Three critical security patches in a month.
PHP2Shell, NGINX2Root and a new critical PHP flaw, all in a month. Here’s how we kept our clients’ websites protected, and what to ask your own agency.


If you’re one of our clients, you’ll have heard from us a fair bit over the last month. Each time, we got in touch before we started patching, kept you updated while we worked, and confirmed once your sites were done.
That’s because it’s been a busy few weeks. Three serious security flaws have gone public in software that sits underneath a huge share of the world’s websites, including many of the ones we build and look after. Each one meant moving quickly to protect client sites before attackers could take advantage.
Here’s what happened, what we did, and what it means for your website.
First, a bit of background
You don’t need to know how any of this works under the bonnet, but two names come up a lot below.
PHP is the programming language behind a large part of the web. WordPress runs on it, and so does Laravel, the framework behind a lot of our bespoke builds.
nginx (pronounced “engine-x”) is the software on a web server that hands your pages to visitors. It runs tens of millions of websites worldwide.
When a serious flaw turns up in either of them, it doesn’t affect one website. It affects millions at once.
October: PHP2Shell
Last month a flaw nicknamed PHP2Shell was announced. You will have heard about this in the mainstream news, with many sites on the internet being taken offline for hours (some days) while patching took place. In plain terms, PHP2Shell could let someone anywhere in the world run their own code on a website’s server, without needing a password. Reports after the patch was released soon followed of attackers already trying it.
The fix was released at 3pm. We let clients know what was happening straight away, then got to work and didn’t stop. The team carried on through the evening and into the night until all 42 servers across our clients were patched, sending updates as we went and confirmation once each site was done. We found no sign that any had been exploited.
Patching on a brand new server running the latest versions of everything is a simple fix. Patching on old hardware that may be running old versions of operating systems, or on hardware we may not have direct access to was far more tricky. But we got through it, and with no vulnerabilities exploited.
This week: NGINX2Root
Yesterday, security researchers published details of five flaws in nginx, together known as NGINX2Root. Chained together, they could let an attacker go from simply visiting a website to taking complete control of the server it runs on.
We started patching straight away. Patching is still underway with the vast majority of client servers fully protected.
This afternoon: PHP again
PHP is releasing another security update between 2pm and 2:30pm today, fixing a flaw it has rated critical. We’re ready, and the team will start rolling it out to client sites as soon as it lands.
Why does this keep happening?
2026 has been an unusually busy year for this. Since May alone there have been several critical flaws found in nginx and PHP, and some of them had sat unnoticed in the code for more than 15 years.
The uncomfortable part is what happens once a flaw goes public. The announcement tells developers what to fix, but it also tells attackers where to look. From then on it’s a race between the people patching and the people scanning the internet for sites that haven’t been patched yet.
That’s why speed matters. A fix applied a week later is a week of leaving the door open.
What we learned: you can only patch what you can reach
The PHP2Shell night taught us something worth sharing.
Servers we manage directly were quick and easy to patch. We know exactly how they’re set up, we have the access we need, and we can roll a fix out in minutes.
The harder ones were sites hosted elsewhere. Some sat with third-party hosting companies, some were on shared hosting where we didn’t have full control, and some needed login details or sign-off from someone outside our team. In those cases, getting access took far longer than the fix itself.
None of that is anyone’s fault. It’s how a lot of websites end up being set up over the years. But in the middle of the night, with a live threat, it’s the difference between being protected and waiting.
It’s worth checking
Keeping your website maintained is about more than keeping it running. If you’re responsible for your organisation’s website, ask whoever looks after it these questions:
- Who is responsible for security updates? Not just in principle, but who actually applies them, and how quickly.
- Do they know where your website is hosted? And do they have the access they’d need to fix it in a hurry?
- Would they know about a critical flaw the day it’s announced? Or would they only find out when something breaks?
- Would they tell you? You should hear from them before, during and after a fix, not from the news.
If the answers are vague, have that conversation before the next one comes along. Because there will be a next one.
A note on Cloudflare
When we onboard a client with us one of the first things we recommend is to put the site behind Cloudflare. Cloudflare is a service which sits between your server(s) and the internet. All traffic goes through Cloudflare, and they have a suite of tools to help us manipulate that traffic (redirects, speed improvements, etc). But one of the main advantages is that Cloudflare is able to identify a lot of these attacks, intercept and stop them before they even reach your server.
Those clients using Cloudflare are typically protected before and during patching. Given these incidents are typically a race to get patched before the exploit attackers hit you, that is a great situation to be in. Using Cloudflare doesn't remove the need to patch, but it does take the pressure off and adds a substantial level of security.
Unsure if your website is at risk?
We’re here to help. Whether we already look after your site or you’d like a second opinion on how it’s set up, get in touch and we’ll talk you through it.
Insights.






