&quot;,
    &quot;https://armeria.dev/release-notes/1.43.0&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-09-pm-en/">Authorization flaws in Grafana 12.4 and path bypass in Armeria: patches available</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
		<item>
		<title>Two critical flaws in MCP servers expose repositories and bypass SSRF</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-09-am-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-09-am-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Mon, 09 Nov 2026 13:07:47 +0000</pubDate>
				<category><![CDATA[AI]]></category>
		<category><![CDATA[Artificial Intelligence]]></category>
		<category><![CDATA[Cloud Security]]></category>
		<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[English]]></category>
		<category><![CDATA[Supply Chain]]></category>
		<category><![CDATA[artificial intelligence]]></category>
		<category><![CDATA[Firewall]]></category>
		<category><![CDATA[Zero Trust]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-09-am-en/</guid>

					<description><![CDATA[<p>Technical analysis of CVE-2026-41902 and CVE-2026-41904: confinement escape in mcp-server-git and SSRF bypass in mcp-server-fetch. Immediate mitigations without official patch.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-09-am-en/">Two critical flaws in MCP servers expose repositories and bypass SSRF</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-09-am-es/" hreflang="es">Leer este artículo en español</a></p>
<p>On November 8, 2026, two security advisories were published in the official modelcontextprotocol/servers repository affecting the tooling layer of the Model Context Protocol (MCP). Both flaws were reported by ShieldPi and are under review by the maintainers, who confirm partial technical validity. No official patch exists at the time of writing; defense relies on network, filesystem, and least-privilege deployment controls.</p>
<h2>Context: MCP and the exposed tooling layer</h2>
<p>MCP defines a standardized protocol for LLM agents to invoke local or remote tools such as filesystem, Git, HTTP, or databases. The reference servers mcp-server-git and mcp-server-fetch implement these capabilities and run with the privileges of the process that launches them. Any failure in their input validation breaks the trust model &#8216;the agent only sees what the server allows&#8217; and opens lateral movement or exfiltration vectors.</p>
<p>The attack surface grows when multiple agents share MCP servers without segmentation by scope or identity. A Git confinement escape or an SSRF bypass in fetch allows a malicious agent—or an injected prompt—to reach resources outside its scope: private repositories, cloud metadata, internal services. The severity lies not in the MCP protocol but in the implementation of its reference servers.</p>
<p>Both advisories were published as GitHub Security Advisories on the same day, indicating coordinated responsible disclosure. The maintainers (PederHP) acknowledge that &#8216;some technical claims appear valid&#8217; and warn that additional automated submissions will result in blocking. That response suggests patches in days, not weeks, but until then exposure is real in any deployment running versions 2026.11.4 or 2026.10.24.</p>
<h2>CVE-2026-41902: confinement escape in mcp-server-git</h2>
<p>The validate_opened_repo function verifies the worktree, git dir, and common dir, but omits .git/objects, objects/info/alternates, and symlinks in objects/. A crafted repository inside the allowed root that points its alternates to an external object database—or has objects/ as a symlink to an outside pool—allows git_show to return foreign commits and blobs, and git_checkout/git_create_branch to materialize external history in the authorized worktree.</p>
<p>The confirmed reproduction describes creating two repositories (allowed/crafted and outside), writing the external object path into crafted/.git/objects/info/alternates, and invoking git_show against the server started with &#8211;repository allowed. The negative control (sibling repo without alternates) correctly fails, isolating the cause in the incomplete validation of alternates.</p>
<p>The impact spans cross-disclosure of secrets, proprietary code, and history, as well as write priming of the authorized worktree from an external store. The proposed fix requires resolving .git/objects and each alternates target after opening the repo, demanding they remain under the authorized root or rejecting alternates outright, and revalidating in every tool that resolves objects. See the full advisory at <a href='https://github.com/modelcontextprotocol/servers/issues/5487'>GitHub Security Advisory #5487</a>.</p>
<h2>CVE-2026-41904: SSRF bypass in mcp-server-fetch</h2>
<p>The _is_public_address guard unwraps well-known IPv4-mapped and NAT64 forms but omits the fec0::/10 range (deprecated IPv6 site-local) and trusts ip.is_global, which returns True for fec0::1. Result: the SSRF allowlist accepts requests to non-public destinations. The proof of concept reproduces on versions 2026.11.4 and 2026.10.24; the 36-case matrix correctly rejects loopback, RFC1918, link-local, 307/308 redirects, userinfo, trailing-dot, uppercase-host, 0.0.0.0, and [::].</p>
<p>The reachability note clarifies that on modern kernels fec0::1 typically times out; the translated form [::ffff:0:127.0.0.1] is normalized and correctly blocked. This is a correctness flaw in the guard&#8217;s classifier (its threat model includes non-public destinations), not a proven exploit against metadata services on default hosts. The proposed fix explicitly blocks fec0::/10 and ::ffff:0:0/96, and applies deny-by-default over not ip.is_global plus reserved/site-local range checks.</p>
<p>The affected surface is any mcp-server-fetch deployment that resolves URLs controlled by the user or agent. The vector allows pivoting to internal networks if services are listening on site-local. Technical details and test matrix at <a href='https://github.com/modelcontextprotocol/servers/issues/5492'>GitHub Security Advisory #5492</a>.</p>
<h2>Immediate mitigations and limits</h2>
<p>For mcp-server-git: mount the served repository on a read-only filesystem for the server process; use &#8211;repository pointing to a bare copy without alternates or symlinks in objects/; run with a dedicated user without read access outside the root. Requires deployment control (Docker, systemd, K8s). Does not protect if the attacker writes to the served repo via another MCP tool with write capability.</p>
<p>For mcp-server-fetch: deny outbound traffic to fec0::/10 and ::ffff:0:0/96 at the firewall/host (iptables/nftables, cloud security groups); enforce deny-by-default on egress with an allow-list of required domains. Does not fix the logical bug; an attacker who controls DNS or uses a valid IPv4-mapped address could pivot if other flaws exist. Restrict scheme to https:// and allowed domains if the server exposes that configuration.</p>
<p>Cross-cutting measures: exact inventory and versioning of each MCP server per environment; principle of least privilege (dedicated container/user, read-only FS except explicit directories, minimal egress); logging of invocations (git_show, git_checkout, fetch) with parameters and requester; alerts on anomalous patterns (private IPs, ../, alternates); image integrity verification (cosign/sigstore, SBOM) and review of fix PR provenance; subscription to the repo’s GitHub Security Advisories and CVE/NVD.</p>
<h2>Next steps and defensive architecture</h2>
<p>Today: identify all deployments of mcp-server-git and mcp-server-fetch (versions 2026.11.4 or 2026.10.24) and apply temporary network and FS mitigations. Within 24-48 h: implement egress deny-by-default rules and domain allow-list for fetch; mount served Git repos read-only without alternates. Once the patch is released: deploy the fixed version (expected in days), validate in staging that it resolves the described PoCs.</p>
<p>Review MCP architecture: assess whether the “server per capability” model (Git, fetch, FS, SQL) is properly segmented and whether each agent has its own server scope (zero-trust between agents). The tooling layer is the most exposed surface when autonomous agents invoke tools; segmentation by identity and purpose reduces the blast radius of any individual flaw.</p>
<p>These flaws are not vulnerabilities in the MCP protocol, but in reference implementations. The prompt publication in GitHub Security Advisories and the maintainers’ response indicate imminent official patches; until then, network, filesystem, and least-privilege deployment controls are the only verified defense. Maintaining visibility into advisories and the supply chain will close the loop when the fixes arrive.</p>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;Two critical flaws in MCP servers expose repositories and bypass SSRF&quot;,
  &quot;summary&quot;: &quot;Technical analysis of CVE-2026-41902 and CVE-2026-41904: confinement escape in mcp-server-git and SSRF bypass in mcp-server-fetch. Immediate mitigations without official patch.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-09T06:01:35.033+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-08&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-09-am-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-09-am-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;AI&quot;,
    &quot;Artificial Intelligence&quot;,
    &quot;Cloud Security&quot;,
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;English&quot;,
    &quot;Supply Chain&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://github.com/modelcontextprotocol/servers/issues/5487&quot;,
    &quot;https://github.com/modelcontextprotocol/servers/issues/5492&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-09-am-en/">Two critical flaws in MCP servers expose repositories and bypass SSRF</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
		<item>
		<title>NCSC and ANSSI: Agentic AI and Ransomware in Critical Infrastructure</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-08-am-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-08-am-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Sun, 08 Nov 2026 05:11:44 +0000</pubDate>
				<category><![CDATA[AI]]></category>
		<category><![CDATA[Artificial Intelligence]]></category>
		<category><![CDATA[Compliance and Regulations]]></category>
		<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[English]]></category>
		<category><![CDATA[Malware]]></category>
		<category><![CDATA[Supply Chain]]></category>
		<category><![CDATA[artificial intelligence]]></category>
		<category><![CDATA[malware]]></category>
		<category><![CDATA[Ransomware]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-08-am-en/</guid>

					<description><![CDATA[<p>NCSC warns: AI will scale ransomware in critical infrastructure before 2028; ANSSI defines seven principles to govern agentic systems with deterministic authorization and immutable audit.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-08-am-en/">NCSC and ANSSI: Agentic AI and Ransomware in Critical Infrastructure</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-08-am-es/" hreflang="es">Leer este artículo en español</a></p>
<h2>NCSC: AI accelerates existing tradecraft and creates new attack surface</h2>
<p>On November 3, the National Cyber Security Centre published its State of the Cyber Nation report and CEO Richard Horne stated that it is highly likely that AI‑driven malware will be used in ransomware against critical national infrastructure before the end of 2028. The threat does not lie in AI discovering new vulnerabilities, but in it scaling the discovery and exploitation of technical debt tolerated for years. <a href='https://www.ncsc.gov.uk/news/cyber-chief-uk-facing-ai-powered-attacks-on-critical-infrastructure-in-two-years'>source</a>.</p>
<p>The NCSC’s highly likely assessment is based on China and Russia routinely employing AI for geopolitical purposes in cyber operations. The complementary report argues that the short‑term risk is twofold: AI makes known tradecraft faster and broader, and by embedding itself in critical infrastructure itself, it generates new attack surface. Sectors with legacy OT are priority targets.</p>
<p>The UK government announced alongside the report sixty million pounds for the first phase of the national Cyber Shield, seventy million for a National Real‑Time Exposure and Vulnerability Centre, and an AI Cybersecurity Centre of Excellence in Manchester. These investments reflect that the response requires coordinated state capability, not just isolated and sustained organizational measures.</p>
<h2>ANSSI: seven principles for securing agentic AI systems</h2>
<p>On October 29, the Agence nationale de la sécurité des systèmes d&#8217;information published Les Essentiels Sécuriser un système IA agentique version 1.0, a two‑page fact sheet with seven operational principles. The starting point is clear: an authorized action can be harmful because the agent may make mistakes, be manipulated, or chain permitted operations with a prohibited and systemic effect. <a href='https://universconvergents.com/2026/11/05/anssi-essentiels-securiser-systeme-ia-agentique'>source</a>.</p>
<p>The first principle requires independent deterministic authorization: an upstream filter for each tool, a policy immutable by the model, and verification of action chains. The second classifies actions into three categories — autonomous execution, mandatory human approval, prohibited action — and stipulates that critical operations always require a human approver. The third principle mandates that the approval mechanism reside outside the agentic system.</p>
<p>Principles four through seven complete the framework: untrusted inputs and outputs, secrets outside the model context, separate memories and external flows via controlled interfaces; emergency stop with access revocation and degraded mode without weakening controls; assume all data sent to a hosted model may be accessible to the provider; immutable audit with tamper‑proof logs, documented mission, annual review, human owner, and deployment blocked if unacceptable risk persists.</p>
<h2>Convergence: when agentic AI operates on critical infrastructure</h2>
<p>The cross‑synthesis reveals a common vector: AI scales the exploitation of technical debt in critical infrastructure per the NCSC, and agentic AI chains authorized actions with harmful effect per ANSSI. The joint priority action is to inventory agentic systems in OT and IT environments and perform threat modeling of action chains to anticipate critical abuse paths.</p>
<p>The shared key control deploys model‑independent policy enforcement points on every tool exposed to an agent: cyber hygiene and segmentation on one side, deterministic out‑of‑band authorization with critical human‑in‑the‑loop on the other. In governance, the UK Cyber Assessment Framework and the National Cyber Action Plan must align with GDPR Article 32 and EU AI Act Article 6 on high‑risk.</p>
<p>The ANSSI fact sheet serves as a concrete benchmark for demonstrating accountability to regulators and auditors. It impacts the supply chain: tender specifications must reflect security requirements derived from the agentic system risk analysis. A single registry of agentic systems with owner, criticality, action matrix, and annual review is needed to maintain continuous and permanently auditable traceability.</p>
<h2>Practical implementation: guardrails, secrets, and immutable audit</h2>
<p>Upstream deterministic authorization is implemented with a policy engine such as OPA or Cedar and an API gateway that validates action, params, and target before execution. It adds latency and requires an exhaustive catalog of tools and actions plus versioned policies to avoid uncontrolled drifts at runtime and ensure consistency across versions and continuous audit and automatic validation.</p>
<p>The classification of actions into three levels requires an action‑by‑criticality matrix reviewed by architecture and legal, and a human approval workflow based on Slack, Jira, or ITSM with timeout and escalation. It does not scale to thousands of trivial actions; only low‑risk actions are automated after rigorous threat modeling and continuous compliance monitoring and exhaustive documentation.</p>
<p>Secrets outside the context require Vault, AWS Secrets Manager, or Azure Key Vault plus a token broker that injects short‑lived credentials into the agent. Operational complexity is high: rotation and revocation must be automatic and audited, and immutable logs use append‑only storage with hash chaining and a SIEM with integrity alerts and legally defensible, continuous retention.</p>
<h2>Applicable recommendations for technical leadership</h2>
<p>Applicable recommendations for technical leadership: first, conduct an immediate inventory of agentic systems in OT and IT environments with an action and criticality matrix. Second, deploy model‑independent policy enforcement points on every tool exposed to an agent with default‑deny and human‑in‑the‑loop for critical operations. Third, establish a centralized kill switch with token revocation and network quarantine tested in quarterly chaos engineering exercises.</p>
<p>Fourth, enforce immutable append‑only logs with hash chaining and integrity alerts in the SIEM alongside legally defensible retention. Fifth, align governance with the Cyber Assessment Framework, GDPR Article 32, and EU AI Act Article 6; maintain a single register with an owner and annual review. Sixth, prioritize patching based on active exploitation according to CISA KEV and NCSC advisories, and enforce OT/IT segmentation with allow‑listing of remote management.</p>
<p>Conditions and limits: many OT systems do not support patching without downtime; Cyber Essentials does not stop nation‑state actors with zero‑days; behavioral detection requires tuning and continuous threat intel; human approval does not scale to thousands of trivial actions; secrets outside the context add operational complexity; hosted model providers may not offer private deployment; investment in guardrails and auditing incurs latency, storage, and specialized personnel costs.</p>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;NCSC and ANSSI: Agentic AI and Ransomware in Critical Infrastructure&quot;,
  &quot;summary&quot;: &quot;NCSC warns: AI will scale ransomware in critical infrastructure before 2028; ANSSI defines seven principles to govern agentic systems with deterministic authorization and immutable audit.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-08T06:01:35.034+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-03&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-08-am-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-08-am-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;AI&quot;,
    &quot;Artificial Intelligence&quot;,
    &quot;Compliance and Regulations&quot;,
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;English&quot;,
    &quot;Malware&quot;,
    &quot;Supply Chain&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://www.ncsc.gov.uk/news/cyber-chief-uk-facing-ai-powered-attacks-on-critical-infrastructure-in-two-years&quot;,
    &quot;https://universconvergents.com/2026/11/05/anssi-essentiels-securiser-systeme-ia-agentique&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-08-am-en/">NCSC and ANSSI: Agentic AI and Ransomware in Critical Infrastructure</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
		<item>
		<title>OpenAI Launches GPT-6 Cyber and Autonomous Pipeline Finds 948 LPE Exploits in Linux</title>
		<link>https://falifuentes.com/rafael-fuentes-2026-11-07-pm-en/?utm_source=rss&#038;utm_medium=rss&#038;utm_campaign=rafael-fuentes-2026-11-07-pm-en</link>
		
		<dc:creator><![CDATA[Rafael Fuentes]]></dc:creator>
		<pubDate>Sat, 07 Nov 2026 19:11:54 +0000</pubDate>
				<category><![CDATA[AI]]></category>
		<category><![CDATA[Artificial Intelligence]]></category>
		<category><![CDATA[Cybersecurity]]></category>
		<category><![CDATA[Cybersecurity News]]></category>
		<category><![CDATA[English]]></category>
		<category><![CDATA[MFA]]></category>
		<category><![CDATA[OpenAI]]></category>
		<category><![CDATA[artificial intelligence]]></category>
		<guid isPermaLink="false">https://falifuentes.com/rafael-fuentes-2026-11-07-pm-en/</guid>

					<description><![CDATA[<p>OpenAI releases an autonomous pipeline that discovers 948 privilege escalation exploits in Linux and launches GPT-6 Cyber, a critical model under Daybreak Red access.</p>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-07-pm-en/">OpenAI Launches GPT-6 Cyber and Autonomous Pipeline Finds 948 LPE Exploits in Linux</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></description>
										<content:encoded><![CDATA[<section lang="en">
<p><a href="https://falifuentes.com/rafael-fuentes-2026-11-07-pm-es/" hreflang="es">Leer este artículo en español</a></p>
<p>OpenAI has published details of an autonomous pipeline that, in a single run in mid-October 2026, discovered, validated, and generated patches for 948 local privilege escalation exploits in the Linux kernel, at a cost of $1.2 million in API credits. The announcement, accompanied by the launch of GPT-6 Cyber — a specialized cybersecurity variant with a Critical designation — marks a paradigm shift in vulnerability research at industrial scale, according to the <a href='https://openai.com/index/a-thousand-fewer-paths-to-root'>OpenAI technical publication</a>.</p>
<h2>Autonomous pipeline: from discovery to patching in days</h2>
<p>The pipeline integrates scout and analyst agents that identify relevant code, scan the kernel, reproduce findings in isolated environments with KASAN, generate functional PoCs, assess exploitability, and produce candidate patches before human review. In the October run, 2,308 candidate findings were reduced to 1,724 dynamically reproduced vulnerabilities (~74.7%) in 30 hours, at a cost of $73,330.</p>
<p>Of those, 291 PoCs were information leaks and 948 functional and distinct LPE exploits, exploitation cost $775,929 (average $359 per successful exploit). Total run cost: $1,213,430 in API credits, ~$1,279 per LPE, according to the OpenAI technical publication.</p>
<p>Within the ten days of the scan, kernel maintainers reacted quickly: 61 LPEs already patched upstream, 43 with patches in staging, and 299 in public discussion, demonstrating that mass disclosure accelerates community response.</p>
<h2>Exploit economics: cost per LPE drops to $1,279</h2>
<p>Cost breakdown: discovery (analysts 12h) $316,549, dynamic validation $73,330, exploitation $775,929. Compared to historical bug bounties of $20,000–$70,000 per LPE, the defensive ROI is 16x–55x, an economic window favorable to the defender.</p>
<p>In kernelCTF (hardened), 25 info-leaks and 21 zero-day LPEs in 112h for ~$32,936, ROI 45x versus 26 LPEs accepted in all of 2026. The economics suggest that less-resourced actors will replicate the capability soon.</p>
<p>OpenAI confirms that more capable models find vulnerabilities that earlier models missed in the same code and harness: GPT-5.6 Sol (July 2026) 166 LPEs for $513,690; October cyber model 948 LPEs for $1,213,430.</p>
<h2>Novel techniques and validation in hardened environments</h2>
<p>Sampling of 166 exploits verified by external experts reveals at least 10 new techniques: 2 persistent elastic objects, 2 page grooming, 5 data-only, 1 arbitrary type confusion. The exploitation surface evolves faster than known mitigations, reducing the effectiveness of signature-based defenses.</p>
<p>Against kernelCTF, 21 confirmed zero-day LPEs. In ExploitGym Hardened success 36.94% vs 48.45% unhardened. GPT-6 Cyber outperforms Astra: SEC-Bench Pro 87.02% pass@1, SRE-Bench 92.84%, ExploitGym 48.45%.</p>
<p>The pipeline detected and exploited a vulnerability in QEMU TCG (patched upstream but not in the build used), demonstrating that validation infrastructure is attack surface and must be audited before being used for offensive research.</p>
<h2>GPT-6 Cyber: critical capabilities under controlled access</h2>
<p>GPT-6 Cyber, a variant of GPT-6 Astra with post-training on authorized offensive/defensive tasks (vulnerability discovery, exploit development, web pentesting). Critical designation in Cybersecurity under the Preparedness Framework (same as Astra), High in Biological/Chemical, below High in AI Self-improvement.</p>
<p>External evaluations Irregular: CyScenarioBench 11/11 challenges (61% average vs 8/11 Astra), FrontierCyber 90/288 (0 successes on fully hardened, 0/15 highest-difficulty), Atomic 22/22 (97% vuln research, 98% network sim, 63% evasion). Honeypot exploitation rate 0.07% without instructions (vs 49.5% GPT-5.6 Sol), 0% with instructions.</p>
<p>Restricted access via Daybreak Red: identity verification, SSO, MFA, IP allowlisting, Private Safety Processing mandatory for Zero Data Retention API, high-risk assistance block (industrial controls, essential services, election systems, weapons) validated with Dragos. Conservative biological mode forced for all users, per the <a href='https://deploymentsafety.openai.com/gpt-6-cyber'>GPT-6 Cyber System Card</a>.</p>
<h2>Implications for defenders: mandatory re-scanning and hardening</h2>
<p>The bottleneck shifts to prioritization, exploit path validation, and patch coordination. Organizations must plan continuous re-audits aligned with frontier model releases, with dynamic validation harness (KASAN, fuzzing, isolated VMs) and triage at scale.</p>
<p>Hardening remains the most effective defense: SLAB_VIRTUAL, V8 sandbox, KASLR, SMEP/SMAP, HARDENED_USERCOPY, stack canaries, CET, PAC/BTI reduce success to 36.94% in ExploitGym Hardened. But 10 novel techniques in partial sampling imply that mitigations based on known patterns lose efficacy.</p>
<p>Mandatory to verify causality in autonomous exploits (20/200 in early run used a different vulnerability), isolate testing/CI/CD infrastructure, define internal patching SLAs in days not weeks. Monitor kernelCTF, Pwn2Own, Patch the Planet as leading indicators of AI offensive capabilities.</p>
<h3>Applicable recommendations</h3>
<ul>
<li>Implement automated re-scanning of critical code (kernel, browsers, parsers, TCP/IP stack) after each generational leap in frontier models, with dynamic validation harness and triage at scale.</li>
<li>Adopt an AI-assisted patching pipeline with mandatory human review (root-cause, minimal PoC, candidate patch, full exploit), accepting that AI patches require iteration and adversarial review.</li>
<li>Harden high-value targets by enabling SLAB_VIRTUAL, V8 sandbox, KASLR, SMEP/SMAP, HARDENED_USERCOPY, stack canaries, CET, PAC/BTI.</li>
<li>Isolate and audit testing/CI/CD infrastructure (hypervisors, emulators, runners, sandboxes) before using it for exploit validation.</li>
<li>Verify causality in autonomous exploits with fine-grained instrumentation and define give-up conditions to limit cost of failed attempts.</li>
<li>Evaluate access to Daybreak Red / GPT-6 Cyber for internal red-team / vulnerability research teams, meeting strict organizational requirements.</li>
<li>Monitor bug bounty programs (kernelCTF, Pwn2Own, TyphoonPWN, Patch the Planet) as leading indicators of AI offensive capabilities.</li>
<li>Invest in secure-by-default code generation, surface reduction, and formal verification.</li>
<li>Define internal patching SLAs aligned with mass disclosure speed (days, not weeks).</li>
</ul>
<h3>Conditions and Limits</h3>
<ul>
<li>The API cost of a full kernel run exceeds $1.2 million; not viable for all organizations.</li>
<li>AI-generated patches may introduce incompleteness, performance regressions, or deviations from style conventions; they do not replace maintainer review.</li>
<li>Hardening reduces but does not eliminate risk: GPT-6 Cyber achieves 36.94% success rate in ExploitGym Hardened.</li>
<li>Access to GPT-6 Cyber only for organizations with SSO, MFA, managed devices, IP allowlisting, PSP, and legal attestation; not available to individual users.</li>
<li>Zero-day exploits against fully hardened targets remain difficult: 0/15 highest-difficulty in FrontierCyber, 0 successes against fully hardened targets.</li>
</ul>
<section class="article-reference">
<h2>Structured reference for readers and AI agents</h2>
<pre><code class="language-json">{
  &quot;format&quot;: &quot;rafael-fuentes-article/1.0&quot;,
  &quot;language&quot;: &quot;en&quot;,
  &quot;title&quot;: &quot;OpenAI Launches GPT-6 Cyber and Autonomous Pipeline Finds 948 LPE Exploits in Linux&quot;,
  &quot;summary&quot;: &quot;OpenAI releases an autonomous pipeline that discovers 948 privilege escalation exploits in Linux and launches GPT-6 Cyber, a critical model under Daybreak Red access.&quot;,
  &quot;author&quot;: &quot;Rafael Fuentes&quot;,
  &quot;publicationDate&quot;: &quot;2026-11-07T20:01:35.033+01:00&quot;,
  &quot;eventDate&quot;: &quot;2026-11-04&quot;,
  &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-07-pm-en/&quot;,
  &quot;translation&quot;: {
    &quot;language&quot;: &quot;es&quot;,
    &quot;url&quot;: &quot;https://falifuentes.com/rafael-fuentes-2026-11-07-pm-es/&quot;
  },
  &quot;topics&quot;: [
    &quot;AI&quot;,
    &quot;Artificial Intelligence&quot;,
    &quot;Cybersecurity&quot;,
    &quot;Cybersecurity News&quot;,
    &quot;English&quot;,
    &quot;MFA&quot;,
    &quot;OpenAI&quot;
  ],
  &quot;sources&quot;: [
    &quot;https://openai.com/index/a-thousand-fewer-paths-to-root&quot;,
    &quot;https://deploymentsafety.openai.com/gpt-6-cyber&quot;
  ]
}</code></pre>
</section>
</section>
<p>La entrada <a href="https://falifuentes.com/rafael-fuentes-2026-11-07-pm-en/">OpenAI Launches GPT-6 Cyber and Autonomous Pipeline Finds 948 LPE Exploits in Linux</a> se publicó primero en <a href="https://falifuentes.com">Fali Fuentes</a>.</p>
]]></content:encoded>
					
		
		
			</item>
	</channel>
</rss>
