For example, one of the first vulnerabilities I found, that was used for the introduction hook in the 39c3 talk, remains unpatched to this day. Instead of being fixed with code, Werner Koch - the main developer of GnuPG - published a blog post declaring the widely-used feature being “harmful”; while they had weeks in advance, they published this on day one of 39c3, not even giving us time to respond.
what the actual fuck
Somewhat reminded of @jonah@neat.computer’s general recommendation against GPG (for other reasons than what this talk is about)
Alright, so just because a plain text signature is along with content, doesn’t mean that the content is signed.
Then how do you check that the content is signed by the plain-text signature given with it?I thought that the signature means that either a whole copy of the text is encrypted or a hash of the text is encrypted using the private key, which would mean that spoofing the sign would require solving complexity equal to either of:
- Finding another string that gives the same hash
- Finding the private key and signing a new hash with it
What am I missing?
The video, probably.
That was what I prefer to read instead, but fine
Edit: I did watch it, they don’t even tell, because it was in the previous disclosure, so I now need to find and watch that one
Edit2: and from reading https://gpg.fail/ with original vulnerability descriptions I can’t understand how they did the trick with ISO, is the ISO signed as if it were plaintext and allowed truncated lines? If so, this does look pretty bad both on implementation side and on user side, imo
The above question came to me after watching the video.
IDK maybe I just have low IQ.Might be slop then. I didn’t watch the video.
Well they did leave out a lot of important information.
This site seems pretty good though: https://gpg.fail/
It has all the given vulnerabilities in text.Now just need to read and understand all of them and find out which one explains the above comment and the answer to my question is probably another headache.
Is there another thing to use? At least for the flow(s) of “encrypt, transfer, decrypt” or “sign, transfer, verify” for binary files. BSD licensed, preferably.
EDIT: minisign was mentioned on the talk… not in the most positive way but would still trust (after they get their shit together of course); however would never trust “age” with what they did
however would never trust “age” with what they did
What happened with
age(or its devs?) that made you not trust it?
Yeah, GPG is a dumpster fire. It’s security theatre at its finest.
Bruce Schneier, who invented the term security theater, would perhaps disagree with you.
