SCARY: “Atom Bomb” Windows Security Hole said to be Unfixable
Windows has an unfixable security flaw. That’s the frightening conclusion of this researcher, who says he can inject code—at will—into browsers and other Windows apps.
It affects all versions of Windows released in the past 16 years, he says. And it can’t be patched, without breaking legitimate desktop apps. It could be bad for containerized server workloads, too.
But is it actually a big deal? In today’s IT Newspro, we loosen the airtight hatchway.
What’s the story? Richard Chirgwin brings us Windows Atom Tables popped by security researchers:
A security researcher has found a way to…inject code into Atom Tables [in] all versions of Windows. … A successful attack could…accomplish quite a lot of evil [by] snooping on the contents of memory…screen-grabs and browser hijack.
But, uhh, Defender will help, right? Jai Vijayan notes that it doesn’t exploit a vulnerability:
[It] is undetectable to existing anti-malware tools. … enSilo’s AtomBombing attack involves the injection of malicious code into the so-called atom tables [in] Windows. … The oldest version…that uses atom tables is Windows 2000.
What does the researcher have to say for himself? Tal Liberman claims it’s a Brand New Code Injection for Windows:
Here’s a new code injection technique. [It] exploits Windows atom tables and Async Procedure Calls.
[We] copy our code to an RW code cave in the target process. … Then use a meticulously crafted…Return Oriented Programming…chain to allocate RWX memory.
Not a big problem. … We can use ZwAllocateVirtualMemory. … The complete implementation can be found on GitHub.
Oh noes. What does Redmond have to say? A spokesperson effected this statement to Charlie Osborne:
We encourage our customers to practice good computing habits online, including exercising caution when clicking on links…opening unknown files, or accepting file transfers. A user’s system must already be compromised before malware can utilize code-injection techniques.
Wait, are you suggesting it’s a lot of fuss about nothing? Ken Hagan seems to agree:
There is nothing you can attack that doesn’t have exactly the same access. … So you can only attack processes that you can already control.
So it isn’t really a security vulnerability. … I have to assume that this is just someone trying to drum up some publicity.
Phew, panic over? Catalin Cimpanu isn’t quite so sanguine Microsoft can’t patch against AtomBombing technique:
The bad news is that this is a design flaw. … Microsoft can’t patch it without changing how the entire OS works.
The last word on the subject? It sounds like something Raymond Chen would counter with It rather involved being on the other side of this airtight hatchway. Lasse V. Karlsen explains:
If you have a cracker that is able to execute code…on your own machine, you have already lost. … So no, this is not a vulnerability.
Now, having said that…there is support for attempts to at least make it harder. [But] if the bad-guy can execute code on your platform, what is there to stop him?
Main image credit: @Tal_Liberman