people spamming cp into people's first #introductions post is a huge problem. looks like we may need to start adding blocklists on media from certain domains.
looks like the ai-cp spammers are using this domain:
libfans.com
just saw apparently real cp on this domain:
https://matrix.org/_matrix/media/v3/download/livingutopia.org
I also contacted abuse at matrix.org for the latter since it looks like non-ai. we need better reporting protocols than email.
Login to reply
Replies (30)
Hi jb55, and welcome to #nostr! 🤓
✅ Great, you've already set up your #lightning address! I just ⚡️zapped you some #welcomesats 😊
Made with 💚 by @minestr.app
Can't tell if it's Federal Agents, butthurt people upset that Nostr exists, or someone stress testing the network.
Regardless, I'm glad you guys are beginning to acknowledge some of the problems.
Fire up the wood chippers.
that's just disgusting.
Had my first dose of that on i think #asknostr last night... very disturbing indeed.... definitely not the first impression you want to leave with new users.
What’s “cp”? Asking for a friend, I mean I totally know what it means… 

made a draft to extend nip56, would appreciate feedback or imrovements.
The logic is that clients can start off users with a whitelist of auto preview domains, but that kind of data can also exist on relays for other users/clients to make decisions from.
```
{
"kind": 10099,
"content": "",
"tags": [
["d", "domain_lists"], // identifier
["white", "nostr.build"],
["white", "void.cat"],
["black", "malicious-site.net"],
["black", "scam-domain.com"]
["unknown", "ask"] // Options: "load" | "block" | "ask"
]
}
```

GitHub
nips/56.md at extend56 · limina1/nips
Nostr Implementation Possibilities. Contribute to limina1/nips development by creating an account on GitHub.
Interesting…

awesome
I approve of this. It would make it quite easy to build automatic lists based on web of trust as well.
definitely the ideal solution. I'm much more ok with friends reporting media domains instead of pubkeys themselves which isn't effective anyways because they are often randomly generated.
That is the matrix media proxy.
wow really
yeah each homeserver has their own media proxy under that path to not reveal user IPs
the domain at the end is the server actually hosting the content
do you think everyone is dumber than you or something? This is pretty obvious.
no?
also, it may not be obvious to many people reading the note, either
if you don’t care, move on
Well, If you want to fix this for long term,
Disable auto-loading media on non-following user.
That's it.
people spamming cp into people's first #introductions post is a huge problem. looks like we may need to start adding blocklists on media from certain domains.
looks like the ai-cp spammers are using this domain:
libfans.com
just saw apparently real cp on this domain:
https://matrix.org/_matrix/media/v3/download/livingutopia.org
I also contacted abuse at matrix.org for the latter since it looks like non-ai. we need better reporting protocols than email.
View quoted note →
This is what I've been saying! I think Primal never takes down their CP images either, so we should add them to the blacklist until they improve.
Until we centralize media hosting, child pornography will remain an existential threat to Nostr.
Did you do something about Primal too?
Thanks for posting. Still wanna shoot that mf after seeing it. I was just looking in introductions to welcome new people.
"they" really want to shut down free speech by making it look horrific.
Is there any feasibility in client AI that auto filters stuff like that?
yep, pushed it this morning
I mean, it's hard coded, but I would add your NIP once it's drafted
Thanks for the heads up on those domains. Right after reading this post I scrolled down and saw a bunch of posts in a row from anon accounts with links using those domains. Thankfully, the previews didn't auto populate.
I'm not sure if I should report, because it could be normal content just using that domain. But I don't wanna click on the link to find out.. 👀
GrapeRank uses follows, mutes, and reports “out of the box” to generate a WoT network of “trustable” users for any pubkey. We got this. @david
GrapeRank’s design allows it to ingest data from any source and create contextual trust scores.
So if liminal’s kind 10099 notes catch on, we could use them to generate curated lists of whitelists and blacklists. In other words, a composite of 10099 notes authored by trusted pubkeys. It would be a simple enough matter to add this feature to grapevine-brainstorm.vercel.app which is currently in production.
Alternatively, these lists can be handled in the background by your personalized grapevine WoT relay, which is what I’m currently working on.
🍇🧠⚡️
ah wtf I just mentione nostr to my work people . im 100% not metioning nostr to people if that can happen that easily ,very unnerving that im a click away from that so yea it's not huge problem its THE problem rn