Wikipedia:Bots/Noticeboard

Bots noticeboard

Here we coordinate and discuss Wikipedia issues related to bots and other programs interacting with the MediaWiki software. Bot operators are the main users of this noticeboard, but even if you are not one, your comments will be welcome. Just make sure you are aware about our bot policy and know where to post your issue. Do not post here if you came to discuss non-urgent bot issues, bugs and suggestions for improvement. Do that at the bot operator's talk page discuss urgent/major bot issues. Do that according to instructions at WP:BOTISSUE discuss general questions about the MediaWiki software and syntax. We have the village pump's technical section for that request approval for your new bot. Here is where you should do it request new functionality for bots. Share your ideas at the dedicated page

User:B-bot

B-bot didn't tag any files for F5 deletion today when it usually tags several dozen per day (see Category:Orphaned non-free use Wikipedia files). B hasn't been around for a few days so I thought I'd bring it up here in case the situation applies to other bots. Liz ^{Read! Talk!} 23:14, 24 July 2024 (UTC)

B-bot seems to run at 2:00 and 17:00 daily (judging from its internal lists at User:B-bot/List of orphaned images/day-0, .../day-1, and .../day2). It pulls its data from Wikipedia:Database reports/Unused non-free files, which updated at 16:32 yesterday, just before the B-Bot run, and 20:21 today, well after it. So I wouldn't start to worry unless B-bot doesn't do anything at 2:00. —Cryptic 23:32, 24 July 2024 (UTC)

Okay, thank you for the information, Cryptic. I think it tagged 96 files yesterday so it was suprising to see 0 today. Liz ^{Read! Talk!} 01:04, 25 July 2024 (UTC)

This bot is again dead for a week. 185.172.241.184 (talk) 13:42, 12 August 2024 (UTC)

Perhaps this is because of the replag in phab:T367856. If so, it'll come back online shortly without needing intervention. –Novem Linguae (talk) 14:33, 12 August 2024 (UTC)

IP running a bot

I blocked IP 5.14.154.187 (talk · contribs · WHOIS) for disruptive editing via a report at AIV and also because they appeared to be running an unapproved bot - bot-like edits and they claimed to be running a bot[1]. I left a message about running an unapproved bot on the IP talk page with a link to the bot policy and 7heGame has popped up saying "You can’t do that"; I assume they are the bot-operator. Does anything more need to be done? -- Malcolmxl5 (talk) 18:12, 27 July 2024 (UTC)

Good block, both because it was obviously some sort of vandalism (same three images added on every draft I checked) and because the user said it was a bot. No idea about 7heGame. Anomie⚔ 18:18, 27 July 2024 (UTC)

Yup, that's a vandal (not convinced about the bot part btw). I would treat 7heGame as innocent; they had one of their drafts vandalised by the IP, and it's not a surprising response. -- zzuuzz ^(talk) 18:25, 27 July 2024 (UTC)

I agree it's unlikely to actually be a bot. — Qwerfjkltalk 19:21, 27 July 2024 (UTC)

• On reflection, 7heGame is a victim and I’ve struck out my mention of them. Nothing more needs to be done then (except keep an eye on the IP). Thanks all. --Malcolmxl5 (talk) 19:58, 27 July 2024 (UTC)

ST47ProxyBot

Is ST47ProxyBot still functioning? I’m thinking it might be useful in curbing the current wave of proxies plaguing the admin boards. -- Malcolmxl5 (talk) 09:21, 2 August 2024 (UTC)

User:ST47ProxyBot has not made any blocks since March 2024. You can follow up with the operator on their talk. — xaosflux ^Talk 10:27, 2 August 2024 (UTC)

Thanks. -- Malcolmxl5 (talk) 11:19, 2 August 2024 (UTC)

Assuming @ST47 isn't willing to maintain this bot task anymore, I would be interested in picking it up. -Fastily 03:35, 5 August 2024 (UTC)

@Fastily I think we should move in parallel while waiting for ST47 to reply on their talk page. The admin boards should not be locked for so long, and there are million of IP addresses to parse and blocked if necessary. – robertsky (talk) 04:56, 6 August 2024 (UTC)

@Robertsky I emailed ST47 to ask for details about how their bot works. I've done some cursory research and there appears to be a plethora of vpn/open proxy services. Would be curious to get folks' thoughts on preemptive blocking of IPs vs reactive (e.g. monitoring RecentChanges and blocking suspicious IPs). Also worth noting that WMF T&S is working on mw:Trust and Safety Product/Temporary Accounts which could complicate things. -Fastily 05:39, 7 August 2024 (UTC)

@Fastily I think for a start we can do a reactive one for the admin boards. Block and revert any IP address with proxy behaviors. Temporary accounts, we can cross the bridge when it comes. – robertsky (talk) 05:54, 7 August 2024 (UTC)

@Robertsky: Sounds good, I've asked for feedback at Wikipedia:Village pump (idea lab)#Bot to block Proxy/VPN IPs (ST47ProxyBot replacement). If there is enough community support for the idea, I'm happy to work on this. -Fastily 21:27, 8 August 2024 (UTC)

@ST47 thank you for all the work and time you've put into maintaining the bot. From our meeting last year I could see that this is a lot of effort, and I'm thankful that you kept it going as long as you did.

@Fastily a few thoughts:

• it would be interesting to compare some indicators (reverts, blocks, suspected vandalism) in the period with and without ST47ProxyBot in action. I'm curious if we can quantify the real world impact of the bot placing the blocks based on Spur data
• We now have https://wikitech.wikimedia.org/wiki/Service/IPoid running which has many more IP addresses available. We have a research project underway to analyze how edits and account creations associated with IPs in Spur data (T360195) are correlated to on-wiki activity. That would be interesting to look at before making more decisions on how to use the IP address data for blocking or other mitigations (like showing a CAPTCHA, via T20110)
• it would probably be more maintainable to use a MediaWiki extension that can talk directly to IPoid service, and make it configurable via CommunityConfiguration, to administer any type of automatic blocks for IPs based on Spur data
• mw:Trust_and_Safety_Product is working on a few projects this year that might allow us to reduce our dependency on IP blocking, and instead use IP reputation as an additional signal among others-- unique device model project (https://phabricator.wikimedia.org/tag/fy24-25_we4.2.1_unique_device_identification_model/), account reputation scores (https://phabricator.wikimedia.org/project/view/7279/), and improved bot detection (https://phabricator.wikimedia.org/project/view/7292/).

@Fastily @Robertsky could we have a chat before bringing back the bot, to discuss what we could improve, what we might want to change, how it can be implemented to minimize burden on community and get the most out of using WMF infrastructure, etc? KHarlan (WMF) (talk) 06:17, 15 August 2024 (UTC)

I do not intend to resume operation of these bots. Modern anonymization infrastructure addresses are too dynamic for a daily dump from spur and a single VM running port scans. WMF has the capability and means to implement a solution similar to mw:Extension:TorBlock based on threat intelligence feeds, as does any major website for at least the last decade. I have on several occasions passed details on my past proxy blocking data sources to WMF staff. Blocking anonymous editing and broadening the use of account creation blocks for known hosting and CGNAT ranges is likely the best stopgap solution. ST47 (talk) 05:20, 6 August 2024 (UTC)

@ST47 i am curious here, is there a phab ticket to see if there's work done by the devs? – robertsky (talk) 05:29, 6 August 2024 (UTC)

I have never been given security access to Phabricator, so I am only aware of tickets that I have been invited to. One is T265845. It has received no substantive discussion since 2021, and remains "Needs Triage". ST47 (talk) 05:41, 6 August 2024 (UTC)

Looks like Robertsky is investigating if the IP information tool's data can be used to block proxies in phab:T371870. –Novem Linguae (talk) 07:46, 6 August 2024 (UTC)

Thanks for sharing. I also have opened meta:Community_Wishlist/Wishes/Blocking_proxy_IP_address. Let's hope that we can push this through. – robertsky (talk) 08:07, 6 August 2024 (UTC)

I'd suggest we follow up in T354599, which proposes to allow using IP reputation data from Spur in AbuseFilter. KHarlan (WMF) (talk) 15:58, 14 August 2024 (UTC)

@ST47: confirming, User:ST47ProxyBot is being retired and can be deflagged (-bot, -sysop)? (A new BRFA would be needed to reactivate). — xaosflux ^Talk 13:47, 9 August 2024 (UTC)

@Xaosflux: Yes, that's correct. ST47 (talk) 15:56, 14 August 2024 (UTC)

BRFA's adminbots page

Wikipedia:Bots/Requests_for_approval/Adminbots exists as a place to list open BRFAs regarding admin bots. The page is technically connected to the bots header.

I'm not sure how much we really use it as of recent, how able we are to keep it updated (for one, there's an adminbot in trial atm not listed there), and whether it has a purpose? Should we retire this page and remove from the header, or if it has use, how do we make sure it's used correctly? ProcrastinatingReader (talk) 15:09, 17 August 2024 (UTC)

I think mostly it's not used much because we have very few adminbot BRFAs, and it doesn't seem to be mentioned at WP:ADMINBOT or in any instructions at WP:BRFA either, just that header template. OTOH, I'm not sure how much point there is to it either. The main point, I guess, would be for people who're interested in new adminbot BRFAs to watch. But how many such people there are I don't know; Special:PageInfo/Wikipedia:Bots/Requests for approval/Adminbots doesn't show that many watchers. Anomie⚔ 18:17, 17 August 2024 (UTC)