,

Patch diffing explained in 2 minutes

What it is. Patch diffing means comparing a piece of software before and after a security fix (a patch) to see exactly what changed, and from that, work out what the flaw was.

An analogy. Imagine a shop quietly adds a new lock to one back window. Anyone who notices which window got the lock now knows which window used to be easy to open, and that every shop with the same building plan, but no new lock, still has the weak window.

Why it matters. Software makers have to publish fixes so people can install them. But each fix also hints at where the problem was. Anyone who has not yet updated is exposed during the gap between the fix appearing and the update being installed. The faster attackers can read a fix, the shorter that safe window becomes. Security researchers have done patch diffing by hand for years; it used to take skill and time, which gave defenders a head start.

Why it showed up in the news this week. On 16 November, the Google Threat Intelligence Group reported finding a tool, called exploit_agent, that used AI models to do this automatically. According to Google, it compared versions of popular web software before and after each patch, worked out the likely flaw, wrote an attack and ran it, with no person involved. Google’s team said in its previous quarterly report that it had not yet seen fully autonomous attack pipelines like this in use.

One thing to remember. When a security update comes out, the clock starts for attackers too. Installing updates quickly, especially on anything connected to the internet, matters more than it used to. For organisations, that means treating web software updates as urgent rather than waiting for the next scheduled maintenance day.

Sources: AIooze brief on Google’s findings; Google Threat Intelligence Group, 16 November 2026


Related