Encryption with Gnu Privacy Guard (GPG) | Hacker News Reader

Encryption with Gnu Privacy Guard (GPG)

digital-era.net

47 pointstriberian30 comments

Open articleView on HN

On all the encryption tutorials I found they always assume 2 and only 2 people trying to talk privately.

I wonder how would one encrypt a conversation between say 15 people.

Multiple -r's:

gpg -e -r recipient1@example.com -r recipient2@example.com

That products some ciphertext which can be decrypted by either recipient.

Thank you. Does it weaken the encryption strength ?

No. The symmetric key that is used to encrypt the message is itself encrypted once for every recipient with their respective public keys, so the only difference is filesize (for most intents and purposes negligable). Of course, more recipients equals more targets for an attacker.

> Of course, more recipients equals more targets for an attacker.

This is especially significant if the recipients have different sized public keys. If you send a message to a 2048-bit key and a 1024-bit key, the message only has 1024 bits of security. So you really do get a weakest link effect.

I learned this the hard way when using gpg+mutt back in 2001. All my Sent mail was being encrypted only with the recipient's key, so I couldn't read it myself. There's an option to also encrypt outgoing email with your own GPG key.

In case anyone is wondering how to make this the default behavior:

"encrypt-to YOUR_KEY_ID"

in your gpg.conf

This article also suggests that you should use a 2048 Bit key, and you really shouldn't. Better to go with 4096 Bit. And I guess every CPU is fast enough.

This does not go into the Web of Trust and having your key signed, maybe the most complicated part of PGP, and perhaps also its most important feature.

Agreed - it also fails teach how one should create a revocation certificate.

No it doesn't. See the section titled "Generating revocation key"

Web of Trust isn't too useful if you don't know how to sign a message or check a signature.

Specifically "check a signature". Even everyone that knows how to sign fail to really check anything. Honestly, if it aint boiled down to a green or red icon then people are going to misuse, assume security when its not, or just click "yes".

If you get a new public key for someone, and it's signed by someone else you don't know (most of the time this is the case) are you going to bother to utilise that signature to increase the trust? No. Are you going to assume increased trust somehow regardless? Yes. FAIL.

Modify it further, if it is signed by someone you do know what are the chances your UI is going to give you any sort of indication that this is more trusted that other message workflows? And if it doesn't give you any indication (most leave it up to you to check) who's gonna bother to look just passed "your friend <yourfiendsname@company>" ASCII and actually check the key? no one. FAIL.

PGP WoT fails miserably.

I use PGP extensively. However, I haven't bothered with the web of trust, for the same reason I don't publish my address book for everyone to read, privacy.

I have already run into this at work. I have customers whose keys are publicly available, and others who should not be, and so I cannot effectively use keysigning internally to manage relationships. This is because I don't want to upload every key I know about to a public keyserver.

Maybe I am mistaken or maybe there's a way to do that without having multiple keyrings or something, but that's kind of the problem. It's really hard to tell what I would or would not be sharing because there's no "interface" to speak of and I am hardly an expert.

btw, how is a person signing a key in the usual WoT model supposed to know whether or not the email address specified in the key uid is actually controlled by the person standing in front of them at the key signing party with passport and key fingerprint in hand? Even if you carefully follow the recommended key signing procedure the binding which actually matters for exchanging encrypted email has very poor validation (or none at all).

Best practice is to send an encrypted message to that address in order to confirm the owner controls it.

If you only meant verifying the email as apart of a keysigning party, agreed. Meaning you are looking at their passport and sending an email at the same time. This is better than nothing but wont protect against many directed attacks. For instance, a journalists has to be more careful. Also, in the end if you are sitting in front of each other might as well just verify key fingerprints. Better privacy and security than a simple email verification.

If the author meant actually sending someone a email to see that it is indeed owned by them remotely: NO WAY. Does not work:

me: "hey, just sending you an email to confirm its really you before sending the secret plans?"

them: "yeah, its me. send the plans."

That is exactly what WoT tries to solve but can't due to misuse.

I think you misunderstand. After meeting someone and exchanging public key fingerprints, you verify their identity. This requires checking their government ID and making sure they really are who they say they are. Then, you verify the uid of their key by sending them a message encrypted with the figerprint you received in person. Only the holder of both that key and the email address can respond to your message. Then you can sign their public key and return it to them.

What kind of attacks is this practice vulnerable to?

clearer, thanks.

> What kind of attacks is this practice vulnerable to?

So long as the trust only translates to the very limited use cases then there is no vulnerability. Those limits basically mean that no trust should be assumed by anyone that did not participate in the WoT in question. This particular style of WoT means anyone wanting to retain anonymity needs to skip it, or am i missing something?

Mainly I think all WoT models i've seen thus far are too susceptible to sybil attacks (impersonation) and as a result instill bad habits.

you only need to look at people use of pgp today to see how much WoT's would be a failure if used. pgp.mit.edu still aint got ssl. journalists are linking to nonssl links for their key via twitter t.to urls

> What kind of attacks is this practice vulnerable to?

I want to pretend to control target@example.com and the legitimate owner of this address is an OpenPGP user who has published a keyring on the public keyservers.

1) I create a keyring and add a single uid with my real name and target@example.com

2) I download the public keyring for the legitimate user target@example.com and extract the encryption subkey.

3) Even though I don't know the private key I can add this public key as the encryption subkey to the keyring created in step #1.

4) I publish this keyring on the public keyservers so that you will find it by querying the fingerprint I give you when we meet.

5) You send email to the real user target@example.com which they are able to decrypt and respond to. Of course there could be some confusion since the real user is not expecting an email which presumably talks about verifying keys.

6) Since the mail was decrypted and responded to, you sign the key and return it to me.

7) I revoke the certification on the encryption subkey I borrowed from the real user and add a new encryption key which I create.

8) People who trust your signature encrypt mail to target@example.com with the false key I've published.

I don't know if it's very useful even if you do know how to sign keys and verify signatures.

There seems to be also a windows implementation: http://www.gpg4win.org/ (Wondering why the site doesn't use ssl, though)

On a related note, I have a hard time understanding why a web site talking about digital security also doesn't have a certificate.

Lots of GPG themed sites do not. Notice that key servers provide web access over plain HTTP as well. My best guess is that they do not want to buy into the CA infrastructure and provide security using GPG itself. Also check out MonkeySphere.

> My best guess is that they do not want to buy into the CA infrastructure and provide security using GPG itself.

This may be true (I really have no idea) but isn't it like travelling on a highway at night with your normal lights turned off because you have a better system based on infrared? After reading about the story of that guy who hacked into a computer by MITMing the notepad++ site, I became even more convinced that all pages must have certificates. Nowadays it's also possible to get a basic ssl cert for free, I can't figure out what the catch is really.

Don't get me wrong, I agree with you. Even my personal 100% static content site is served over HTTPS only. I am just commenting on the pattern that any GPG-themed sites I've seen follow.

Did anyone else see reop recently? It's a sort of simple reimagining of easy-to-use public key crypto. http://www.tedunangst.com/flak/post/reop

"To encrypt a File:"

    gpg -o encrypted_file.gpg –encrypt -r original.file

# they didn't put recipient after -r ?

In this case it will just prompt you for for the recipient. This command is too verbose though. You can type:

gpg -e file

The output will be file.gpg. It will ask you for a recipient unless you have "default-recipient" or "default-recipient-selt" set in gpg.conf.

Best blog on using GPG that I have seen. Thanks!

Reply on news.ycombinator.com