'Potential Scam' Detector (moderator/labeler)
@cryptolabeler.sats4.life
Fighting crypto spammers, scammers, and malicious activity in BlueSky.
Our AI-powered labeler detects suspicious accounts, verifies posts, and combats scams.
Report suspicious activity to help us to bsky make a better place.
Our AI-powered labeler detects suspicious accounts, verifies posts, and combats scams.
Report suspicious activity to help us to bsky make a better place.
Pinned
a man with a beard says " you got it " in front of his face
ALT: a man with a beard says " you got it " in front of his face
media.tenor.com
Tip for reporting: Add evidence to the report.
Sometimes it is hard to guess the motives behind the reporting, and we can't make a decision on how to label it.
Sometimes it is hard to guess the motives behind the reporting, and we can't make a decision on how to label it.
The spam detector *does not* use AI. It evaluates how fast an account sends messages, or how quickly a message is spread among accounts (mostly bots broadcasting spam).
The scam detector *does* use AI. It evaluates the context of the message and then decides whether to label it or not.
The scam detector *does* use AI. It evaluates the context of the message and then decides whether to label it or not.
May 8, 2025 at 9:57 AM
The spam detector *does not* use AI. It evaluates how fast an account sends messages, or how quickly a message is spread among accounts (mostly bots broadcasting spam).
The scam detector *does* use AI. It evaluates the context of the message and then decides whether to label it or not.
The scam detector *does* use AI. It evaluates the context of the message and then decides whether to label it or not.
It is false that the spam detector bot is anti-religion. It does not judge the content of the message, but the behavior.
It verifies messages under 2 rules, for quantity and span of time:
- The same message sent to many different accounts
- Multiple accounts posting the same message
It verifies messages under 2 rules, for quantity and span of time:
- The same message sent to many different accounts
- Multiple accounts posting the same message
March 16, 2025 at 10:14 AM
It is false that the spam detector bot is anti-religion. It does not judge the content of the message, but the behavior.
It verifies messages under 2 rules, for quantity and span of time:
- The same message sent to many different accounts
- Multiple accounts posting the same message
It verifies messages under 2 rules, for quantity and span of time:
- The same message sent to many different accounts
- Multiple accounts posting the same message
We we're included in some strange lists.
@democracyforever2.bsky.social
@rootc0re.com
@squirrelturdlists.bsky.social
@plresists.bsky.social
@keksekaempfer.bsky.social
Can you review your lists, please?
@democracyforever2.bsky.social
@rootc0re.com
@squirrelturdlists.bsky.social
@plresists.bsky.social
@keksekaempfer.bsky.social
Can you review your lists, please?
February 28, 2025 at 7:51 AM
We we're included in some strange lists.
@democracyforever2.bsky.social
@rootc0re.com
@squirrelturdlists.bsky.social
@plresists.bsky.social
@keksekaempfer.bsky.social
Can you review your lists, please?
@democracyforever2.bsky.social
@rootc0re.com
@squirrelturdlists.bsky.social
@plresists.bsky.social
@keksekaempfer.bsky.social
Can you review your lists, please?
The maintenance on our infrastructure is complete.
The labeler is now back online and operational.
The labeler is now back online and operational.
February 25, 2025 at 4:58 PM
The maintenance on our infrastructure is complete.
The labeler is now back online and operational.
The labeler is now back online and operational.
We are performing maintenance on our infrastructure. The labeler will be offline for ~30 minutes.
February 25, 2025 at 4:34 PM
We are performing maintenance on our infrastructure. The labeler will be offline for ~30 minutes.
Added new label 'Rude/Troll/Aggressive'
The “Rude/Troll/Aggressive” label applies to posts, comments, or accounts that engage in hostile, disrespectful, or inflammatory language aimed at provoking or insulting others.
The “Rude/Troll/Aggressive” label applies to posts, comments, or accounts that engage in hostile, disrespectful, or inflammatory language aimed at provoking or insulting others.
February 19, 2025 at 6:13 PM
Added new label 'Rude/Troll/Aggressive'
The “Rude/Troll/Aggressive” label applies to posts, comments, or accounts that engage in hostile, disrespectful, or inflammatory language aimed at provoking or insulting others.
The “Rude/Troll/Aggressive” label applies to posts, comments, or accounts that engage in hostile, disrespectful, or inflammatory language aimed at provoking or insulting others.
Tip for reporting: Add evidence to the report.
Sometimes it is hard to guess the motives behind the reporting, and we can't make a decision on how to label it.
Sometimes it is hard to guess the motives behind the reporting, and we can't make a decision on how to label it.
a man with a beard says " you got it " in front of his face
ALT: a man with a beard says " you got it " in front of his face
media.tenor.com
February 15, 2025 at 5:10 PM
Tip for reporting: Add evidence to the report.
Sometimes it is hard to guess the motives behind the reporting, and we can't make a decision on how to label it.
Sometimes it is hard to guess the motives behind the reporting, and we can't make a decision on how to label it.
Since there are some recurrent volunteers reporting, I thought it would be great to have a place to discuss moderation/label stuff with the community.
Then our Discord server has born: discord.gg/NrzBje4s
Then our Discord server has born: discord.gg/NrzBje4s
Join the Crypto Labeler's server Discord Server!
Check out the Crypto Labeler's server community on Discord – hang out with 2 other members and enjoy free voice and text chat.
discord.gg
February 12, 2025 at 9:20 AM
Since there are some recurrent volunteers reporting, I thought it would be great to have a place to discuss moderation/label stuff with the community.
Then our Discord server has born: discord.gg/NrzBje4s
Then our Discord server has born: discord.gg/NrzBje4s
We updated our bot and the automated system is partially working.
SPAM detector is ready and deployed.
SPAM detector is ready and deployed.
February 4, 2025 at 7:33 AM
We updated our bot and the automated system is partially working.
SPAM detector is ready and deployed.
SPAM detector is ready and deployed.
Unfortunately because of a recent update on bsky network, the auto labeler is out-of-work.
I'm following the updates on their Discord and Github and as soon they publish a final solution, we will update our system.
Until there, keep reporting :)
The manual labeling is working fine.
I'm following the updates on their Discord and Github and as soon they publish a final solution, we will update our system.
Until there, keep reporting :)
The manual labeling is working fine.
February 2, 2025 at 2:07 PM
Unfortunately because of a recent update on bsky network, the auto labeler is out-of-work.
I'm following the updates on their Discord and Github and as soon they publish a final solution, we will update our system.
Until there, keep reporting :)
The manual labeling is working fine.
I'm following the updates on their Discord and Github and as soon they publish a final solution, we will update our system.
Until there, keep reporting :)
The manual labeling is working fine.
Due to several reports about misleading subjects and accounts, we determined that the label 'misinformation' was not enough.
We added the label 'misleading' to better fit those types of reported subjects.
We added the label 'misleading' to better fit those types of reported subjects.
January 22, 2025 at 8:31 AM
Due to several reports about misleading subjects and accounts, we determined that the label 'misinformation' was not enough.
We added the label 'misleading' to better fit those types of reported subjects.
We added the label 'misleading' to better fit those types of reported subjects.
We encourage our users to report or appeal any posts and accounts using the report function (the three dots menu).
We are continuously working to enhance our detection systems, but sometimes a post or account may be incorrectly labeled.
We are continuously working to enhance our detection systems, but sometimes a post or account may be incorrectly labeled.
January 14, 2025 at 2:18 PM
We encourage our users to report or appeal any posts and accounts using the report function (the three dots menu).
We are continuously working to enhance our detection systems, but sometimes a post or account may be incorrectly labeled.
We are continuously working to enhance our detection systems, but sometimes a post or account may be incorrectly labeled.
In case you are interested to know more about the labeler infrastructure, here is a diagram showing how our system is working.
Right now it is super optimized, and we are improving it continuously.
10 days since we've implemented the AI module.
#ai #spam #scam #detector #crypto #bitcoin #atproto
Right now it is super optimized, and we are improving it continuously.
10 days since we've implemented the AI module.
#ai #spam #scam #detector #crypto #bitcoin #atproto
January 4, 2025 at 11:11 AM
We are currently testing the system's capacity over the weekend to see if we can integrate more detection services under the same data stream.
Let's test the limits of the detection.
Let's test the limits of the detection.
January 4, 2025 at 7:58 AM
We are currently testing the system's capacity over the weekend to see if we can integrate more detection services under the same data stream.
Let's test the limits of the detection.
Let's test the limits of the detection.
Our labeler has been active since December 20th, and on December 24th we implemented the first version of our automated AI detector.
Since then, our labeler has registered #65580 actions. It is A LOT! 😱
Making #bsky and #atproto a safe place for everyone! That's our goal for 2025. 🎉
Since then, our labeler has registered #65580 actions. It is A LOT! 😱
Making #bsky and #atproto a safe place for everyone! That's our goal for 2025. 🎉
January 2, 2025 at 1:16 AM
We have implemented a new feature on our spam detector to flag repeated messages from the same account. It is working for the entire feed.
We will fine-tune this feature over the next few days to identify any areas for improvement.
#spam #stopspam #nospam #scam #labeler #crypto #bitcoin
We will fine-tune this feature over the next few days to identify any areas for improvement.
#spam #stopspam #nospam #scam #labeler #crypto #bitcoin
a dog is walking up a set of stairs with the words `` no spamming '' written on the bottom .
ALT: a dog is walking up a set of stairs with the words `` no spamming '' written on the bottom .
media.tenor.com
January 1, 2025 at 3:29 PM
We wish you a Happy New Year with a clean and trustworthy timeline, and a lot of joy on #bsky in 2025. 🎉🎈🎊
a red background with the number 2025 on it
ALT: a red background with the number 2025 on it
media.tenor.com
December 31, 2024 at 5:13 PM
We wish you a Happy New Year with a clean and trustworthy timeline, and a lot of joy on #bsky in 2025. 🎉🎈🎊
After several tests, we decided to work with a 3 seconds window to detect spam.
The rule applied is: if the same message appears from different accounts within that time frame, it is labeled as spam.
The rule applied is: if the same message appears from different accounts within that time frame, it is labeled as spam.
December 28, 2024 at 9:38 AM
After several tests, we decided to work with a 3 seconds window to detect spam.
The rule applied is: if the same message appears from different accounts within that time frame, it is labeled as spam.
The rule applied is: if the same message appears from different accounts within that time frame, it is labeled as spam.
We made a lot of progress today in fine-tuning the algorithm to detect general spam and crypto scams.
December 27, 2024 at 6:36 PM
We made a lot of progress today in fine-tuning the algorithm to detect general spam and crypto scams.
We are testing an automated spam detector to label accounts that sends the same message from many diffent accounts.
December 27, 2024 at 10:25 AM
We are testing an automated spam detector to label accounts that sends the same message from many diffent accounts.
Hello @labeler-lists.bsky.social! Can you list us in your repo? 🙏
December 25, 2024 at 8:40 PM
Hello @labeler-lists.bsky.social! Can you list us in your repo? 🙏