moderation queue can be easily flooded by a malicious user

[antibot4navalny]

A malicious user (eg state-sponsored Matryoshka bot tinsleygold.bsky.social for the @ab4n-labeler.bsky.social labeler) can easily flood labeler moderation queue with illegitimate reports, and Mute User doesn't stop those reports from appearing in Reports tab.

[SteveTownsend]

We need Ozone tooling to block them for good and flush the malicious reports out of the DB.
Have you got heuristics to detect Matryoshka realtime, or are you relying on user reports? I have a firehose client that scans for disinfo tropes and displays them in Grafana. I am working to have it do realtime detection and alerting. Ideally it will be adaptive without code updates or a restart to pick up new filtering info, in the end state.

[antibot4navalny]

Have you got heuristics to detect Matryoshka realtime, or are you relying on user reports?

As long as Ozone is vulnerable to flooding by any account, it is not that important whether or not we are able to detect a malicious actor of a specific type.

[SteveTownsend]

*Blacklist* as a permanent mute that blocks forever and flushes out reports from a malicious user in Postgres would be helpful

…

On Sat, Jan 4, 2025 at 10:12 AM antibot4navalny ***@***.***> wrote: Have you got heuristics to detect Matryoshka realtime, or are you relying on user reports? As long as Ozone is vulnerable to flooding by any account, it is not that important whether or not we are able to detect a malicious actor of a specific type. — Reply to this email directly, view it on GitHub <#274 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADRDNC3A76XEYZ7S6TME2ED2I7245AVCNFSM6AAAAABUTBZAOSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNZRGMZTAMJWGE> . You are receiving this because you commented.Message ID: ***@***.***>