Thinking about a noteguard filter that broadcasts nostr notes of the IP address of bots that spam my relay. Then other public relays could read that note and add it to their block list automatically (if they trust me).

Or could be at least be interesting as a diagnostic tool, could follow that account to monitor attacks.

Would be happy if the person spamming me right now gets banned from all public relays. Spammers don’t get censorship resistance on public relays, sorry. nostr:note14dq8aa4whf3wh6ud3qda60qjmmpnddukfu8mnxm0pfu8e3egkp2sqq36ez

Reply to this note

Please Login to reply.

Discussion

Like it! 👍

No solutions only trade-offs.

I've seen this 'rainbow spam' stuff on a few notes, seems to be on trending/popular notes, no? I think it's more than just you. Your posts are popular tho lol

i have 6 spam notes in my notifications per minute

You can only really do this if it was a targeted attack due to my ip based relay rate limits, but it’s possible its slipping through via other relays. I will test this

No sure I get it. But I was just saying it may not be specifically targeting you/personal. Doesn't make it less annoying or anything lol. But I suppose it's just not someone who hates you, so that's... something :) ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯

Based on analysis of all the incoming spam on my relay it does specifically targeting me, the replyguy spam is not effective on my relay due to rate limits, but targeted attacks are still possible.

Ok, I'm just guessing here cos I've also been seeing it on on vitor's post and fiatjaf's. The rainbow impersonation stuff?

What do you mean by rainbow?

That's the spam I'm seeing. Let me get a screenshot as an example

Oh interesting, i see its targeting vitor and fiatjaf as well. I will turn off all other relays and see if its some specific relay not stopping it

I'm seeing this sort of thing a lot. Impersonator rainbow spam.

Like impersonator accounts that repost old notes maybe (?) with lots of rainbows in the text.

It seems to be on popular notes / accounts.

I assumed it was just me because they aren’t adding lots of p tags which would have been smart, but maybe that would be easier to block

I guess 6/minute ratelimit is still not enough to stop this. Even on a slow drip its still annoying. Next thing will be tracking numbers of notes per ip i guess and blocking the ones that are doing this.

It's kinda tradeoff, bcs i as not that popular npub could be counted as a spam if i post lot of notes i guess.

Only if you post more than 6 times a minute persistently

Yeah, hope thats okey. Sometimes when im high i post like 4 notes in minute, or to be complete- 4 comments, idk if comment counts

That sounds more reasonable from a privacy point of view, only broadcasting IP addresses with some reasonable suspicion.

Relays can use that info directly, but clients could also tell their relay to filter from a given IP address (though that's a bit dangerous as an attacker could use it to confirm that an IP address is used by some npub).

垃圾邮件发布者在公共中继中应该受到审查抵制!

This could also help normalize relays having npubs of their own :)

If you built something like this I'd love to build a simple dashboard for it in noStrudel

yeah I definitely want relay pubkeys, not sure if there is relay metadata for that yet

There is a pubkey field in nip-11 but its for the owner. Although depending on how you look at it you could consider the npub that is moderating the relay the "owner"

Also nip-66 has support for relays that are addressed by their npubs

i thought about this in the past. the danger is centralized banning.

The relay is accountable in this case, if i broadcast a note that censors a legitimate user there could be a public investigation and people would no longer trust my ip ban lists. Still tied to trust and reputation.

if a legit user and a spammer uses the same vpn, the legit user will be banned in more places than 1 relay

VPN users should probably not be using public relays, vpns will likely banned eventually due to spammers

That’s a bummer. I always connected thru Mullvad.

just use a paid relay like nostr.wine or the damus purple relay when its out.

Purple relay would be nice. Don’t wanna pay just for the relay service at this stage.

Won't spammers generally use cloud resources? Or are we including those in the VPN group?

I suspect it will get to that stage, and then I will have to start temp banning ip ranges.

Total transparency of relay filters helps address this because everyone involved has access to the same information.

Relay operators are not trying to suppress information without others knowing.

What if you did a combination of a block List and allow List? so the users can subscribe to your list by default in your platform but then they could submit a set of users or IP to their personal allow list on their client and it permits these to come through or you could reference their followers if they’re following somebody you’ve blocked, you could allow that to come through. If there’s inappropriate person responsible because they specifically added it to their list.