That code doesn't even have to be part of the normal scanning process, let alone a "signature". It could be an update for the memory-resident part that looks for interesting documents in the background, when the computer is not in use.
-
-
Replying to @VessOnSecurity @KyleHanslovan and
Yes, AV programs are privileged and can do a lot of nasty stuff, if they decide to. But doing it by modifying a scan string is a stupid, inefficient, and ineffective way to do it - and nobody would do such a stupid thing, even if they wanted to do mischief.
1 reply 0 retweets 3 likes -
Replying to @VessOnSecurity @KyleHanslovan and
I could not disagree more with this statement. Moding a sig in memory seems to be one of the more stealthy ways to do this IMO. Make the signature silent and alter the exfil location to something you control and you have a pretty infallible and undetectable collection system.
1 reply 0 retweets 3 likes -
Replying to @flyryan @KyleHanslovan and
Scanners update themselves. It doesn't matter if they are going to change one signature or some code component - it's equally normal (or stealthy). So it's not worth it doing it like this at the price of only catching a small subset of documents.
1 reply 0 retweets 2 likes -
Replying to @VessOnSecurity @KyleHanslovan and
You are taking a very narrow use case and treating it as if it’s the ONLY use. He showed you could manipulate signatures in memory to gather things you care about. Just because he only targeted an RTF doesn’t mean it can’t be used otherwise. Patrick was just proving a concept.
2 replies 0 retweets 2 likes -
Replying to @flyryan @KyleHanslovan and
No, you are missing the point. His method is limited to what he did precisely because scanners don't work as he seems to think they do. He CANNOT "modify a signature" to "make the scanner detect a string anywhere in any document".
1 reply 0 retweets 2 likes -
Replying to @VessOnSecurity @KyleHanslovan and
You’ve laid out a lot of “CANNOT”s mixed in with “stupid way to do it”s (which conflict with each other) but you haven’t discussed the reason WHY you think he is missing the mark. What did Patrick miss if what he’s shown doesn’t actually work?
1 reply 0 retweets 1 like -
Replying to @flyryan @KyleHanslovan and
The two are related. Scanning the whole file (while doing normal scanning) is a STUPID way to proceed (regardless of the goal) which means that a scanner CANNOT use this approach to look for secret documents - because people will stop using the scanner (due to the slowdown).
2 replies 0 retweets 2 likes -
Replying to @VessOnSecurity @KyleHanslovan and
In the classified document use case, it doesn’t have to scan the whole document. Just the header. Even so, if it doesn’t scan the whole document, how does it detect malicious content embedded within? I’ve read
@matalaz’s book on AV hacking and what you’re saying doesn’t jive…2 replies 0 retweets 1 like -
Replying to @flyryan @VessOnSecurity and
Hi Ryan! The troll
@VessOnSecurity got me blocked but, nevertheless: he is wrong, as always. Scanning the whole file is common and finding a string anywhere in a file is common. It's usually done by finding a prefix and calculating a CRC of a given size. He is full of it.3 replies 0 retweets 11 likes
Yep, unlike Vess, Joxean actually knows what he's talking and has actually looked. Vess worked on "F-PROT" a *long* time ago, and guesses about everything else. (I haven't read what Vess wrote, I blocked him).
-
-
He was basically saying that AV engines don’t scan entire files in real time because that would cause too much of a slowdown for the user (which would discredit the feasibility/practicality of
@patrickwardle’s Kaspersky POC).1 reply 0 retweets 0 likes -
Hah, Translation: F-PROT didn't do that on DOS 6.22 in 1984. I wrote an exploit for Kaspersky that *appended* a ZIP file to a DLL https://googleprojectzero.blogspot.com/2015/09/kaspersky-mo-unpackers-mo-problems.html …
0 replies 0 retweets 14 likes
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.