Feature Request: Flag instead of vote.

[Cam-B]

I know you are concerned about flag bots in the wrong hands. I agree a flag bot has potential to do much damage but I would rather see a flag/spam bot for everyone instead of encourage people to delegate SP to "popular" spam fighters.

I imagine a good use case for now would be to set the algorithm to flag:

Blacklisted accounts that are repeating the same or almost identical comment in high volume.

The typical. "I upvoted you. Follow me please" for example, is something I would set I bot to flag If the account has posted this comment more then maybe 5 times that day.

I think we should be very careful we don't misconfigure our bots and let them run wild unattended, but I do think this would be a good approach to decentralizing spam control even further.

[ghost]

Adding a flag will not get you this use case because there are no metrics to support this. You would need at least this:

1. Processing comments (would be implemented in Vote on Comments - Feature request #104)
2. Content checking for either self similarity, blacklisted words or phrases, or similar

Point 2 needs to be developed.

Unless you can develop this example or come up with an alternative example that uses the existing metrics, I do not support adding flagging because I am not convinced it will be useful.

[Cam-B]

For point 2 I was thinking to use existing metrics to scan for comments by blacklisted authors containing a blacklisted phrase/string of keywords.

Now I'm not sure if the default option "content word blacklist" allows blacklisting of phrases. Does it only work with single keywords?

[ghost]

Currently it only works with single words, but this is just because the list UI is pretty basic. It could be updated to handle phrases.

[Cam-B]

That would be pretty sweet. I'd try to do it but I have no idea how.