Got two notifications this morning where I was tagged in an issue, but just a few minutes later when I tried to view the report it was already gone.
Really, props to the Codeberg moderation team for their hard work.
I was tagged, immediately clicked to see the link so I could report abuse, and got a 404. Very fast response. I appreciate that and the message Codeberg sent out that OP posted here. Came to p.d to see if anyone posted it yet.
Here someone abused some feature to automatically create comments mentioning hundreds or more users so that the server send the comment notification via email. The email I got had just the N word followed by "balls" and lots of mentioned accounts. Not sure if others got longer messages?
More likely due to extreme AI scraping. That has been an issue with selfhosted Forgejo instances for months now and it is a complete PITA to deal with.
There is not currently a way to differentiate between web crawlers that are indexing sites for search purposes, vs crawlers that are training AI models. ANY SITE THIS SOFTWARE IS APPLIED TO WILL LIKELY DISAPPEAR FROM ALL SEARCH RESULTS.
I guess an easy fix for that particular issue is to severely rate limit mentions. E.g. if a user mentions more than 100 users in 1 hour then delay them and flag the account. Then you can whitelist it if it's a legit CI bot or whatever.
I had a codeburg account I never did anything with and was worried they got hacked or something. However, since I had never really used it, I just kinda shrugged (separate email, randomly-generated pass). They sent an email mentioning it was this sort of spam.
Got two notifications this morning where I was tagged in an issue, but just a few minutes later when I tried to view the report it was already gone.
Really, props to the Codeberg moderation team for their hard work.
I was tagged, immediately clicked to see the link so I could report abuse, and got a 404. Very fast response. I appreciate that and the message Codeberg sent out that OP posted here. Came to p.d to see if anyone posted it yet.
Seems like codeberg got significant attention for spammers to come and do this.
I'm looking for a repo to host a couple of projects. I hadn't considered codeberg as a home until now.
I guess I'll donate when I sign up.
Exactly. I've already hosted a small script project there but this blog post reminded me I'd forgotten to add them to my December donations list.
I currently have an outage from my own forgejo insrance (which codeberg runs as well afaik) since yesterday. I wonder if that is connected.
Here someone abused some feature to automatically create comments mentioning hundreds or more users so that the server send the comment notification via email. The email I got had just the N word followed by "balls" and lots of mentioned accounts. Not sure if others got longer messages?
same email here.
That's exactly what I got too
So that's what it was...
More likely due to extreme AI scraping. That has been an issue with selfhosted Forgejo instances for months now and it is a complete PITA to deal with.
Time for the LLM crawler trap: https://zadzmo.org/code/nepenthes/
interesting project, thx for sharing! though:
Is there an easier way to give them money than becoming a member?
Take a look here
I guess an easy fix for that particular issue is to severely rate limit mentions. E.g. if a user mentions more than 100 users in 1 hour then delay them and flag the account. Then you can whitelist it if it's a legit CI bot or whatever.
this could be gamed though - mention 99 users, switch accounts, rinse and repeat
I assume there's some barrier to creating accounts that makes it difficult? If not there's pretty much nothing they can do.
I had a codeburg account I never did anything with and was worried they got hacked or something. However, since I had never really used it, I just kinda shrugged (separate email, randomly-generated pass). They sent an email mentioning it was this sort of spam.
Quite childish behavior.
"Signed, a free speech absolutist"
Anti Commercial-AI license
I'm glad I saw this post and codeberg's statement before the spam notifications.