Security

Business as usual, or something special?

Every so often a vulnerability arrives that makes the whole industry look up from its dashboards. I wrote about this during the Heartbleed days, for an internal audience. PHP2Shell gives me a reason to revisit the same question, and this time I am writing it for everyone.

What happened

On 21 October 2026, the PHP project shipped emergency patches for two critical flaws, CVE-2026-109618 and CVE-2026-109619. Together they are called PHP2Shell, and both carry a CVSS v4.0 score of 9.2. The fixed releases are PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35.

Attackers did not wait for the paperwork. F5 Labs recorded thousands of exploit attempts within hours of the disclosure.

Is this not just BAU?

Here is the question I keep coming back to. If a vulnerability is well described, and a team can either patch or apply a workaround inside the required timeframe, isn’t that business as usual?

Mostly, yes. We already have machinery for this. Severity sets the urgency, internet-facing systems go first, and the change rides the normal window. A well-described vulnerability with a fix is exactly what that machinery was built for.

PHP2Shell fits that description in most respects. The advisory is clear, fixed versions exist, and the affected branches are known. So where does it stop being routine?

Where it stops being routine

Heartbleed and PHP2Shell share the same shape. Both are pervasive. PHP is not necessarily a product you can point to in an inventory. It is the engine underneath, embedded in things that never show up in a normal search: hosted sites, appliances, container base images, and old internal tools that still work, so nobody touches them.

Two details make this worse. First, updating WordPress or its plugins does not fix it, because the flaw sits in PHP itself. Updating PHP does. Second, end-of-life products may never receive a patch, which means no fixed version to move to and no product-specific guidance to follow.

Once you reach that point, routine patching is not enough. The organisation needs a programme that does three things:

  1. Builds an understanding of where the vulnerability could exist across every platform, including the ones nobody remembers.
  2. Turns that understanding into action, with a named owner for each place and a fast way to reach them.
  3. Plans for the systems that will never get a fix, through isolation, mitigation or retirement.

Vendor guidance still matters. It just becomes one input to that programme, rather than the whole plan.

Questions to ask this week

  • Do we know every host that runs PHP, including the ones we did not build?
  • Which of those face the internet, and were they updated first?
  • Are developer machines and CI base images pinned to a vulnerable PHP version?
  • For anything end-of-life, what is the plan, and who signed it off?

Further reading

Business as usual is a fine place to live, right up until the usual stops working. PHP2Shell is a reminder that patching is the easy part once you know where the thing is. Finding it is the hard part.

Leave a Reply