Threat Intelligence

Threat Thursday: November 19th, 2026

Threat Thursday: November 19th, 2026

Welcome to Threat Thursday, Galactic’s weekly threat intelligence roundup.

Every Thursday we cover the cybersecurity stories that matter most for protecting organizations from emerging threats, and we break each one down into what happened, what it could mean for your organization, and what to do about it.

This week’s critical fixes were for software that’s easy to overlook. It often comes built into the products a company buys or runs on services someone else manages, so updating it starts with finding out where it lives. Each story below covers what to look for as well as how to fix it.

This Week’s Stories

01. Popular Web Server Software nginx Gets Fixes for Five Serious Takeover Flaws

nginx is free web server software that delivers a large share of the world’s websites and sits inside many cloud services, appliances, and hosting platforms. On November 18, the security firm Xint disclosed five serious flaws it calls NGINX2Root, and the nginx team released fixed versions the same day. The most dangerous, CVE-2026-123007, affects nearly every release since 2015 and lets an outsider run commands on the server without logging in, as long as a common speed feature called HTTP/2 is turned on. Researchers have already published code that demonstrates it. The other four flaws only apply to specific setups. Xint also showed that an attacker who gets in can pair that access with a separate, still unfixed Linux flaw to take full control of the machine.

Potential impact: Web servers are the front door most organizations leave open to the entire internet, so a flaw that needs no password is about as serious as it gets. The notable detail is how hard nginx can be to find: it’s often built into products that never mention it, and some Linux vendors apply fixes without changing the version number, so a quick version check can give the wrong answer. Xint also notes that this attack leaves little trace in normal web logs, which means an organization may never know it was hit.

What to do: Administrators should update nginx to 1.30.7 or 1.31.9, or to their Linux vendor’s fixed package, and fully restart it, since a simple reload keeps the old code running. Ask vendors whether their appliances or hosted services include nginx and when they’ll ship a fix. Where updating has to wait, turning off HTTP/2 blocks the most serious flaw, though it may slow some sites or break apps that depend on it.

Source: Xint NGINX2Root write-up

02. Hackers Broke Into N-able’s Email Filtering Service, Taking Its Admin Tools Offline

N-able sells IT management and security tools mainly to IT service providers, including two email filtering services, Mail Assure and SpamExperts, that screen messages for spam and malicious content before they reach the inbox. On November 16, customers lost access to the services’ admin tools, which N-able first blamed on a configuration problem. Two days later it confirmed that an unauthorized person had gotten into one of its servers, and it shut off the admin website and automated connections in the Americas, Asia-Pacific, and Europe. Email kept flowing and filtering kept working the whole time, but features like releasing held messages and searching logs were still offline as of November 19. N-able says it has seen no evidence that customer data or mailboxes were accessed, though its investigation is ongoing.

Potential impact: The service that screens your email before it arrives is something an attacker can target too, even when another company runs it. Most organizations would struggle to name every outside service that touches their mail, let alone say what each one can see. The immediate risk is practical: with held-message tools offline, staff may miss legitimate mail, and scammers like to send fake ‘release your messages’ emails during disruptions like this one. Whether the intruder saw anything beyond that single server is the open question, and the answer depends on N-able’s final findings.

What to do: Organizations that use Mail Assure or SpamExperts, directly or through an IT provider, should look for N-able’s notice and follow its guidance, or contact N-able support if no notice arrived. Set up a manual way to request held messages until access returns, then review admin accounts and filtering rules for changes. Treat any email about releasing held messages or resetting a password as suspect until it’s confirmed through a known contact.

Source: The Hacker News

03. PHP, the Language Behind Millions of Websites, Fixes Seven Critical Flaws

PHP is the programming language behind a huge number of websites, including every WordPress site. This week the PHP project released fixes for seven critical flaws, each rated 9.2 out of 10, in versions 8.2.36, 8.3.37, 8.4.28, and 8.5.13. The flaws could let an attacker run their own code on a website’s server, though most only work under certain conditions, such as a particular way the web server hands requests to PHP. One of them, CVE-2026-129496, can be triggered by anyone on the internet without needing an account. None are known to be under attack yet. This is the second batch of critical PHP fixes in about a month, after two flaws in October drew attack traffic around the time they were announced.

Potential impact: Few organizations know exactly how many PHP sites they run, because PHP comes along with marketing sites, customer portals, plugins, and hosting plans that different teams set up over the years. Each copy has to be updated separately, and hosting companies update on their own timeline. There’s also a deadline worth noting: PHP 8.2 stops receiving security fixes on December 31, so sites still on it won’t get a fix for the next critical batch, and with two in two months, betting on a lull would be a gamble.

What to do: Start by listing every website and app that runs PHP, including ones run by hosting companies or outside agencies, then update each to the fixed release for its version, and restart the web service so the update takes effect. Ask hosting providers when their update will land. Sites still on PHP 8.2 should move to a newer version before the end of the year.

Source: The Hacker News

04. Attackers Start Probing Java4Shell, a Java Flaw Hidden Inside Many Business Apps

Java is software that many business applications run on behind the scenes, from project trackers to sign-in systems to network management tools. On November 17, the research firm Calif disclosed a flaw it calls Java4Shell (CVE-2026-114938), and Oracle and the Java open source project released fixes the same day. The flaw only affects computers with certain newer Intel and AMD server processors, and it lets an attacker crash the application or, with extra work tailored to each product, run their own code. Calif showed it working without a login against popular tools including Jira, Confluence, Jenkins, and Keycloak. Security firms saw attackers testing it within a day, though no successful break-ins have been confirmed.

Potential impact: What makes this one hard is that most organizations never installed Java on purpose. It arrives built into products from Cisco, Atlassian, Elastic, Siemens, and many others, and the fix has to come from each of those vendors, several of which haven’t released one yet. That leaves a stretch where the flaw is public and attackers are probing, but the affected systems can’t be updated directly. The organizations that move fastest will be the ones that already know which of their products carry Java inside.

What to do: Defenders should find every system running Java, including Java built into vendor products, and check whether those systems use the affected processors. Apply the November Java updates where Java is installed directly, and follow each vendor’s advisory for products that bundle it, using their temporary safeguards until a fix ships. Limit internet access to tools like Jira, Confluence, Jenkins, and Keycloak until they’re updated.

Source: SecurityWeek

05. Optimizely Website Tool Hijacked to Steal Card Details on 178 Organizations’ Sites

Optimizely is a tool companies use to test different versions of their web pages, and it works by loading a small piece of code on every page. On November 16, an attacker abused a flaw in Optimizely’s system to slip malicious instructions into that code for many customers, so it reached visitors through Optimizely’s normal delivery network. Security researcher Scott Helme of Report URI found infected code tied to 178 organizations, including Zoom, CNN, the BBC, HSBC, and Salesforce. The code tried to swap cryptocurrency wallet addresses, capture card numbers typed into some checkout forms, take screenshots, and steer iPhone and iPad users toward hacking tools. The infected files were cleaned up by the end of that day, and Optimizely says it has fixed the flaw.

Potential impact: None of these websites were hacked in the usual sense, which is exactly the problem. A modern website loads code from analytics, chat, advertising, and testing services, and each of those companies can change what runs on the page without the site owner touching anything. Most organizations don’t keep a list of these outside scripts or watch them for changes. For an affected business, the harm lands on its customers, and stolen card details or drained crypto wallets can still bring questions from card companies, customer notices, and damage to its reputation.

What to do: Organizations that use Optimizely should review their project change history for November 16 and check what code their pages served that day, since a clean file today doesn’t show what visitors received then. Every organization should keep a list of the outside scripts on its website, limit who can change them, require multi-factor authentication (a second sign-in step) on those accounts, and set browser rules that only allow code from approved sources.

Source: Hackread

06. Suspected Chinese Spies Used a VMware Flaw for Months Before a Fix Existed

VMware vCenter is the control panel organizations use to run their virtual servers, the software-based machines that host many core business systems. On November 19, Broadcom, which owns VMware, fixed a flaw (CVE-2026-123890, rated 9.8 out of 10) that lets someone who can reach vCenter on the network skip the login and gain full administrator access. The same day, Google’s threat intelligence team and Mandiant reported that a suspected China-linked espionage group had used the flaw since early September, roughly two months before any fix existed. The group got in through a remote support device from another vendor, BeyondTrust, moved to vCenter from inside the network, then created hidden administrator accounts and stole the keys used to verify logins. Older vCenter 7 systems, no longer on regular support, can be affected too.

Potential impact: What stands out is that being up to date didn’t help: the targeted server already had July’s vCenter fix, and the attackers didn’t need the internet to reach it because they were already inside. Whoever controls vCenter can copy, shut down, or encrypt every virtual server at once, which is why both ransomware crews and spies go after it. Older vCenter 7 systems are easy to overlook because they keep working, but they’ve fallen out of the normal update cycle.

What to do: vCenter owners should install Broadcom’s fix right away (vCenter 8.0 U3m, 9.0.2.0200, 9.1.1.0200, or Cloud Foundation 5.2.5), reading Broadcom’s upgrade notes first. Limit which computers on the internal network can reach vCenter at all, and look for unexpected new administrator accounts. Anyone running vCenter 7 should plan an upgrade or ask Broadcom about an extended support fix. If signs of misuse turn up, treat it as a full incident rather than just rebuilding the server.

Sources: Mandiant and Google Threat Intelligence, The Hacker News

07. Phishing Emails Posed as Microsoft 365 Outage Alerts During Last Week’s Disruption

Microsoft 365, which runs email, Teams, and file sharing for many organizations, went down worldwide for more than 22 hours starting November 12. While Microsoft was still restoring service, scammers launched a large phishing campaign, meaning fake emails built to trick people into clicking a link or handing over a password. The messages looked like official service alerts, claimed that emails held during the outage would be deleted within 24 hours, and pushed a big ‘Restore Messages’ button. Each link was customized to the person who received it, and the emails carried hidden text designed to slip past spam filters. Microsoft published details on November 16 and said it hadn’t directly observed anyone entering a password or losing control of an account.

Potential impact: Scammers follow the news as closely as anyone, and a real outage hands them a ready-made story that employees half expect to see in their inbox. Timing matters too: during an outage, people are frustrated, help desks are swamped, and a message promising to recover lost email feels helpful rather than suspicious. A single password typed into the wrong page can open up email, files, and chats, which is a common starting point for invoice fraud. Expect the same trick after the next major outage at any large provider.

What to do: IT teams should search email for these fake outage notices, remove any that were delivered, and check whether anyone clicked. Anyone who entered a password should have it reset and be signed out of every device. Remind staff that service updates belong on the official status page or with the IT team, and that any email demanding fast action to recover messages deserves a second look.

Source: Infosecurity Magazine

The big picture

An inventory of the products an organization bought is a good start, and this week showed how much can sit underneath it. When attackers hijacked Optimizely’s system, 178 organizations ended up serving malicious code on their websites without anyone on their side changing a thing, because the code came from a service they’d already invited onto every page.

The habit that helps is taking that inventory one layer deeper before the next advisory lands. Vendors can usually tell you which open source components their products carry and how they’ll let you know when one needs a fix, but it tends to take someone asking. The same question works for the outside services that load code onto your website, and for any IT provider managing systems on your behalf. Anything nearing the end of vendor support, like PHP 8.2 at the end of this year, belongs on that list with a replacement date attached.

Make sure to check back here each week for another Threat Thursday update. See you then!

Back to All Articles

Copy Link

See where your own environment stands.

A recurring penetration test finds this before an attacker does.

Schedule a Demo