And yeah the way it is advertised (yes, advertised is the correct word) makes it sound more like it's a problem with the standard or with the crypto itself (bad RNG in gcrypt, etc), which would be pretty world-ending.
If all my PGP messages could suddenly be decrypted by anyone today, I don't know what I'd do. At least QC is something we all see coming, but a bug?
@TomK. It always afraid me when I see that newspaper are sometimes far from reallity on topic I know, because I wonder how far from reallity they are for topic I don't know...
Unrelated: wasn't there some research showing how the stock market could be used to reveal confidential information like finding out what company got a contract to use some kind of proprietary rocket fuel?
It makes me wonder if the NSA's Utah datacenter's general specs could be estimated, since it would be prohibitively expensive if they built all the drives in-house. Some company like Seagate or WD must have nailed a jackpot contract.
In which case it would still show. I mean you really can't assemble that many drives in secret without it showing, even if you split it up among several companies.
(And yeah I know, the sheer number of shell companies the NSA uses... The majority of the NSA is actually split up into 11 or 12 distinct shell companies that do all their work for them IIRC)
I had a pentester here yesterday who whill be continuing his test today. We couldn't really start yesterday, because the web application that had to be tested had response times over 2000 ms long
Our setup looked like this:
test system <-> some firewall (allow any/any) <-> web app
the web app is also reachable from the internet, but there is a web application firewall in front of it
so a normal user has to take this route:
user <-> WAF <-> web app
so if you try to reach the web app from the internet, the response time was fine (<50 ms). but with our setup the response time was >2000 ms
the theory of the pentester was, that the web app made a reverse lookup, and because the test system had no host name, the reverse lookup couldn't find anything, timed out and then proceeded as normal
any normal user would have have a hostname because of the routing through the WAF
@Anders: "Stop encrypting email in your email client (e.g. by disabling the PGP plugin), and instead copy paste the encrypted data into a separate program to do the encryption." - say what now?
Did you mean: "Stop decrypting email in your email client (e.g. by disabling the PGP plugin), and instead copy paste the encrypted data into a separate program to do the decryption."?
I wonder if we made up a name, a website and a logo and faked a half good looking paper how far we could take a new vuln/attack that was called "Flooding with electricity"
3
don't present it to the academics, directly give it to news organizations
also, we should massively invest after releasing our cryptocurrency to the newspapers. lot of folks will think it's worth it, we sell all our coins and we are rich
I think this answer on Crypto.SE contains some very relevant criticisms of OpenPGP in light of EFAIL. While not disputing that obviously mail clients are to be blamed, it also highlight some pretty important design flaws.
There are a few parts to the EFAIL attacks. Some are the fault of the mailer authors; some are the fault of the OpenPGP and S/MIME designers for failing to heed modern cryptography engineering principles.
HTML interpretation.
If you receive a message with <img url="http://efail.de/0a7492fc62d...
Hmm... If you are just guessing on your own, without no interaction, I wouldn't say it counts as social engineering. Tricking someone into disclosing their birth date so you can test if it works as a PIN, on the other hand, would qualify. Other might disagree.
@AJHenderson Well, If you make a good comment over several lines or one that makes only sense in a conversation, I would have to star every single line. Which is a bit weird IMO
Just staring "sure, it just type coerces in to insanity" would look kinda weird on the wall