The DMZ

« first day (2706 days earlier) ← previous day next day → last day (2472 days later) »

forest

01:18

A: Is ccrypt 'AES 256' file encryption fundamentally flawed

I'm the author of ccrypt. Sorry for responding so late to this thread, but I was only made aware of it recently. As always, questions about ccrypt can also be directed to me (and are likely to receive a faster response) in the ccrypt forums on SourceForge. To address the original question: let m...

I wish more people realized how much of this is bullshit.

I keep seeing people referencing this ccrypt trash from time to time.

Really, why would anyone use something that intentionally doesn't use a salt?

And a nonce generated with fucking gettimeofday() and getpid().

forest

01:39

> in other words the input of haststring consists of a comma separated ASCII string of the hostname (255 characters max), the seconds in time-of-day, the microseconds in the time-of-day, the process ID and a static / process-wide counter.

I wonder if I should be more proactive and warn the distros that ship this.

7 hours later…

Anders

08:29

@forest Wow.

> But the problem is that they only make dictionary attacks harder by a constant factor, and this is not good enough.

"Should we take the train or should we walk?"
"Well, taking the train is only faster by a *constant factor*, so we might as well walk."

Anders

09:03

What distros ship that, by the way?

Benoit Esnard

09:33

We'll publish critical vulnerabilities in PGP/GPG and S/MIME email encryption on 2018-05-15 07:00 UTC. They might reveal the plaintext of encrypted emails, including encrypted emails sent in the past. #efail 1/4

Tweeted by seecurity on May 14, 2018 at 6:00 AM

Because there much fuss about efail I posted a quick summary. Note that the GnuPG team was not contacted by them in advance; I got the info from a paper to the Kmail developers. https://lists.gnupg.org/pipermail/gnupg-users/2018-May/060315.html

Tweeted by gnupg on May 14, 2018 at 7:59 AM

1 hour later…

Tom K.

10:50

efail.de

It was just published there

1 hour later…

Benoit Esnard

12:00

Thanks!

Benoit Esnard

12:51

wordsofheart.com

(this is quite old, but still genious)

Kepotx

13:36

I know that's not the better way to attack, but are fingerptint brute-forceable ?

Benoit Esnard

I've read a long time ago that there was a 1:64 billion probability that you may have the same fingerprint than someone else

64 billion is quite easy to bruteforce in an offline scenario

I assume that the real number is way lower, since reading fingerprints might not be a "perfect art"

Anders

Somewhat harder to bruteforce if you are making fake ones out of gelatine. :-)

Kepotx

well, if it work the same way as forensic, they use limited amount of "minutiae" so it's not even 64 billion

Anders

On a separate note: Waiting for the site to turn into efail.SE...

Thinking about trying to ask a good question on it, but not sure what the right angle would be.

Kepotx

is there a way to do something against efail ?

I don't really use encrypted email so I'm not personally harmed, but can those who have sensitiva data do anything to prevent someone from reading it?

Anders

13:49

@Kepotx Yeah, I think that is a valid first priority.

Anders

14:00

Question asked: security.stackexchange.com/questions/185777/…

Benoit Esnard

14:41

Do 10k users see removed messages in chat?

Anders

@BenoitEsnard No

Or at least I can't see them. If thats because I'm not supposed to, or if it's because I don't understand how chat works... I don't know.

Benoit Esnard

Chat is such a strange place

M'vy

15:46

Oh I finally made it out of the star wall

forest

16:11

That "Efail" thing seems pretty overhyped.

Before I opened the website I was freaking out thinking that it was something like ROCA but for GnuPG... Turns out it's just a silly bug that exploits HTML parsing email clients that support live embedded links that also automatically decrypt PGP.

@BenoitEsnard You sure that was for a fingerprint and not a key ID?

M'vy

@forest mhmm, that seems to be the consensus

forest

Fingerprints are SHA-1 whereas key IDs are heavily truncated.

You aren't gonna find a SHA-1 collision in a mere 64×10^9 operations.

M'vy

@M'vy and yeah, I so asked for this one lol :P

Tom K.

long time no see ~

forest

@M'vy What is, that the Efail thing is overhyped or that a fingerprint is easy to collide?

Tom K.

16:15

@forest I think the mail by Koch sums it up pretty well

M'vy

ooops I messed up the reply. It was about the Efail yeah

Anders

There seems to be a lot of drama on Twitter over EFAIL. Let the blame game comence.

forest

It... really doesn't seem very bad.

What about the next HTML parser RCE? Will it get all this hype?

I mean at least this isn't SadLock...

Tom K.

well, it's only been... less than 8 hours

Anders

Hype will never be proportional to impact, I'm afraid. (See: KRACK)

Tom K.

16:18

maybe the hype will already die down tomorrow

forest

Well at least KRACK actually has a nasty impact on systems that never update.

Tom K.

@Anders even Spectre & Meltdown. We didn't see any Wannacry-like impact in terms of exploitation so far

Anders

To me, it seems like OpenPGP is a lot like Wordpress, in that there is nothing wrong with the core product itself, but the ecosystem surrounding it is flawed.

forest

Well Wannacry is RCE. But Spectre/Meltdown are... really really really bad. And yes malware is actively using it now to exfiltrate private information.

Tom K.

They are, but where is the widely used malware that really impacts users?

Anders

16:23

@forest I'd say EFAIL is really really really bad as well. We could argue about whos fault it is, but it will certainly be used to extract private info.

forest

But it's so conditional.

Any old RCE in the email client (which are a dime a dozen) is a thousand times worse.

I mean the bug is clever, don't get me wrong.

Tom K.

I'd say 5% of users still use Win XP. So there are always worse exploits

forest

But in terms of actual impact in the real world, I'd say things like good old unicode reversal tricks on filenames or IDN homograph attacks are worse.

Tom K.

how many people actually use S/MIME, OpenPGP?

forest

@TomK. Exactly. And how many people use it with an email client that supports network connections from email? I mean already allowing nearly arbitrary attacker-controlled HTML in an email is a recipe for disaster.

Anders

16:27

One bad thing about EFAIL is that it is so easy to exploit if you have the encrypted emails. You can decrypt them en masse without having to build a complicated payload that targets the client and system your victim is using.

forest

Huh? You can't decrypt it just by having an encrypted email.

The bug is exploited on the victim's mail client.

Anders

@forest Yes, off course.

forest

(Also, I've always decrypted email with gpg in the command line)

Anders

I bet that very few journalists do that, though. Decrypt on the command line, that is.

forest

gpg -d foo.asc | cat -v

@Anders I bet very few journalists even use PGP :P

Anders

16:30

He he

I guess one reason people blame PGP is because people already hate it for having UX from hell.

forest

I don't even get that.

I mean it's only slightly not intuitive.

Tom K.

@forest Everything that is not "click here and everything works like I want to" equals "UX from hell" for Joe Everydayuser

Anders

Yeah, but imagine someone with less intuition than you do, no patience.

forest

If their free antivirus doesn't encrypt their emails for them, it's not intuitive :D

Tom K.

Using a CLI is unacceptable for 99,99% of all people

(probably more 9s)

Anders

16:34

EFAIL or not, I think we need better platforms for end-to-end encrypted conversation. Because let's face it, the people who bother to understand PGP are not always the people who need it the most to avoid persecution.

forest

Indeed. Even with PGP, email is a shitty medium for communication.

If only because it leaks metadata like crazy.

Tom K.

yeah, just declare an end to email

Anders

This escalated quickly.

But yeah, email is a mess.

Tom K.

donate to Signal

forest

We should all go and use Telnet BBSes.

Tom K.

16:39

only ever call people, but rot13 spoken language

Uryyb sberfg, ubj ner lbh?

forest

Rot26 is better.

I use it for all my communications 24/7.

Tom K.

genius

Anders

Personally, I prefer rot4π.

forest

Q: What was/is the purpose of the 'MONLIST' command in NTP?

Having looked through multiple RFC documents about NTP without success (of finding an answer), and a ton of blog posts stating how dangerous, evil and useless this command is, but do not explain why it exist or what the real purpose is, I started to wonder: Why does the MONSLIST command exist in...

ntp

I wonder how often this feature is actually enabled on real production servers.

5 hours later…

NH.

21:22

EFAIL is just another manifestation of the problem of partially-encrypted content. Same reason you shouldn't trust webpages that use HTTPS for all but one little script.

I hate how the news makes it sound like public-key encryption is broken, though.

NH.

22:08

Hey look! I totally predicted that a vulnerability would be found in a poor implementation of PKE!

(I just wrongly guessed that HTTPS would be affected, rather than email encryption)

« first day (2706 days earlier) ← previous day next day → last day (2472 days later) »

all rooms