[ursuscamp]

So, I was part of the Nostr community for quite a while and was the author of a popular Nostr extension for Safari, before eventually giving up on Nostr for various reasons.

I haven't read that entire paper. Mainly, I skipped to the section you mention here:

> The event protocol that drives the system doesn't authenticate public keys, so asymmetric signatures are performative: attackers that can intercept messages (Nostr servers, the presumed adversary of an E2EE messaging system) can just swap out keys and re-sign.

I think you and the authors perhaps misunderstand the Nostr protocol. Nostr is, effectively, an identity system tied to a public key. The cryptography is sound. Your identity is your public key. When you request a user's profile, or their events, you request it specifically by their public key. That is unforgeable (assuming no bugs in the implementation, like what the authors found in Damus).

This does present UX issues that can manifest as security issues, such as "how can you verify that a user with a certain public is who they say they are instead of an impostor". That is a separate issue from whether the cryptography itself is sound.

[tptacek]

If you read the entire paper you'll see that the paper presents a formalized set of security goals that acknowledge Nostr uses public keys as identities. They haven't misunderstood the system. Meanwhile: the cryptography is obviously unsound: it relies on unauthenticated CBC, and signatures that aren't verified, and provides attackers with the ability to coerce users into following links.

[jonstaab]

The cryptography was thrown together in the very early days as a proof of concept, that reached some level of adoption because of how nostr suddenly grew at the end of 2022. The community has since largely switched to a new standard (NIP 44) which has been independently audited, although there are some popular clients that haven't yet transitioned.

[tptacek]

From a brief scan, NIPS 44 seems reasonable; it's just AEAD ChaCha20, which is boring, which is good.

[eduction]

Interesting way to spell “i was extremely wrong”

[skinkestek]

I think it is an HN standard:

I am less on HN these days, but as far as I have seen:

Telegram is still judged by its very early releases, still called "unencrypted" while it is about as encrypted as your bank transactions (they definitely aren't e2ee either).

Signal can do what they want including dabbling in crypto currency without being open about it. Signal can also have extremely "interesting" bugs (didn't it at some point send messages to random people?) and glaring security issues (relatively trivial remote code in the desktop client IIRC a few years ago).

Last I checked WhatsApp was supposedly also good since they now use good encryption despite now being owned by Facebook, sending my social graph to them and sending peoples entire backups (including chats with me) unencrypted to Google for "free" (IIRC) backup.

That said these days I am definitely looking for Telegram alternatives.

[tptacek]

Your bank doesn't operate in Telegram's threat model! You are never concerned that your bank's servers are attacking your transactions: if you can't trust your bank, you're fucked anyways. That's precisely what's not supposed to be the case about a messaging service!

[tptacek]

Is the protocol the paper was written about no longer deployed anywhere, or is this just a dunk?

[jonstaab]

It was (and still is) deployed in a number of places, so it's a valid criticism in that context. That said, it's mostly dunk.

[vnuge]

nostr cryptographic developer here (author of libnoscrypt C library)

Nip04 has been deprecated, and to be clear, in practice the nip04 payload is in a signed nip01 event wrapper.

nip44 replaced nip04, which has been reviewed/audited. Does use authenticated encryption in the message payload with forward secrecy, again in practice wrapped in a nip01 event, singed by the author, usually by the same cryptographic software used to encrypt the message.

nip44 is becoming more widely used for direct messages and other "private" metadata stored on relays. It's chacha20 + hkdf.

[tptacek]

Hey, I should say this, and wish I had earlier:

I don't really so much care whether Nostr is good or bad. I'm a connoisseur of cryptographic vulnerabilities, and the ones in that paper are fun. We host a podcast (me, Deirdre Connolly, and David Adrian) that is mostly about good crypto vulns. If there's someone affiliated with Nostr that would want to chat for an hour or so about how applicable the vulns in this paper are or aren't, and how they're addressed in NIP44 --- we'd love to talk. My email address is in my profile. Whoever showed up, they'd be in good company!

https://securitycryptographywhatever.com/

[duskwuff]

How robust is nostr against downgrade attacks - can an attacker induce users to use nip04, or are there safeguards against this?

(I have no significant knowledge of the protocol; if this is a meaningless question just say so.)

[vnuge]

Hard to say how relevant that is. DMs are simply a collection of events sitting on a relay. It's not really a mutual tunnel, most clients implement nip44 via nip19 (giftwrap DMs) so your ni04 message wouldn't likely make it to them. It's not considered backward compatible such that you could send a user a DM, then cause their client to downgrade to the DM scheme that uses nip04.

It's also worth noting, the user _must_ be made aware of the encryption method that was used, their "signer" application, which is also responsible for encryption and decryption, would require their permission to do an operation in either direction. Users may often choose to grant a trusted client application the permission to decrypt all nip04 or nip44 messages alike, automatically, or generally manually with a popup. That's up the signer application how granular the permissions get.

To be clear this is a client implementation detail, im not a client developer, so I can't say in practice how many have handled the UX on this, but know that the signer, and the user had the final say on which algorithm was granted permission.

Clients and signers alike could choose to block obsolete encryption methods if they choose.

[some_furry]

> It's also worth noting, the user _must_ be made aware of the encryption method that was used, their "signer" application, which is also responsible for encryption and decryption, would require their permission to do an operation in either direction.

Let me ask a more pointed question about downgrade attack resistance then:

Is the algorithm being used determined by the encrypted message contents? Or is it determined by the key controlled by the signer app?

[vnuge]

Regardless of the signer interface the procedure call remains the same. The client application determines what method it wants to use, then the plaintext is passed to the signer (web extension, nip46 remote signing, android etc) with the nip44.encrypt or nip04.encrypt procedure calls.

The user is then requested to confirm the encryption operation. So a "downgrade" could happen in two ways. The client selects nip04 without the user's instructions, and the signer does not properly guard or notify the user that the message to be encrypted is using nip04. Still not really an attack I don't think, since no "sessions" exist in DMs there shouldn't be any way a remote user gets to cause a client to change algorithms.

To answer directly, the client app chooses, makes a remote procedure call with the desired algorithm, user confirms, message is encrypted, returned, signed (another rpc round-trip), then written to relays.

The signer application is ALWAYS authoritative, if it chooses to.

[sealeck]

> how can you verify that a user with a certain public is who they say they are instead of an impostor

This sounds awfully much like a cryptography problem to me!