SmokeDetector Info for Moderators (SIM) update: Version 0.7.0 (GitHub) (install). Add a "Report post" button which will post a !!/report for the post in Charcoal HQ.
The quota_remaining property was not in the API response. Error: no method found with this name at 02:31:11 UTC. Backing off on requests for the next 12 seconds. Previous URL: https://api.stackexchange.com/2.2/questions
@SmokeDetector This edit is just from one useless piece of filler text to another piece of useless filler text. While the filler text shouldn't be permitted, I can't really say this is actually vandalism. OTOH, the easy evaluation from a machine learning POV is that the edit should be blocked, as it does go from words that make sense (but are useless filler) to just repeating "a" characters.
@CodyGray That pattern looks like it's already caught by Potentially bad keyword in body and Potentially bad keyword in answer Append -force to the command word(s) if you really want to add the pattern you provided.
First instinct is omit the "facebook.com" part and just watch the slug. It looks unique enough.
The only issue is whether bookending or some other footgun exists that would prevent that from matching as expected. And.... I never know the answer to that.
Basically... as far as I can tell, the "answer" contains both a Thai version and an English translation, confusingly interspersed. But it seems to be some sort of attempt to answer (but still NAA?)
Except that it's rude with phrases like "including the 9th brother who shot the 8th brother" (which appears in the original Thai, too)
Well... that's the name of their file on GitHub?
Even if it was spam, I don't think that's something we'd expect to recur, since it's just an arbitrarily chosen file name
But, I'm admittedly confused about the whole mess.
So... what was the other thing we were discussing? I'm trying [and failing] to do too many things at once.
OK, yeah, that Facebook thing. We're good with a naive watch: https://metasmoke.erwaysoftware.com/search?utf8=%E2%9C%93&title_is_regex=1&title=(%3Fs)windows11freedownloadactivatediso(%3F%3C%3D(%3F%3A%5E%7C%5Cb)windows11freedownloadactivatediso)(%3F%3A%5Cb%7C%24)&body_is_regex=1&body=(%3Fs)windows11freedownloadactivatediso(%3F%3C%3D(%3F%3A%5E%7C%5Cb)windows11freedownloadactivatediso)(%3F%3A%5Cb%7C%24)&username_is_regex=1&username=(%3Fs)windows11freedownloadactivatediso(%3F%3C%3D(%3F%3A%5E%7C%5Cb)windows11freedownloadactivatediso)(%3F%3A%5Cb%7C%24)&or_search=1
[ SmokeDetector | MS ] Blacklisted website in body, link at beginning of body, mostly non-latin body, mostly non-latin title, potentially bad ip for hostname in body, +4 more (302): Pgslotสล็อตเว็บตรง by PG slotgame on travel.SE
[ SmokeDetector | MS ] Blacklisted website in body, link at beginning of body, mostly non-latin body, mostly non-latin title, potentially bad ip for hostname in body, +4 more (302): Pgslotสล็อตเว็บตรง by PG slotgame on travel.SE
@metasmoke Would it be feasible to exclude mathematical symbols from the emoji detection? @Makyen
As long as we get a spamwave involving emoji every so often, the confidence score will stay pretty high. But I don't think the filter was meant to catch those.
Especially on the math-related sites we'll continue to get FP by it.
@Mast if there isn't a defined subset within the regex library it might be a nightmaries solution to maintain our own, let alone hope for no performance impact.
It has probably happened, maybe a handful of times, but were those posts we wanted to catch with this reason?
AFAIK, this was mainly implemented to catch telephones, christmas trees, yellow stars and the other crap those quickbooks / coinbase / binance programs threw at us.
Where it was simply not possible to watch for single icons since they switched it up every post.
@EJoshuaS-StandwithUkraine The !!/blacklist command has been deprecated. Please use !!/blacklist-website, !!/blacklist-username, !!/blacklist-keyword, or perhaps !!/watch-keyword. Remember to escape dots in URLs using \.
