> Blacklisted website in body, blacklisted website in title ---------- Title - Position 1-20: guideusermanual.com Body - Position 1-20: guideusermanual.com
@tripleee one of the blue ones here should be able to confirm if they can pull up a list of 'already resolved' red flags. IIRC something like that existed, and the GD mods can use it to nuke users.
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, link at end of body, pattern-matching product name in body, pattern-matching website in body: Magic Rock RX Male Enhancement by ruthevai on superuser.com
@Glorfindel I guess "What facial cream to use?", "Alphaplex you must lose weight by" and "brain power How will they" are the greatest hits, so I wouldn't be surprised if they posted some informative spam.
@Glorfindel maybe that's sufficient, I'm mainly stressed when I see all other recent Smokey notices go gray and the GD ones scroll off without greying (which is however now already unusual)
... plus seven which were shown as not having a "deleted_at" field but are deleted now
@SmokeDetector @ArtOfCode any idea how that was triggered? I see that the user uses his own name as a variable, I suspect that's it, but I don't see a URL.
I tend to nuke feedback conflict notifications once I've dealt with them so that the other admins don't come scrolling through later and try to resolve one that's already done
It would be nice if MS could serialize the entire Post object so that if we ever added a !!/test-ms command we wouldn't need to make additional API calls if MS doesn't have everything
@Cerbrus It's no where near as advanced as yours. Maybe I should just link you to the git repo first. I've probably made all sorts of userscript faux pas :P
I personally want to let Smoke Detector flag as many posts as possible with my account, so I've set my conditions so that I'm as close to the 99.5% accuracy limit as possible
But I noticed that over time, as more posts are incorporated into the testing corpus, the predicted accuracy of condition settings changes a little
What happens if I set my conditions so that today, they're just barely over the 99.5% accuracy minimum, but then in the future, without me changing my conditions, my settings end up having an estimated accuracy less than 99.5%?
Does my account continue flagging like normal? Is it pulled from the pool of accounts used for flagging? Or does something else happen?