On 21 October 2026, the NCSC urged UK organisations to address critical security flaws in PHP, software used to run websites and other applications. The flaws, CVE-2026-109618 and CVE-2026-109619, can allow an attacker to run commands on a server without signing in. PHP has released security updates for both.
If you run a business and somebody else looks after your website, I would turn that alert into a specific question: which of our systems use PHP, and what have you done about these flaws?
You do not need to learn server administration to ask it. You need an answer that identifies the affected services, the action taken and anything still outstanding. That gives you something to act on, rather than a general assurance that security is covered.
Find out who can answer
Start with whoever manages your website hosting or business applications. That might be your IT provider, web agency or an internal colleague. Ask them to confirm who is responsible for checking the underlying software and applying security updates.
If the agency manages the website and another company supplies the hosting, ask them to agree who will give you the complete answer. I would keep the request together in a support ticket so that a hand-off between suppliers does not leave part of it unanswered.
Include services you still use but rarely discuss: an old customer portal, a separate recruitment site or an internal application. Use these as prompts for discovery and ask the provider to establish what each runs on. The NCSC advises checking for PHP across the infrastructure, including software that bundles it.
Ask for evidence of the fix
I would send a request along these lines:
Please check our services against CVE-2026-109618 and CVE-2026-109619. Which systems use PHP, which are reachable from the internet, and what update or mitigation has been applied to each? Please include the date, the installed version or supplier advisory confirming the fix, and any remaining action with a named person responsible.
Ask for a separate answer for each service. If a provider says a service does not use PHP, ask them to record that too. It prevents the same question being reopened because nobody can find the conclusion.
For affected systems, I would follow this order. The graphic draws on the NCSC’s priority actions; the questions around it are how I would turn those actions into a business record.
- Find PHP installationsAsk the provider to identify affected services and their exposure.
- Apply security updatesGet confirmation of the fix, dealing with internet-facing services first.
- Track further vendor updatesAsk the provider to review subsequent advice for the software you run.
Do not make the version number your only test. PHP says some operating-system and long-term-support suppliers provide fixes for older releases. If that is the provider’s answer, ask for the supplier notice and confirmation that the fixed package is installed on your service. An unexplained version number is difficult for an owner to assess; a documented answer gives your IT adviser something to verify.
Also ask the provider to check PHP’s current security notices. These include follow-up configuration advice for an unusual additional attack path and updated Windows packages addressing a separate issue. I would ask the technical team to assess that advice against your setup, rather than copy settings into a hosting panel yourself.
Agree the business checks around an update
Tell the provider which functions matter to the business. For a website, I would name enquiry forms, bookings and customer sign-in where applicable. Ask how they will check those functions after the update, and who in your team will confirm that an enquiry or booking reaches the right place.
Agree how they will restore service if the change breaks something, including what they will do to avoid returning to an exposed setup. Keep testing proportionate, and ask the provider how it fits around the urgent security work. A small functional check should help confirm the change worked.
Keep the result with the support ticket: what changed, when it changed and what was checked. That is also useful when you next review the support arrangement or prepare evidence for Cyber Essentials.
If the provider cannot update
Ask for the reason, the interim action and a firm next decision. “We are looking into it” needs a follow-up time and somebody responsible for answering.
PHP says updating is the only complete solution to these flaws. I would ask the provider to explain the limits of any temporary measure and record how they will move to a fixed version.
The NCSC says that where updates and temporary mitigations cannot be applied promptly, organisations should consider taking the system offline. I would make that a conversation between the business owner and the technical provider: what must be restricted, what work it interrupts and how customers will be directed while it is unavailable.
Keep possible compromise separate from patching. An update confirmation alone would not answer a concern about earlier suspicious activity. If the provider finds something concerning, ask for an incident assessment and a clear explanation of what is known.
Start the request now if you have no recorded answer. The useful outcome is a checked list of services, evidence of the action taken and a decision on anything still exposed. You can ask for that without becoming the person who maintains the server.