< BACK

Your Cyber Attacker Doesn’t Want to Be Clever. They Want to Be Repeatable.

repeatable hackers cyber attacker Emerge Managed IT Services Greater Cincinnati NKY

Key Takeaways

  • The most common way cyber attackers get in right now is not a sophisticated exploit. It’s a fake verification page that talks a person into pasting a command into their own computer, a technique called ClickFix, responsible for nearly half of the access attempts Microsoft tracked last year.
  • Once inside, attackers increasingly rely on tools already installed on the machine, the same administrative software your own IT team uses, rather than bringing in anything new.
  • Both techniques succeed for the same reason: they work the same way for every victim, letting a criminal group run the same playbook thousands of times without adapting it.
  • This is also why AI has not reshaped ransomware the way many expected. A tool that improvises produces a different result at every target, which is the opposite of what a repeatable, franchise-style criminal operation needs.
  • The most important defense isn’t a smarter tool. It’s making sure someone is actually watching what your existing tools already catch.

The Break-In That Doesn’t Look Like a Break-In

A recent analysis from The Hacker News makes a case worth sitting with: the attackers causing the most damage right now are not the ones with the best tools. They’re the ones who found something that works the same way at every company they target, and they run it as many times as they can.

The clearest example is a technique called ClickFix. A web page tells a visitor to prove they are not a robot. While the instructions display, the page quietly copies a command to the visitor’s clipboard. The page then walks the visitor through opening a terminal and pasting it in themselves. By Microsoft’s own tracking, this was the single most common way attackers gained initial access last year, accounting for nearly half of the incidents its team observed.

Notice what’s missing from that description. There’s no malicious attachment to scan. There’s no software vulnerability to patch. The entire technique depends on a person following instructions, which is the one weakness present in every organization on earth, with no patch available.

Once They’re In, They Use What’s Already There

The pattern continues after the initial break-in. Rather than installing custom malware, repeatable cyber attackers increasingly rely on legitimate administrative tools already on the machine: scripting engines, remote management utilities, and built-in system software your own IT team uses every day. Recent analysis of high-severity security incidents found that the large majority involved exactly these kinds of pre-existing, legitimate tools rather than anything an attacker brought with them.

This isn’t primarily about evading detection, although it can have that effect. It’s about repeatability. An attacker who learns a sequence of built-in commands once can run that same sequence against the next victim without adapting it to a different piece of software or environment. Nothing needs to be rebuilt. Nothing needs to be tested against an unfamiliar system.

Why This Explains the AI Question Too

This framework also explains something that has puzzled a lot of security teams: why hasn’t AI transformed ransomware the way it has transformed other industries?

The answer comes down to economics and structure, not capability. Scanning for a known vulnerability and running a public exploit against it costs an attacker almost nothing. Adding an AI model into that process adds real cost to every single attempt, in a business model built entirely around driving that cost toward zero.

There’s a deeper structural reason too. Ransomware largely operates on an affiliate model, where a criminal organization writes a procedure once and expects it to produce the same result in the hands of many different operators. An autonomous AI tool that improvises its way into a network does something different at every target. That variance is a liability to a business whose entire advantage comes from consistency, not an asset. AI is more likely to show up behind the scenes, helping develop and refine a playbook offline, rather than improvising live against a victim.

What Actually Helps Defend Against Repeatable Cyber Attacks

The encouraging flip side: a standardized attack is a standardized defense problem. You’re not defending against unlimited creativity. You’re closing a specific, limited set of doors that a repeatable playbook depends on.

  • Patch the exposed, remotely exploitable stuff fast. The gap between a vulnerability going public and a working exploit landing on the internet is often just days. Closing that window matters more than trying to patch everything.
  • Restrict what can run, not just what can arrive. Application control and script execution policies interrupt techniques like ClickFix at the exact moment a pasted command tries to become a running process.
  • Limit who can use your own administrative tools. Most employees have no legitimate reason to invoke the remote management and scripting utilities attackers rely on. Deciding who can is a policy choice, not a technical limitation.
  • Look at events together, not one at a time. A remote tool running is normal. An account signing in is normal. Data moving to cloud storage is normal. Three of those things happening on the same host, in that order, in a short window, is not. Catching that pattern requires correlation, not a single alert.
  • Make sure someone’s actually watching. The most common failure in real incident investigations isn’t missing technology. It’s technology that raised the right alert to nobody with the authority to act on it in time.

That last point is where many organizations quietly fall short, and it’s the same theme behind why insurers are now underwriting based on detection and response speed rather than prevention alone (we’ve written about that shift here). It’s also connected to a risk we’ve flagged before around unsanctioned AI tools spreading inside organizations, since every new tool your team adopts, sanctioned or not, is one more thing that needs to be watched, not just installed.

None of this requires outsmarting a criminal organization. It requires closing the small number of doors their business model depends on, and having someone actually paying attention when they try the handle.

Scroll to Top