Charcoal HQ

12:14 AM

!!/whois admin

@Undo I hit an error while trying to run that command; run !!/errorlogs for details.

!!/errorlogs 30

return _default_decoder.decode(s)

File "/usr/lib/python3.5/json/decoder.py", line 339, in decode
obj, end = self.raw_decode(s, idx=_w(s, 0).end())

File "/usr/lib/python3.5/json/decoder.py", line 357, in raw_decode
raise JSONDecodeError("Expecting value", s, err.value) from None

json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)
2018-05-24 00:14:47.390949 UTC
File "/home/ubuntu/ws/SmokeDetector/chatcommunicate.py", line 346, in f
result = func(*processed_args)

File "/home/ubuntu/ws/SmokeDetector/chatcommands.py", line 1001, in whois

o.O

it's because the filters are wrong

12:30 AM

... do we have a Slack channel?

Seeing a Slack integration to #charcoal from @Catija on the Charcoal team, curious what that is.

no we don't

WELZ

@Undo What's the problem with it?

chat.stackexchange.com/…

WELZ

@WELZ @WELZ @WELZ

Yes @Catija, I can still see it :D

Catija

@Undo That's a private notifications channel I was testing. For science! I'm not using it and it's really not good that it shows up for everyone. I've been talking somewhat with the Teams people about it. They're considering changing how they work but I'm guessing it's not high priority. I can forward you the emails if you're interested.

12:39 AM

Nah, I'm just curious. Not concerned about anything.

tpu- by Tetsuya Yamamoto on which wordpress plugin is used for seo? [MS]

[ SmokeDetector | MS ] Bad keyword with email in answer: Does Google log your search to your Google account? by Ally on webapps.SE

fp- by Tetsuya Yamamoto on How do I center a search bar in a nav, with elements on each side as well? [MS]

tpu- by Tetsuya Yamamoto

fp by Tetsuya Yamamoto on Couldn't Download! Could not retrieve data from URL [MS]

fp by Tetsuya Yamamoto on Could no retrieve data from URL [MS]

1:07 AM

[ SmokeDetector | MS ] Offensive answer detected, potentially bad keyword in answer, toxic answer detected: How can I check the condition of an SD card? by Linux Dr on unix.SE

tpu- by Makyen

[ SmokeDetector | MS ] Potentially bad keyword in body: Longitudinal study - conversion long to wide - implications by LeGeniusII on datascience.SE

fp- by Makyen

Merging MS changes for four flags; setting to 99.9

[ metasmoke-deploy ] deploy started by Undo

[ metasmoke ] ci/circleci success on 17f5703: Your tests passed on CircleCI!

"0,0,0,99.9,101,101"

^ That means "First three flags require 0 accuracy, a fourth flag requires 99.9% accuracy, and five or six flags is impossible (can't hit 101% accuracy)"

Max flags to four, updating site records to 4 except the non-fun sites.

max_flags_per_post set to 4 across the board, except SFF

We should send 4 flags on confident (>99.9%) reports. If you think it's doing something bad, !!/stopflagging & ping people.

12

MS admins - kill switch on this is to just set max_flags FlagSetting back to 3

(or, I guess, get Art to deploy something non-bad)

1:33 AM

[ SmokeDetector | MS ] Bad keyword with email in answer: Can I use Quikrete FastSet Repair Mortar to replace the mortar between stones in a fireplace hearth support? by Emily Cook Cookie Cooks on diy.SE

sd n

Catija

Is that "get back to me by email" thing... concerning?

Doesn't strike me as particularly odd for an NAA

Catija

Yeah... I don't know, something about the email sort of sounds phishy.

But whatever.

eh, a little weird I guess. But people have weird email addresses.

Fresh new metasmoke FR if anyone's looking for something to do...

1:43 AM

[ SmokeDetector | MS ] Pattern-matching website in answer: Who is Satoshi Nakamoto? by Dreams_Okorie on bitcoin.SE

[ SmokeDetector | MS ] Bad keyword with email in body: Chimenea/Outdoor Fireplace Crack Repair Mortar-type by Emily Cook Cookie Cooks on diy.SE

[ SmokeDetector | MS ] Korean character in title, mostly non-Latin body, mostly non-Latin title: 오피쓰☽강남오피ő ps『Š』₂。 ņET부평오피⇰부평건마⊔강남건마☮️동대문마사지☮️태릉밤문화⌟양재OP⥷상계안마 ∆ ≌안성휴게텔 ┌ 종로유흥 ≪목동키스방 by pettigrewtennille78 on gaming.SE

er

I'm going to guess that's spam, since it keeps mentioning massage

sd k

fp- by Makyen on Chimenea/Outdoor Fireplace Crack Repair Mortar-type [MS]

[ SmokeDetector | MS ] Blacklisted username, toxic answer detected: How do I prevent identity spoofing in a multiplayer game? by fuk u on gamedev.SE

tpu- by Makyen

2:39 AM

[ SmokeDetector | MS ] Korean character in title, mostly non-Latin body, mostly non-Latin title: 주식디비ヲ▶【스ㅋㅏㅇㅣ프 ＯＫＤＢ７９】◀ロ주식디비ヨ증권디비ま선물디비み해외선물む최신주식DBめ주식디비판매も by user391773 on meta.SE

tpu- by NobodyNada

NobodyNada

Google Translate isn't helpful, but it's helpful enough to reveal it's spam

tp by Undo

2:55 AM

Should we reevaluate metasmoke.erwaysoftware.com/post/120381? User has been nuked (not me); had several Indian-packer-mover posts from 2016

SmokeDetector's flag was marked helpful

Post on its own is probably fine.

tpu- by Tetsuya Yamamoto on Who is Satoshi Nakamoto? [MS]

!!/test cashprof.com

> Would not be caught as a post, title or username.

!!/watch cashprof\.com

@TetsuyaYamamoto Added cashprof\.com to watchlist

[ SmokeDetector ] SmokeDetector started at rev 9ce607a (SmokeDetector: Auto watch of cashprof\.com by Tetsuya Yamamoto --autopull) (running on ArtOfCode/EC2)

Restart: API quota is 17185.

tp by Tetsuya Yamamoto on How do I add the "Open With..." option to the right click menu for multiple file selections on Windows 10 64 bit? [MS]

tp by Tetsuya Yamamoto on What is the best program for 3D jewellery design and CAD Models? [MS]

tp by Tetsuya Yamamoto on C++ Project Error [MS]

3:06 AM

!!/test amanqq.com

> Potentially bad keyword in body, potentially bad keyword in title, potentially bad keyword in username
----------
Title - Position 1-11: amanqq.com
Body - Position 1-11: amanqq.com
Username - Position 1-11: amanqq.com

!!/test abellaescorts.com

> Pattern-matching website in body, pattern-matching website in title
----------
Title - Position 7-17: escorts.co
Body - Position 7-17: escorts.co

3:26 AM

!!/unwatch \wopss\d+\W*(com|net)\w

@Makyen Removed \wopss\d+\W*(com|net)\w from watchlist

!!/watch *o[^a-z0-9-]*p[^a-z0-9-]*s[^a-z0-9-]*s[^a-z0-9-]*\d+[^a-z0-9-]*(?:\d+[^a-z0-9-]‌*)*\w*(?:c[^a-z0-9-]*o[^a-z0-9-]*m|n[^a-z0-9-]*e[^a-z0-9-]*t)*

@Makyen Added *o[^a-z0-9-]*p[^a-z0-9-]*s[^a-z0-9-]*s[^a-z0-9-]*\d+[^a-z0-9-]*(?:\d+[^a-z0-9-]‌*)*\w*(?:c[^a-z0-9-]*o[^a-z0-9-]*m|n[^a-z0-9-]*e[^a-z0-9-]*t)* to watchlist

[ SmokeDetector | MS ] Potentially bad keyword in body: Will Facebook Pixel fire if user goes through blog? by Alex Lomov on stackoverflow.com

bwDraco

@Makyen sigh

fp- by Makyen

bwDraco

3:27 AM

Spammers have discovered that regexes are their worst enemy.

[ SmokeDetector ] SmokeDetector started at [rev 3bb8e7a (SmokeDetector: Auto watch of _*o[^a-z0-9-]*p[^a-z0-9-]*s[^a-z0-9-]*s[^a-z0-9-]*\d+[^a-z0-9-]*(?:\d+[^a-z0-9-]‌)*\w*(?:c[^a-z0-9-]*o[^a-z0-9-]*m|n[^a-z0-9-]*e[^a-z0-9-]*t)_* by Makyen --autopull*)](//github.com/Charcoal-SE/SmokeDetector/commit/3bb8e7a) (running on ArtOfCode/EC2)

Restart: API quota is 17033.

@bwDraco Yeah, that spammer is a bit much. They are obfuscating their domain name. Here's a metasmoke search for that one in titles.

3:42 AM

[ SmokeDetector | MS ] Bad NS for domain in answer, bad keyword in answer, blacklisted website in answer, pattern-matching website in answer: Where should I post questions about natural resources? by user391776 on meta.SE

tpu- by Tetsuya Yamamoto

[ SmokeDetector | MS ] Potentially bad keyword in body: Iphone/windows transfer by Loli on security.SE

@SmokeDetector why

@TetsuyaYamamoto Body - Position 409-449: hack my iphone through a cell phone call

fp- by Tetsuya Yamamoto

[ SmokeDetector | MS ] Offensive body detected: Error compiling because of prototypes by MIGGYEE on stackoverflow.com

@SmokeDetector why

fp- by Tetsuya Yamamoto

@TetsuyaYamamoto Post - Offensive keywords: shits, shits, shits, shits, shits, shits, shits, shits, shits, shits, shits, shits, shits, shits, shits

[ SmokeDetector | MS ] Potentially bad keyword in username: Will anyone here give me money? by Pussy Pleaser on math.SE

[ SmokeDetector | MS ] Mostly dots in answer: FM Modulation Index and Frequency trippler by user35926 on dsp.SE

3:59 AM

@SmokeDetector The offending string is part of directory path in OP's machine. Not offensive in this context.

tpu- by quartata on Will anyone here give me money? [MS]

@SmokeDetector n

[ SmokeDetector | MS ] Bad NS for domain in body, bad keyword in body, bad keyword in title, blacklisted website in body, link at end of body, +2 more: Keto Ultra best healthy way to lose weight by lauraerwin on meta.SE

tpu- by Tetsuya Yamamoto

fp by Nisse Engström on How do I center a search bar in a nav, with elements on each side as well? [MS]

tp by Nisse Engström on which wordpress plugin is used for seo? [MS]

[ SmokeDetector | MS ] URL-only title, bad NS for domain in body, bad NS for domain in title, bad keyword in body, bad keyword in title, +5 more: www.bluesupplement.com/keto-tone-shark-tank/ by pisaoreew on askubuntu.com

tpu- by Tetsuya Yamamoto

[ SmokeDetector | MS ] URL in title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +2 more: simp3.xyz/2-tone-drip-ybn-almighty-jay-mp3-song-download/ by autokime on english.SE

tpu- by Tetsuya Yamamoto

@bwDraco I don't think there's a lot of insight into specific techniques on the spammer side. Understanding what you are doing and what its consequences are would cause you to stop spamming

but there are guerrilla marketers who understand regex, bayesian filtering etc

I think they regard bayesian filters as the #1 deterrent that they need to work around

hence the spam posts with random word salad

and yet with proper feature engineering I think it could boost our accuracy significantly, too

4:18 AM

Henders/EC2 received failover signal.

[ SmokeDetector ] SmokeDetector started at [rev 3bb8e7a (SmokeDetector: Auto watch of _*o[^a-z0-9-]*p[^a-z0-9-]*s[^a-z0-9-]*s[^a-z0-9-]*\d+[^a-z0-9-]*(?:\d+[^a-z0-9-]‌)*\w*(?:c[^a-z0-9-]*o[^a-z0-9-]*m|n[^a-z0-9-]*e[^a-z0-9-]*t)_* by Makyen --autopull*)](//github.com/Charcoal-SE/SmokeDetector/commit/3bb8e7a) (running on ArtOfCode/EC2)

Restart: API quota is 19999.

Restart: API quota is 16776.

!!/standby henders

[ SmokeDetector ] SmokeDetector started in standby mode at rev 43ecf27 (metasmoke: Merge pull request #2164 from Charcoal-SE/auto-blacklist-1526177255.5076299 --autopull) (running on Henders/EC2)

@Makyen this is a dangerous regex, it should be refactored to avoid exponential backtracking

I suspect it might be responsible for the failover signal just now

John Dvorak

How good is Smokey's regex engine?

oh wait, nvm, I think it's safe actually

4:21 AM

@tripleee OK. you want to remove it, or have me unwatch it?

@JohnDvorak it's the Python regex module which is basically slated to replace the built-in re at some point ... "good" how exactly?

@Makyen no sorry, I think it's fine actually ... sloppy reading on my part

OK. np.

@JohnDvorak it's above average feature wise but shit performance wise

[ SmokeDetector | MS ] Bad keyword with a link in answer: How do I avoid the specification of the username and password at every git push? by Wiki Hotmail on stackoverflow.com

tpu- by tripleee

[ SmokeDetector | MS ] URL in title, bad keyword in body, bad keyword in title, potentially bad NS for domain in body, potentially bad NS for domain in title: videoconverter.media/cashing-out-suigeneris-mp3-song-download/ by delvont on english.SE

!!/watch hotmailwiki\.com

4:30 AM

tpu- by tripleee

@tripleee Added hotmailwiki\.com to watchlist

!!/watch (?:(?:g|hot)mail|outlook)\W*login

CI on 3841946 succeeded. Message contains 'autopull', pulling...

[ SmokeDetector ] SmokeDetector started at rev 759787c (SmokeDetector: Auto watch of (?:(?:g|hot)mail|outlook)\Wlogin by tripleee --autopull*) (running on ArtOfCode/EC2)

Restart: API quota is 16703.

[ SmokeDetector | MS ] Korean character in title, mostly non-Latin body, mostly non-Latin title, blacklisted user: 주식디비판매ヱ▶【텔그램ｏｋｄｂ７９】◀レ주식디비판매エ선물디비ハFX디비ヒ에프에스디비フ증권디비ヘ최신주식시세ホ by user391773 on meta.SE

tpu- by Tetsuya Yamamoto

Although really we don't demand much from a regex engine -- if it des trie optimizations then it's gonna run our stuff great

Can't we add some Korean characters to watchlist?

@AJ absolutely, if you know which ones to watch

4:35 AM

pretty sure SRE can do that

Though, Korean character detection works quite efficiently.

I managed to isolate a common Chinese phrase for "college certificate" (if I remember correctly) by comparing frequent character sequences in our Chinese spam collection

which is hence now blacklisted separately

@tripleee It has a keyword "Gangnam"

@AJ sounds like something we could probably even blacklist but you can certainly watch that

[ SmokeDetector | MS ] URL in title, potentially bad NS for domain in body, potentially bad NS for domain in title: youtubemp3.press/ay-ay-hakan-turkan-mp3-sarki-indir/ by lofkrir on puzzling.SE (@Mithrandir)

tpu- by Tetsuya Yamamoto

4:37 AM

!!/watch 강남

@tripleee That pattern looks like it's already caught by Mostly non-latin answer and Mostly non-latin body; append -force if you really want to do that.

@tripleee Here is Chinese character in title reason list. I dunno which characters are frequently used, some are shared with Japanese kanji.

these spammers arent even linking anything anymore theyre just furiously mashing characters together and praying to the SEO gods

@tripleee "텔그램ｏｋｄｂ７９" can also be watched/blacklisted or "okdb79". Though, a link is already under watch that I found in the title.

毕业证 is on line 396 of bad_keywords.txt

@AJ if it's blacklistable I would prefer that, but can you do the FP/TP check and SE search?

4:42 AM

@tripleee How about "成绩单"? It's often found in college spam posts.

@tripleee I think that's a tricky part. This Korean troll promotes new things every time.

@TetsuyaYamamoto Google Translate gives me "transcript", is this something which could occur in programming questions on the Chinese SO?

I thought we could watch Gangnam, but they have left promoting it long ago.

John Dvorak

Also there's that viral song...

@AJ we have some experience of that from the pharma spam gang(s), just heap on phrases into the watch list, then remove the ones which produce FPs, and promote the good ones to proper blacklist

and eventually clean out the no-ops from the watch list

4:44 AM

@JohnDvorak Gangnam is actually a city.

@tripleee "Transcript" sounds rarely used in programming context. Also with this candidate: "学历认证" (academic certificate).

zero hits in SE stackexchange.com/…

!!/test warmvape.com

> Would not be caught as a post, title or username.

59/59 and zero hits on SE metasmoke.erwaysoftware.com/…

blacklist time

4:46 AM

!!/watch warmvape\.com

!!/blacklist-keyword 学历认证

@TetsuyaYamamoto Added warmvape\.com to watchlist

@tripleee That pattern looks like it's already caught by Mostly non-latin answer and Mostly non-latin body; append -force if you really want to do that.

!!/blacklist-keyword-force 学历认证

@tripleee Blacklisted 学历认证

though most of the hits already have 500+ weight

4:48 AM

!!/test bouhannstore.com

[ SmokeDetector ] SmokeDetector started at rev 65cf442 (SmokeDetector: Auto blacklist of 学历认证 by tripleee --autopull) (running on ArtOfCode/EC2)

Restart: API quota is 16593.

!!/test bouhannstore.com

> Would not be caught as a post, title or username.

!!/watch bouhannstore\.com

@TetsuyaYamamoto Added bouhannstore\.com to watchlist

4:49 AM

Found an entry of the site from here.

It looks like this Korean troll attacks for a couple of days like in every month.

CI on d27e4a4 succeeded. Message contains 'autopull', pulling...

They have a pattern of including a link or particular name of thing they want to promote in the title.

[ SmokeDetector ] SmokeDetector started at rev d27e4a4 (SmokeDetector: Auto watch of bouhannstore\.com by Tetsuya Yamamoto --autopull) (running on ArtOfCode/EC2)

Restart: API quota is 16584.

!!/watch okdb79

4:52 AM

[ SmokeDetector ] SmokeDetector started at rev ac0cb1f (tripleee: Add press to MP3 regex --autopull) (running on ArtOfCode/EC2)

2 TPs yet

@AJ You don't have code privileges, but I've created PR#2188 for you.

!!/watch youtubemp3\.press

Restart: API quota is 16575.

@tripleee That pattern looks like it's already caught by Potentially bad ns for domain in body and Potentially bad ns for domain in answer; append -force if you really want to do that.

Merged SmokeDetector #2188.

CI on 1b97c9b succeeded. Message contains 'autopull', pulling...

@tripleee How to watch a keyword if they use non-English characters in links such as this one and this

4:54 AM

[ SmokeDetector ] SmokeDetector started at rev 1b97c9b (metasmoke: Merge pull request #2188 from Charcoal-SE/auto-blacklist-1527137540.0515733 --autopull) (running on ArtOfCode/EC2)

Restart: API quota is 16558.

There is a word "opss" in there

@AJ sounds like it would easily misfire on Russian links etc ... but those are not proper links, maybe you can frame them with some context conditions

though I guess the monster regex Makyen submitted recently might already cover that

@tripleee I have tried to watch some words that can easily be caught, but others are typical. metasmoke.erwaysoftware.com/reason/9

!!/test okdb79

> Potentially bad keyword in body, potentially bad keyword in title, potentially bad keyword in username
----------
Title - Position 1-7: okdb79
Body - Position 1-7: okdb79
Username - Position 1-7: okdb79

!!/test OKDB79

5:00 AM

> Potentially bad keyword in body, potentially bad keyword in title, potentially bad keyword in username
----------
Title - Position 1-7: OKDB79
Body - Position 1-7: OKDB79
Username - Position 1-7: OKDB79

ｏｋｄｂ７９ in many of the samples uses full-width characters

!!/watch ｏｋｄｂ７９

@tripleee Added ｏｋｄｂ７９ to watchlist

$ xxd <<<'ｏｋｄｂ７９ '
00000000: efbd 8fef bd8b efbd 84ef bd82 efbc 97ef  ................
00000010: bc99 200a                                .. .

so ef bd 8f is the full-width "o" in UTF-8, etc

CI on 066fc72 succeeded. Message contains 'autopull', pulling...

[ SmokeDetector ] SmokeDetector started at rev 066fc72 (SmokeDetector: Auto watch of ｏｋｄｂ７９ by tripleee --autopull) (running on ArtOfCode/EC2)

@tripleee Oh.

5:05 AM

Restart: API quota is 16499.

looks like maybe the Metasmoke search does some normalization, not sure if Smokey can match just okdb79 ... but let's keep both in watch for a while and see what happens

[ SmokeDetector | MS ] Bad keyword in body, bad keyword in link text in body, blacklisted website in body, pattern-matching website in body, potentially bad keyword in body: how can custom essays be useful by LindaSmith on ux.SE

tpu- by tripleee

@tripleee oh okay

!!/blacklist-website traditionessaysonline\.com

@tripleee That pattern looks like it's already caught by Pattern-matching website in answer, Blacklisted website in answer, Pattern-matching website in body, and Blacklisted website in body; append -force if you really want to do that.

[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, body starts with title and ends in URL, link at end of body, pattern-matching website in body, +2 more: The choice would be simple: other than by xlodehose3 on askubuntu.com

tpu- by tripleee

5:15 AM

should we have a tag for "celebrity wealth check" or something like that, I recall we used to have a number of spams like this? metasmoke.erwaysoftware.com/domains/21491

Well, the one I watched earlier could actually be blacklisted, as it has 11 hits and no FP. If it could go on the websites blacklist I would have done so, but SD doesn't recognized it as a website due to obfuscation. I'm also not much of a fan of directly blacklisting a relatively complex/loose regex without having it spend at least a bit of time on the watchlist, just to make sure nothing's severely broken.

sounds good

[ SmokeDetector | MS ] Korean character in title, mostly non-Latin body, mostly non-Latin title, potentially bad keyword in body, potentially bad keyword in title, +1 more: 주식디비팝니다ウ▶【탤래그렘 okDB79】◀ル주식디비팝니다ユ주식시세디비は증권디비ひ선물디비ふ주식DB판매へ최신주식DBほ by user391773 on meta.SE

tpu- by tripleee

[ SmokeDetector | MS ] Bad keyword in body, pattern-matching website in body: why wordpress is important for bussiness website? by Tricity Infotech on rpg.SE

tpu- by tripleee

!!/watch tricityinfotech\.com

@tripleee That pattern looks like it's already caught by Pattern-matching website in answer and Pattern-matching website in body; append -force if you really want to do that.

5:24 AM

^ this Korean spam didn't have the full-width characters so probably good to keep both variants ...?

@tripleee That's a good idea. We need to catch it if they do it in either way.

5:42 AM

[ SmokeDetector | MS ] URL in title, bad NS for domain in body, bad NS for domain in title, bad keyword in body, bad keyword in title: videoconverter.media/cashing-out-suigeneris-mp3-song-download/ by Ishikil on graphicdesign.SE

tpu- by suraj

[ SmokeDetector | MS ] Potentially bad keyword in body: How to make Onion Raita by Being Nawab on stats.SE

tpu- by Glorfindel

[ SmokeDetector | MS ] Bad keyword in link text in body, potentially bad keyword in body: Why do we need reference when doing essay writing? by Hank Han on english.SE

tpu- by Tetsuya Yamamoto

[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, link at end of body, pattern-matching website in body, +1 more: Lumineri Skin - Does it Really Works? by user833457 on meta.SE

tpu- by Tetsuya Yamamoto

!!/test viniscookbook.com

> Potentially bad keyword in body, potentially bad keyword in title, potentially bad keyword in username
----------
Title - Position 1-18: viniscookbook.com
Body - Position 1-18: viniscookbook.com
Username - Position 1-18: viniscookbook.com

@SmokeDetector k

6:03 AM

[ SmokeDetector | MS ] Bad keyword in answer: How to Record Screen in 4K on 1920x1080 monitor? (OBS Studio) by monica66 on video.SE

[ SmokeDetector ] SmokeDetector started at rev 7872f61 (tripleee: Add .media to MP3 TLDs --autopull) (running on ArtOfCode/EC2)

Restart: API quota is 16135.

[ SmokeDetector ] SmokeDetector started at rev bf9ba4a (tripleee: Add .download to MP3 TLDs --autopull) (running on ArtOfCode/EC2)

tpu- by Zoe

Restart: API quota is 16128.

!!/test SlimForce Forskolin

[ SmokeDetector | MS ] Bad NS for domain in body, bad keyword in body, bad keyword in title, blacklisted website in body, link following arrow in body: LunaTrim Exercise at home by karenbormann on apple.SE

> Bad keyword in body, bad keyword in title, bad keyword in username
----------
Title - Position 11-20: Forskolin
Body - Position 11-20: Forskolin
Username - Position 11-20: Forskolin

tpu- by Tetsuya Yamamoto

tpu- by Tetsuya Yamamoto on How to Record Screen in 4K on 1920x1080 monitor? (OBS Studio) [MS]

[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, link at end of body, repeated URL at end of long post: And, if you’re curious about it, don’t by agner nally on workplace.SE

!!/watch-force (?:안마|마사지|알몸사진|애인대행)

@Makyen That pattern looks like it's already caught by Mostly non-latin answer and Mostly non-latin body; append -force if you really want to do that.

tpu- by Tetsuya Yamamoto

@Makyen Added (?:안마|마사지|알몸사진|애인대행) to watchlist

[ SmokeDetector ] SmokeDetector started at rev c6137f1 (SmokeDetector: Auto watch of (?:안마|마사지|알몸사진|애인대행) by Makyen --autopull) (running on ArtOfCode/EC2)

Restart: API quota is 16093.

[ SmokeDetector | MS ] Bad keyword in link text in body, pattern-matching website in body: Somebody to Buy Pinterest Followers? by user391787 on meta.SE

tpu- by Tetsuya Yamamoto

!!/test bestcheaplikes.com

6:10 AM

> Pattern-matching website in body, pattern-matching website in title
----------
Title - Position 1-18: bestcheaplikes.co
Body - Position 1-18: bestcheaplikes.co

[ SmokeDetector | MS ] Korean character in title, mostly non-Latin body, mostly non-Latin title, potentially bad keyword in body, potentially bad keyword in title, +1 more: 디비판매ヰ▶【스카ㅇㅣ프 ｏｋｄｂ７９】◀ リ디비판매イDB판매ナ최신DB판매ニ각종DB판매ヌ디비문의ネ디비구매ノ by user391773 on meta.SE

tpu- by Tetsuya Yamamoto

Martin Sleziak

!!/report mathoverflow.net/questions/300986/…

[ SmokeDetector | MS ] Manually reported answer: How much joyful to choose a best place for your vacation? by trifonkryukov on mathoverflow.net

Martin Sleziak

!!/report mathoverflow.net/questions/300986/…

tpu- by Tetsuya Yamamoto

!!/watch exploresrilanka\.com

6:17 AM

[ SmokeDetector | MS ] Manually reported question: How much joyful to choose a best place for your vacation? by trifonkryukov on mathoverflow.net

@TetsuyaYamamoto Added exploresrilanka\.com to watchlist

tpu- by Tetsuya Yamamoto

CI on cb82278 succeeded. Message contains 'autopull', pulling...

[ SmokeDetector ] SmokeDetector started at rev cb82278 (SmokeDetector: Auto watch of exploresrilanka\.com by Tetsuya Yamamoto --autopull) (running on ArtOfCode/EC2)

@SmokeDetector Still categorized as spam because the title exactly same as previous report.

Martin Sleziak

BTW again the (traveling-salesman-problem) tag. There was a short discussion about tags yesterday. chat.stackexchange.com/…

Restart: API quota is 16009.

tp by Nisse Engström on SlimForce Forskolin Thin SlimForce Forskolin [MS]

[ SmokeDetector | MS ] Username similar to website in answer: python fabric output hangs by Tej on stackoverflow.com

fp- by Tetsuya Yamamoto

[ SmokeDetector | MS ] Potentially bad keyword in username: how to load single row in list view android by Umar on stackoverflow.com

fp- by Tetsuya Yamamoto

@SmokeDetector fp-

[ SmokeDetector | MS ] Manually reported answer: Custom top level category landing page in magento by bhavya on stackoverflow.com

tpu- by Tetsuya Yamamoto

6:30 AM

!!/watch instabuilder\.com

!!/watch instabuilder\.com

Dammit

@Zoe Added instabuilder\.com to watchlist

@TetsuyaYamamoto Added instabuilder\.com to watchlist

I hate when this happens ._.

[ SmokeDetector | MS ] Bad keyword in answer, blacklisted website in answer: What happened to netatalk? MacOS<>Ubuntu filesharing by Evelyn on askubuntu.com

CI on 9b2db4e failed.

tpu- by Tetsuya Yamamoto

6:35 AM

@SmokeDetector 2190

@Zoe Just !!/unwatch

@Makyen Can't.

It's not pulled

!!/unwatch instabuilder\.com

@Zoe No such item instabuilder\.com in watchlist.

Suraj Rao

@Makyen will it remove only one instance in this case?

Duplicate items need to be manually removed in cases like this

6:38 AM

@SurajRao Not sure. The last test I remember it removed all.

Suraj Rao

It should do the same while adding ideally...

maybe include a check to not add if already there

hang on, I'll do a manual fix

@SurajRao I think this happens in two separate branches so it's not in the current branch until they are merged

maybe we would need to refactor it so it checks after merge and tolerates but ignores duplicates

CI on eb441e4 succeeded. Message contains 'autopull', pulling...

[ SmokeDetector ] SmokeDetector started at rev eb441e4 (tripleee: -duplicate --autopull) (running on ArtOfCode/EC2)

@Zoe Have you previously tried !!/pull?

Restart: API quota is 15821.

6:41 AM

hmmm, binfer.com is Microsoft ...?

in Charcoal Test, 7 mins ago, by PulseMonitor

halflife:2018-05-24 06:31:38,628:120792: binfer.com: ip 52.151.43.114 AS 8075 (MICROSOFT-CORP-MSN-AS-BLOCK - Microsoft Corporation, US/US)

I guess it just means Azure?

@Makyen Duplicates makes the CI build fail. The build failing means it can't pull. Since it can't pull, and since pulling is required for the watch to be added, it can't be !!/unwatched

the whois record certainly doesn't look like it's official MS

@Zoe That's more or less what I'd assumed, but hadn't exhausted all possible permutations which didn't require a manual fix. :-)

:)

It'd be nice if there was a !!/unwatch-force which could be used to resolve the issue.

6:46 AM

@tripleee There's a chance that the site just use Azure hosting, the site description refers to file transfer/sharing software.

!!/unwatch \wh+a+\Wh+a+\Wh+a+\w*(?!(?:\n|.)?</code>)

@Makyen Removed \wh+a+\Wh+a+\Wh+a+\w*(?!(?:\n|.)?</code>) from watchlist

!!/watch \wh+a+\Wh+a+\Wh+a+\w*(?!(?:[^<]|<[^c]|<c[^o]|<co[^d]|<cod[^e]|<code[^>])?</c‌ode>)

@Makyen Added \wh+a+\Wh+a+\Wh+a+\w*(?!(?:[^<]|<[^c]|<c[^o]|<co[^d]|<cod[^e]|<code[^>])?</c‌ode>) to watchlist

JAD

wha?

6:48 AM

hahaha, but not in code. Reduces FP by >50%

[ SmokeDetector ] SmokeDetector started at [rev bafeef7 (SmokeDetector: Auto watch of \wh+a+\Wh+a+\Wh+a+\w*(?!(?:[^<]|<[^c]|<c[^o]|<co[^d]|<cod[^e]|<code[^>])*?</c‌ode>) by Makyen --autopull*)](//github.com/Charcoal-SE/SmokeDetector/commit/bafeef7) (running on ArtOfCode/EC2)

Videonauth

[Ww]h+a+ .....

no that's \w which is word, and \W which is non-word

Restart: API quota is 15768.

Videonauth

ah ok

not really awake already :)

6:55 AM

[ SmokeDetector | MS ] Bad keyword with a link in answer, potentially bad keyword in answer, potentially bad keyword in username: What are hiberfil.sys and pagefile.sys files, and can I safely delete them? by hotmail login on superuser.com

tpu- by Tetsuya Yamamoto

Suraj Rao

^do all spammers follow a common template.. Not that its a bad thing

It's easy to fall into a pattern. If it works for them, why not keep using it?

^

[ SmokeDetector | MS ] Link at end of answer: Refactoring JavaScript and PHP code [Job Interview] by Sunil on stackoverflow.com

tpu- by PeterJ

!!/watch phpinterview\.in