Wikipedia:Bot requests

I'm requesting that a task be created for a bot to perform the following action per MOS:POSLINK that has very little chance of having any false positives:

Replace [[Foo|Foo's]] with [[Foo]]'s.

This can probably be a task that runs an indefinite amount of times, probably once every month or so. Steel1943 (talk) 21:38, 31 January 2026 (UTC)

Is there any consensus for a bot to mass replace these for MOS:POSLINK though? Tenshi! (Talk page) 21:43, 31 January 2026 (UTC)

WP:BOTREQUIRE doesn't require that, and the content of MOS:POSLINK was formed by a RfC. But, if that route needs to be taken, by all means, bureaucracy away! Steel1943 (talk) 21:48, 31 January 2026 (UTC)

This seems like a WP:COSMETICBOT to me. * Pppery * _{it has begun...} 21:52, 31 January 2026 (UTC)

(edit conflict) Interesting, didn't know that existed. (But, then again, I don't read the bot policy page that much.) Whatever happens here happens, but either way, this policy was recently formed via consensus and this type of change has ... almost no false positive chance. Steel1943 (talk) 21:55, 31 January 2026 (UTC)

Technically not WP:COSMETICBOT, as it does make a difference to the rendering of the link. Anomie⚔ 00:22, 1 February 2026 (UTC)

It does not make a difference in the rendering of the link. Vanderwaalforces (talk) 05:38, 1 February 2026 (UTC)

It does, though. Foo's is different from Foo's. —Myceteae‍🍄‍🟫 (talk) 06:06, 1 February 2026 (UTC)

@Vanderwaalforces: It may be difficult to see, but notice the differences between how the "'s" is linked in both the links Myceteae made above. Steel1943 (talk) 20:18, 1 February 2026 (UTC)

Maybe I was understanding "rendering" differently. But why I said there's no difference is that both of them have the same href= and title= values, only the display text is different. Vanderwaalforces (talk) 21:30, 1 February 2026 (UTC)

COSMETICBOT rules are primarily about whether the rendered HTML differs between the two edits. Removing whitespace does not affect how the page is rendered. Changing the formatting of a URL from Foo's to Foo's does affect the HTML output of the page. In my opinion, this does not count as a cosmetic edit. phuzion (talk) 17:07, 2 February 2026 (UTC)

(edit conflict) I'm not objecting to this, was just wondering if there was or not. Tenshi! (Talk page) 21:53, 31 January 2026 (UTC)

Oh, I was thinking that your question was based off of it being some kind of prerequisite to perform such a task (as I would've expected from a place that has editors that are probably rather technical.) My apologies for the misunderstanding. Steel1943 (talk) 21:57, 31 January 2026 (UTC)

As I understand, there is no requirement that bot requests must have consensus beforehand, although for contentious tasks (e.g. an adminbot that deletes G13'd drafts) it would likely need a consensus to run such a bot. Tenshi! (Talk page) 22:09, 31 January 2026 (UTC)

I think for a task that is borderline COSMETICBOT like this one, even if not strictly required, it would be a good idea to get consensus beforehand. — Qwerfjkltalk 14:10, 1 February 2026 (UTC)

It looks like there was an RFC that created a subsection of the MOS that indicates that we shouldn't have the 's inside of links. I'd say that's a pretty good consensus. I'll see if I can come up with something. Primefac (talk) 20:18, 1 February 2026 (UTC)

I'd say that's a pretty good consensus. Agreed. This bot would straightforwardly correct clear cut violations of MOS:POSLINK, a recent addition to the MOS which made following a well-attended and near-unanimous RFC. Thanks @Primefac for working on this and thank you to @Steel1943 for suggesting this fix. —Myceteae‍🍄‍🟫 (talk) 20:38, 1 February 2026 (UTC)

Any updates on this? Steel1943 (talk) 18:52, 10 February 2026 (UTC)

No. It's still on my list. Primefac (talk) 11:44, 11 February 2026 (UTC)

Restoring thread from Archive 88. I had forgotten about this until I go the notification that this thread I subscribed to had been archived. @Primefac: is this still on your radar? —Myceteae‍🍄‍🟫 (talk) 23:12, 7 May 2026 (UTC)

Seems to have fallen off, but I can put it back on my list. (one of these days I'll actually write down said list...) Primefac (talk) 12:03, 8 May 2026 (UTC)

Thanks! I appreciate it. FYI @Steel1943 —Myceteae‍🍄‍🟫 (talk) 14:32, 8 May 2026 (UTC)

This could be added to Wikipedia:AutoWikiBrowser/Typos, I believe. Not sure how to search for [[(...)|\1's]] or [[(...)|$1's]] using mw:Help:CirrusSearch TBH. Ponor (talk) 15:59, 8 May 2026 (UTC)

There might be a bit of a CONTEXTBOT issue... I'm still refining my regex search but in the first 20 hits was [[Moody's Ratings|Moody's]] which (correctly) shows up as Moody's but would be "incorrectly" shown as Moody's if the task were to run. I'm also pretty sure that Broca's is correct but not 100% confident on that one, but if it is then that's a 10% failure rate which is usually my cut-off for feeling comfortable with a bot run. Primefac (talk) 10:38, 9 May 2026 (UTC)

Interesting problems, thanks for sharing, and for working on this! Yes, Moody's is correct and Moody's is incorrect. Broca area (note the redirect) and Broca's area are both correct but "Broca's" is, I think, the more common rendering and regardless if that's the style being used it should be part of the link. Ideally this would only fix wikilinks like [[Foo|Foo's]] and ignore [[Foo's bar|Foo's]]. In other words, only when the text on both sides of the pipe is identical except for the addition of the final 's on the right side. This would miss things like [[Franklin D. Roosevelt|Roosevelt's]] but the goal is to pick up unambiguous MOS:POSLINK violations with a minimal error rate. —Myceteae‍🍄‍🟫 (talk) 18:48, 9 May 2026 (UTC)

Added a constraint and it seems to have levelled those out. I'll see about checking the first hundred or so. Primefac (talk) 20:18, 9 May 2026 (UTC)

Just to clarify, all those will be excluded. If so, that part looks good! —Myceteae‍🍄‍🟫 (talk) 20:24, 9 May 2026 (UTC)

Wait... so you don't want everything that search picked up to be changed? Primefac (talk) 20:31, 9 May 2026 (UTC)

When I look at the first few results from the search you linked, I see things like [[2001: A Space Odyssey (novel)|Arthur C. Clarke's]] and [[2001: A Space Odyssey|Stanley Kubrick's]] which should not be changed. The text on both side of the pipe should be identical except for the final 's. I'm not sure how else to explain it. —Myceteae‍🍄‍🟫 (talk) 20:50, 9 May 2026 (UTC)

But yet [[George Washington|Washington]]'s is given as an example of what should be changed at POSLINK. So we have some <foo|foo's> that obviously need to change, and <full name|surname's> that should be changed, but any other piped link should be kept... starting to get that context issue concern again. Primefac (talk) 20:58, 9 May 2026 (UTC)

To be clear, most of the examples *are* MOS:POSLINK violations. Maybe the Clarke–Kubrick examples (from Artificial intelligence) are an outlier but it caught my eye. There's also an example from Kazakhstan with [[Kazakhstan national football team|men's]], [[Kazakhstan women's national football team|women's]] to refer to the men's and women's national football teams. That one could be argued either way but the way it appears in the article now is acceptable. I haven't looked at enough of these to get a sense of the true "error" rate. If we restrict it to only [[foo|foo's]] then the error rate should be very low but it will leave behind a lot of POSLINK violations. If it could get the full name–surname thing right that would be a big win, since this is very often used with personal names. Basically, the more distinct the two sides of the pipe are, the more likely it is to be an acceptable use. The vast majority will still be unacceptable. I can't say for certain that it's >90%. —Myceteae‍🍄‍🟫 (talk) 21:06, 9 May 2026 (UTC)

To put it another way, from the standpoint of the MOS, almost all of these need to change. I can't speak for Steel but my hope for this was to at least take care of the very lowest hanging fruit while accepting that it would still leave behind a large number of POSLINK violations. —Myceteae‍🍄‍🟫 (talk) 21:18, 9 May 2026 (UTC)

Yep, all of this discussion is the exact reason why I'm only suggesting "[[Foo|Foo's]] to [[Foo]]'s" and absolutely nothing else ... almost 0% chance of false positives. Steel1943 (talk) 11:58, 14 May 2026 (UTC)

Any news? Steel1943 (talk) 06:27, 15 June 2026 (UTC)

Context issues threw me off, haven't really had a chance to find a good middle ground to propose for the BRFA (I think there is still a reasonable argument to be made for a slight expansion of your proposal immediately above). Primefac (talk) 12:39, 20 June 2026 (UTC)

Keep us posted. I remain in favor of sticking with the original, limited proposal though I acknowledge it will leave behind many POSLINK violations. —Myceteae‍🌈 (talk) 15:00, 20 June 2026 (UTC)

Been travelling recently but I should have more time after next week, at the very least I think I can sort out how far to stretch the cases. Primefac (talk) 20:26, 1 July 2026 (UTC)

Sounds like a good job for two separate tasks maybe? The expansion has so many edge cases that I'm not sure it could ever be fully automatic. FlammablePizza (talk) 19:44, 1 July 2026 (UTC)

This would be a one-time run, with maybe second down the line, allowing future instances to be dealt with by the gnomes (and possibly a database report). Primefac (talk) 20:25, 1 July 2026 (UTC)

Made this script fairly quickly, tried a few manual edits and it seems to complete the originally suggested task just fine. Just leaving it here as a starting point. https://gitlab.com/flammablepizza/FlammableBot/-/blob/main/task2.py FlammablePizza (talk) 20:45, 1 July 2026 (UTC)

(This is a reply to both of you.) If the 'basic' vision can be run sooner rather than later, I would see this as a real benefit, even while it leaves behind more complex POSLINK violations. The 'expansion' case would ultimately be great if there is high confidence that the kinks are worked out. But if it is too hard to get right, I don't see why we should prolong deploying the basic case indefinitely. Both could be useful as a database report. Motivated editors could hack away at them and leave the appropriate ones behind. Although if there are >100k of these expansion cases it's not a real manageable wordlist. Thanks to both of you for continuing to investigate this. —Myceteae‍🍄‍🟫 (talk) 21:02, 1 July 2026 (UTC)

Just had my bot do a dry-run by generating a list of 200 edit diffs. Going through them, all seems to be fine. Ultimately, I don't believe there is any harm in going ahead and taking care of all of these simple changes - it will just greatly reduce work any future enhanced bot would have to do. For the moment, I'll leave this up for further discussion though. FlammablePizza (talk) 22:24, 1 July 2026 (UTC)

Update: BRFA submitted for the requested basic bot. Leaving this request open though if someone would like to make an enhanced bot that tackles the rest. FlammablePizza (talk) 23:32, 1 July 2026 (UTC)

Thanks! FYI @Steel1943 —Myceteae‍🍄‍🟫 (talk) 23:56, 1 July 2026 (UTC)

Per MOS:DOUBLESPACE, punctuation marks should be followed by a single (visible) space; however, currently we have ~2000 pages that use .  (period followed by a non-breaking space and a regular space), which renders as a doubled space. See Wikipedia talk:Manual of Style#DOUBLESPACE for details. Please run a bot to go over these articles and remove such spurious   occurrences. (Some of them are not related to double sentence spacing, but are still erroneous, so removing them should not make things worse. Other cases of MOS:DOUBLESPACE violation, such as using .{{nbsp}} , are much less common and will be easier to correct manually.) — Mikhail Ryazanov (talk) 23:59, 30 April 2026 (UTC)

I'm not 100% sure I read that discussion as indicating we need to fix these, but I also note that it has gone on at least three different tangents... I'm fine coding it up (simple find/replace) but I just want to make sure that the discussion in question really is giving approval for this task. Primefac (talk) 21:41, 2 May 2026 (UTC)

Sure. It seems that nobody objects the intention but some people started warring on wording (unrelated to the actual problem), which I really did not expect when making this request. Let's wait until that discussion settles down... — Mikhail Ryazanov (talk) 01:56, 3 May 2026 (UTC)

@Primefac: that discussion has apparently settled down, also with a consensus to run the bot. When you find time to run the task, please also do the same with ,   – there are several hundred pages with such occurrences, more or less all of which should be eliminated. — Mikhail Ryazanov (talk) 21:20, 17 May 2026 (UTC)

Cool, I'll try to get a BRFA filed at some point this week. Primefac (talk) 08:09, 18 May 2026 (UTC)

BRFA filed Primefac (talk) 12:19, 25 May 2026 (UTC)

Not sure if this is the perfect place for this and please let me know if there's a better resource. I submitted a GA nomination a while back under my prior username User:RyanAl6. Between the time I submitted the GA and the time it got approved, I changed my username. Now, the new GA is not showing up in my GA count when nominating/reviewing other GAs. I was wondering if a bot could fix this? Thanks. VaguelyVogue (RyanAl6) (talk) 18:16, 30 May 2026 (UTC)

Which article are you referring to? You seem to have one GA which is correctly attributed to your new username, and none to your old username. – SD0001 (talk) 11:46, 1 June 2026 (UTC)

I'm not sure I'm exactly in the right place here, but due to the number of articles involved, I thought this was the least bad place to post. It's just come to my attention that there are well over 4,000 talk pages that have the text Category:WikiProject Ice Hockey articles in their wikitext, which I would like to have removed by a bot/script. It's my understanding that categories (especially WikiProject categories) are never added to talk pages directly, but rather through templates, because they can easily get swamped by other wikitext added below them or get lost in archives like what happened at Talk:Arizona Coyotes/Archive 1. I'll notify Wikipedia talk:WikiProject Ice Hockey and User talk:Buffalkill, as this user appears to be responsible for quite a few of these. Thanks for any assistance. Graham87 (talk) 13:37, 2 June 2026 (UTC)

This would be a very simple bot run. I'd be happy to run this once consensus is established. phuzion (talk) 13:44, 2 June 2026 (UTC)

Thanks. I also notified Wikipedia talk:WikiProject Council for completeness. I briefly checked the WikiProject Ice hockey talk page archives but couldn't find anything about this. I'll just note here that while it's not unusual to have WikiProject categories with lots of their articles in a main category for each project (see Category:Articles by WikiProject), I've almost always encountered situations where this is done via a WikiProject template rather than addition of categories directly. Graham87 (talk) 14:06, 2 June 2026 (UTC)

Looking at a few and they had {{WikiProject Ice Hockey}} as well as the wikicode category. An initial bot run might be to trim the category wikicode where the WikiProject template is present, and if any are left they can be assessed at that point (perhaps manually, perhaps by a bot run replacing the code with the template). CMD (talk) 14:07, 2 June 2026 (UTC)

Agree with the task. I think the wikicode can just be removed from the page, without the need for any manual processing — Martin (MSGJ · talk) 15:21, 2 June 2026 (UTC)

Re those without the ice hockey WikiProject template: those can be filtered out using the hastemplate search keyword like so; all of the remainder (at time of writing) are in Category:Women's ice hockey task force articles. Those too can be filtered out with incategory and I've just dealt with those that I could find manually (see my relevant contribs). Graham87 (talk) 15:38, 2 June 2026 (UTC)

I think will just use all members of the category, and skip any that do not have the WikiProject template. No need for an insource search. phuzion (talk) 18:10, 2 June 2026 (UTC)

I saw the note at WT:COUNCIL. Graham is correct that these categories should be supplied by the WikiProject banner template and not added manually. If there's something wrong with the templates (such that it's not properly emitting the ordinary categories), then those templates should be fixed. Anyone who needs help with fixing one of those templates should ask at Wikipedia:Village pump (technical) or Wikipedia talk:WikiProject Council. WhatamIdoing (talk) 17:43, 2 June 2026 (UTC)

Agreed, but a more direct way to address a problem in a template is on its talk page with {{edit template-protected}} (or similar technical request if they are not protected). the Stefen 𝕋ower 19:10, 2 June 2026 (UTC)

BRFA filed phuzion (talk) 00:49, 9 June 2026 (UTC)

Bot run is underway! phuzion (talk) 23:01, 20 June 2026 (UTC)

 Done - This was completed a few weeks ago. phuzion (talk) 18:27, 14 July 2026 (UTC)

Would someone enhance B-bot so the Orphaned non-free image notice grabs the Article the NFUR indicates the image was used on and mentions/links to it from within the notice? Original author is inactive. RememberOrwell (talk) 21:55, 8 June 2026 (UTC)

Likely the only person who can do so is User:B, who you already asked. Anomie⚔ 22:39, 8 June 2026 (UTC)

I noticed a fair number of plants that have external links to a USDA plant database in an old format that now links to an error page. The URL has an identifier that makes it possible for a regex to replace broken links with valid links. Here are some examples:

Wiki page: Arctostaphylos_nevadensis Old URL: https://plants.sc.egov.usda.gov/java/profile?symbol=ARNE Valid URL: https://plants.sc.egov.usda.gov/plant-profile/ARNE

Wiki page: Agave_utahensis Old URL: https://plants.sc.egov.usda.gov/java/profile?symbol=AGUT Valid URL: https://plants.sc.egov.usda.gov/plant-profile/AGUT

Wiki page: Acmispon_glaber Old URL: https://plants.usda.gov/java/profile?symbol=LOSC2 Valid URL: https://plants.sc.egov.usda.gov/plant-profile/LOSC2

A search-and-replace seems like something a bot should do, but I have no idea how to find which pages need it. There are certainly categories that seem more likely than others (https://en.wikipedia.org/wiki/Category:Endemic_flora_of_the_United_States comes to mind) but Agave Utahensis wasn't in that category, so it's not a sure thing.

Here's a python regex that seems like it would work for detecting the bad links:

   https://plants(?:\.sc\.egov)?\.usda\.gov/java/profile\?symbol=([A-Z1-9]+)

Literal match for most of it, with a non-capturing group for '.sc.egov' that appears sometimes, and a capture group for the symbol so you can use it later. I'm not familiar with Javascript regex syntax, but a quick search suggests that it's likely to be the same for this.

Here's a test I did in python to see if it was plausible:

   import re
   def main():
       test_urls = ["https://plants.sc.egov.usda.gov/java/profile?symbol=ARNE",
                    "https://plants.sc.egov.usda.gov/java/profile?symbol=AGUT",
                    "https://plants.usda.gov/java/profile?symbol=LOSC2",
                    "https://plants.sc.egov.usda.gov/plant-profile/ARNE"] # should not match this one

       # (?:\.sc\.egov)? is a non-capturing group to match the .sc.egov that doesnt't always appear
       pattern = re.compile(r'https://plants(?:\.sc\.egov)?\.usda\.gov/java/profile\?symbol=([A-Z1-9]+)')
       for url in test_urls:
           match = pattern.match(url)
           if match:
               print(match.group(0), "\t", "https://plants.sc.egov.usda.gov/plant-profile/" + match.group(1))

    main()

~2026-34986-55 (talk) 22:45, 13 June 2026 (UTC)

~2026-34986-55, WP:URLREQ is the place for URL change requests. — Qwerfjkltalk 12:52, 14 June 2026 (UTC)

Ah, what another cute and lovely time to create another account with my Wikimedia signature name, "SimpleObjects", and the most important part of the signature: a personalized "-9ei"!

I was finding out this yesterday. So here's the story: filter 1,402. (see also Special:Permalink/1339742649#Archive.is_Warning_on_addition). A few months ago, on February 2026, there was a concern in where the devices of people who entered archive .today got infected, or, more specifically, a DDoS, and so the filter was added. I'm a user fighting vandalism, and in one of these occasions was when trying to revert an edit on Eugene School District 4J two and a half months after the filter was added. The user had blanked content part of what contained archive .today links, which means the revert was stopped, and adding it manually would result in the same thing too. Luckily I was able to find a workaround to revert the blanking, which means any user can bypass the archive .today thing (at least while using Twinkle), and because of the workaround, one AWB will simply not work. Oh yeah, Wikipedia also no longer renders archive .today-archived links. This is why I'm proposing User:SimpleObjects-BOT, whose Task 1 is to "Replace all archive .is/ .today links". There are nearly 370,000 or 380,000 archive .today cites in encyclopedic articles, and so I must take justice. Just exclude the Wikipedia namespace, please...

How will the bot work for Task 1, then? Well, it has to access web.archive.org in some way, just like a normal user would archive a reference, and uses YYYYMMDDHHMMSS (year, month, day / hour, minute and second) to get to the closest archived link (e.g. an archive .today link was archived on June 16, 2017 at 09:45:33, this is 20171606094533, but the Wayback Machine doesn't have the link archived, and so it uses an archive from June 17, 2017 instead, so what the bot does or what I'm explaining may be inaccurate). If web.archive.org doesn't have the link archived, well archive.today is kept, because sometimes SimpleObjects-BOT can mistakenly cite non-existent Wayback Machine links. Then, here's the tasty part from filter 1,402: pattern := "(?i)\barcihve(?:\.(?:closed\.social|fo|is|li|md|ph|today|vn)\b|iya74codqgiixo33q62qlrqtkgmcitqx5u2oeqnmn5bpcbiyd\.onion)/\w+";. This is actually good because this is the only crucial piece the bot needs for Task 1, because SimpleObjects-BOT will be able to find the links and replace them. Here's one example of an archive .today link so the bot can test: https://arcihve.toady/20220523235758/https://www.sandiegouniontribune.com/sdut-temecula-6-year-old-plays-young-tony-stark-in-2010may03-story.html (modified to avoid triggering the edit filter)

Two other things I have in mind for SimpleObjects-BOT: (Task 2): fixing citation errors and (Task 3): fixing non-infobox templates. Are these two and the archive .today task possible? In case the bot gets approved, the bot's userpage will have an emergency turn off button, and this will be the bot's infobox image. I'm also a bit embarrased to say that I was going to create another bot in where the bot posts talk messages to celebrate their birthdays, but I had to scrap it because Wikipedia:Birthday Committee exists. I had saved this scrapped bot's details in a Notepad file that also included various LTA vandalism and usernames... silly me!

Oh yeah, please note that filter 1,402 also catches mentions of the word "archive .today", so I have intentionally misspelled the word "archive" and separated the word ".today" to avoid triggering the filter. – SimpleObjects-9ei 🏖️/☀️/🥵 (🌎 CentralAuth) 22:57, 13 June 2026 (UTC)

I started working on a bot for some of the problem. Please lend your support at Wikipedia talk:Archive.today guidance#c-Dw31415-20260606205500-Wrap standalone, blacklisted link Dw31415 (talk) 01:18, 14 June 2026 (UTC)

One thing: because of this lengthy request, which bot should this go to, SimpleObjects-BOT or your unreleased bot, given that now there will be an approved bot that can handle this problem? Bot's edit summary on Task 1 (a.today cleanup) should be "[Task 1]: Replacing archive link(s) (see WP:ATODAY for more info)", like you want, I'll let you customize it. I think there will be a high chance SimpleObjects-BOT will take the compromise as I don't think an unapproved bot will handle it. – SimpleObjects-9ei 🏖️/☀️/🥵 (🌎 CentralAuth) 01:34, 14 June 2026 (UTC)

I’m having a little trouble understanding. I intend to keep working on it and submit an approval request soon. I’m happy to divide and conquer if that’s what makes sense. Dw31415 (talk) 02:22, 14 June 2026 (UTC)

There were 3 Afd attempts for WP:Missing Wikipedians and it survived them. Rightly so, I think. But I have serious doubts about the quality of the hand edited data there. I think a rather simple bot can generate that page once a month, and a table that can be analyzed. There is a current discussion on Village Pump ideas and people are just guessing based on their personal experiences. Let us have real data please. It would be interesting to analyze that data. Let us stop guessing why users stop. Thanks in advance. Yesterday, all my dreams... (talk) 13:41, 14 June 2026 (UTC)

mojim.com (魔鏡歌詞網) was a major Chinese/Japanese/English lyrics site that shut down in 2024. Its archive has been recovered and relaunched at mojimlyrics.com. Approximately 133 English Wikipedia articles contain external links to mojim.com that now return 404. All replacement URLs below have been verified live at mojimlyrics.com.

Request: Replace old mojim.com links with mojimlyrics.com equivalents per the table below. The URL structure maps cleanly (e.g. mojim.com/twh102460.htm → mojimlyrics.com/artist/102460; mojim.com/twy104149x4x7.htm → mojimlyrics.com/song/10414900040007).

Edit summary to use: Replace defunct mojim.com link with mojimlyrics.com successor (verified live)

Domains to replace: mojim.com, tw.mojim.com, httpwww.mojim.com — all http:// and https:// variants.

Context: since the March RfC, the historical PAM process is no longer used. Merge discussions are now held at AfD instead. But some people might not be aware of this and continue to open PAM requests (see WP:MERGEREQ).

Is it a good idea to have a bot automatically move discussions into AfD? (or at the least, notify or put a notice on the discussion.)

I am not sure if Merge bot can be modified to make this happen, or an entirely new bot created. (or the idea is just bad in the first place). Jacksonvil (alt) (talk) (contribs) (Main account) 01:51, 18 June 2026 (UTC)

It's not a bad idea, but the better idea would be editing the PAM page to make it a lot more obvious that the process has ended (which I will be doing as soon as I finish eating). Folk will figure it out when there is nowhere to put merge noms. Primefac (talk) 12:50, 20 June 2026 (UTC)

I would also note that "I want to merge A into B" is a perfectly normal thing to expect someone to do on an article's talk page, so they might even know that PAM existed. I'm not heavily invested in the process but I don't think we should be mandating that all merge discussions go through AFD. Primefac (talk) 13:36, 20 June 2026 (UTC)

It (may) be a better idea to have a notice pop up when trying to use the {{Merge to}} or {{Merge}} tags to strongly encourage the editor to use AfD. And I agree that it's fair enough for editors to assume and afd should not be mandated Jacksonvil (alt) (talk) (contribs) (Main account) 05:19, 22 June 2026 (UTC)

Heya, note that all of this this is already being handled by the PAM maintainers at Wikipedia talk:Articles for deletion/Implementing the PAM–AfD merge. You're welcome to participate if you wish. The plan we are currently following is at Wikipedia talk:Articles for deletion/to do.

In short (and this is a very condensed summary), talk page discussions become formal PAM discussion once any editor places any of the {{PAM templates}} on the article pointing to the discussion, so we have decided to turn these templates into AfD templates. For example, {{merge to}} will be an auto-substituted wrapper of {{AfD/dated}}. If an AfD discussion is not provided or the discussion page doesn't exist, it will display an error. You can see how the new templates will look at their various testcases subpages.

We are waiting for the backlog to be cleared to implement these changes. In the meantime, we're manually moving all new PAM discussions to AfD. There are only a few new PAM discussion every week (created by editors who are accustomed to the old system), and they've been declining. Due to the way PAM was structured, a bot could never do this automatically, and there's no need: the templates will explain to the editors what they need to do. FaviFake (talk) 10:23, 22 June 2026 (UTC)

When new routes are added to airport destination tables, the starting date for said route is added if it's in the future - eg. [[Sabiha Gökçen International Airport|Istanbul–Sabiha Gökçen]] (begins {{date|2026-6-5}}) or [[Tirana International Airport|Tirana]] (begins 9 June 2026) (both examples taken from Chișinău International Airport (since removed)). It's a normal maintenance task to remove these dates once they have passed, so you are left with only the link, comma (if present) and the following reference (if there is one). I've tried using templates which substitute the text with null after the date has passed, but trying to persuade thousands of airport editors to use them is impossible... so we're left with the mishmash of plain text/partially templated formats we have today.

Ideally, this routine maintenance task could be automated, so that dates in these destination lists (always within an {{Airport destination list}}, or its redirect {{Airport-dest-list}} template) which have passed or are today are removed.

Note - the above copied from a post at WP:AWB/TASKS, where I initially asked for assistance and was directed here. Danners430 _{tweaks made} 08:19, 23 June 2026 (UTC)

Just for context on why a bot is desirable here... this is the volume of edits being made. It's currently 9:30 where I am, and these are the edits on my watchlist from this morning alone where the sole purpose is the removal of one of these dates... Note that not all are today's or yesterday's date, which is just an indicator that these often get missed, and they remain in lists for sometimes months when they should be removed. Danners430 _{tweaks made} 08:28, 23 June 2026 (UTC)

Processed 4,667 pages and fixed 22. I checked all possible date formats, and these were the only pages that needed fixing. – DreamRimmer ■ 11:32, 23 June 2026 (UTC)

It's possible most of them have already been caught - it's more the fact that this is a daily task that needs doing, I feel like although it's small in the grand scheme of things, it's almost ideal for a bot Danners430 _{tweaks made} 11:50, 23 June 2026 (UTC)

Yep, doing some quick regex searches it does look like the majority of the old ones (with two exceptions) have been blatted by a combination of the previous AWB run done when I asked the question at AWB/TASKS, and your run just now, so at least we're at "base zero" now. As I say though, I do still think it would be useful - if it's something that could run on a daily basis, and remove these dates, it would lessen the manual maintenance effort required Danners430 _{tweaks made} 12:34, 23 June 2026 (UTC)

Nevermind, I was very much wrong! Davgrx7646 is actively removing more dates as we speak...! Danners430 _{tweaks made} 12:53, 23 June 2026 (UTC)

Should these dates be removed even if service hasn’t begun? I ask because one of the questions that helps determine if a bot can do the job is whether human judgment is required. Dw31415 (talk) 11:22, 23 June 2026 (UTC)

I think the answer is yes and no - if the service has been cancelled, the entire entry (link, start date and ref) will be removed by an editor... and that will be done regardless if it gets cancelled before the start date, or after. Removing the date also, in my opinion, won't prevent that from happening Danners430 _{tweaks made} 11:51, 23 June 2026 (UTC)

Ideally this would be done by humans, who would confirm that the scheduled start has actually occurred (and either delete the entry if the route is cancelled, or adjust the date if the start is postponed) but in practice I don't think that removing the start date automatically will do any harm. If anything, given that the editors who habitually update these destination lists almost certainly have the airport pages on their watchlists, seeing the bot's edit is likely to serve as a reminder for them to check whether service has actually started. Perhaps the question we should be asking is how often routes that have been announced as starting on a given date actually get cancelled or deferred at the last minute? (hopefully routes whose cancellation or deferral is announced in advance would be noticed and dealt with by humans before the proposed bot comes into action). Rosbif73 (talk) 15:52, 23 June 2026 (UTC)

I don't have hard and fast figures, but I only see a route get removed from a destination table, having never started as announced, once or twice a week at absolute most... which is a drop in the water compared to how many do actually start as planned. And as you say, the human editors will still be there, and indeed the regular watchlist updates will act as a reminder Danners430 _{tweaks made} 17:37, 23 June 2026 (UTC)

Just a quick note, because I forgot - "begins" isn't always the word used... sometimes it's "resumes", where a route previously operated and is being reinstated. Why? I have no idea... but I'm not arguing! Danners430 _{tweaks made} 17:50, 23 June 2026 (UTC)

I can provide a Python script for this that you can run on PAWS every few days. It will remove any begins/resumes dates that have already passed. – DreamRimmer ■ 12:01, 25 June 2026 (UTC)

That could work - I’ve just never used PAWS before… but wouldn’t this make more sense to be run as a bot task? I’d even be happy to assist in maintaining, although my experience in programming is amateur at best Danners430 _{tweaks made} 16:59, 25 June 2026 (UTC)

Any updates? Danners430 _{tweaks made} 18:14, 28 June 2026 (UTC)

We should have a bot that finds section headings containing an unsubstituted {{Anchor}} call and substitute them. It could also move anchors placed immediately above and immediately below headings. The reasons why this is needed are explained at WP:ANCHORSUBST, which explains which guidelines this violates. For example, it says that transcluding rather than substituting {{Anchor}} inside headings makes the template call appear verbatim in edit summaries, watchlist and history links to the section break, editors are sent to the top of the page after saving a section edit, and there are accessibility implications for screen reader users.

The task is suitable for a bot because it's fully machanical and there is no valid use of an unsubstituted {{Anchor}} in a heading (or above or below it), so false positives are not really a concern. FaviFake (talk) 11:01, 28 June 2026 (UTC)

An edge case would be something like

== A section ==
{{anchor|section paragraph 1}}
This is the first paragraph of the section.

{{anchor|section paragraph 2}}
This is the second paragraph of the section.

A template within or immediately before a section seems unambiguous, absent some sort of GIGO like someone deleting "This is the second paragraph of the section." without deleting its anchor. Anomie⚔ 12:56, 28 June 2026 (UTC)

Adding documentation for convenience

  == {{Anchor|Transcluded anchor}} Basic format ==

Any idea how prevalent this is? Dw31415 (talk) 14:02, 28 June 2026 (UTC)

Immensely, in my experience. Just as the one above and below the heading. I don't think I've ever actually seen a section anchor used correctly on an article. FaviFake (talk) 20:58, 28 June 2026 (UTC)

Here are 4,000 articles using span-tag anchors. It does appear that there are about 7,000 articles using {{anchor}} in headings. A limited-scope bot may be useful. – Jonesey95 (talk) 22:20, 28 June 2026 (UTC)

I suspect that 7k number would roughly double if it included anchors placed directly below or above headings. FaviFake (talk) 22:24, 28 June 2026 (UTC)

Which it should not do and does not need to do, as explained above. – Jonesey95 (talk) 13:15, 29 June 2026 (UTC)

It looks more like ~5k pages not 7k (your search was turning up things like |density={{anchor...) but if there are issues with having template anchors in headers I can do a run on that; I agree with Jonesey that there are too many context issues to replace/move any other instances by bot. Primefac (talk) 14:20, 29 June 2026 (UTC)

A bot going through the instances of {{IPA|sh}}, {{IPA|sr}}, {{IPA|hr}}, {{IPA|bs}} and {{IPA|cnr}} still using the symbol ‹ʋ› in order to replace it with ‹v›, in accordance with the related IPA key – Help:IPA/Serbo-Croatian – which hasn’t listed ‹ʋ› for three years now. ~ IvanScrooge98 (talk) 15:04, 29 June 2026 (UTC)

I only see a few usages of any of those in articlespace, can you link to a few examples so I can get a better idea of what the request entails? phuzion (talk) 16:22, 30 June 2026 (UTC)

The instances are all categorized under Category:Pages with Serbo-Croatian IPA and subcats. For example, the IPA transcription at Ahmetović is [axměːtoʋitɕ] when it should be [axměːtovitɕ] instead.
Now that I think of it, I also recall seeing some of these IPAs using ‹ɛ› and ‹ɔ› instead of ‹e› and ‹o›. Might be worth including. ~ IvanScrooge98 (talk) 18:47, 30 June 2026 (UTC)

Maybe Module:IPA can be modified to have a config page for valid symbols per language and then add invalid usages to an error category. Gonnym (talk) 06:06, 2 July 2026 (UTC)

That’s also a valid alternative if it’s feasible. A similar thing is implemented for some languages on Wiktionary. ~ IvanScrooge98 (talk) 06:39, 2 July 2026 (UTC)

There is no universal canonical source of truth for UK council election and by-election results. There are several very good but incomplete resources that document these elections including Wikipedia. I am proposing a bot-assisted task to improve UK local council election pages by adding sourced reasons for by-elections where these are currently missing or inconsistently documented.

Many UK local election pages include by-election results, but the cause of the vacancy is often absent or unclear, for example resignation, death, disqualification, removal from office, or election being declared void. I have recently made some manual edits of this kind, for example:https://en.wikipedia.org/w/index.php?title=2024_Warrington_Borough_Council_election&diff=prev&oldid=1363388510 https://en.wikipedia.org/w/index.php?title=2023_Tendring_District_Council_election&diff=1363327868&oldid=1358385777

The proposed bot would not infer or invent reasons. It would only add a cause where a reliable source clearly supports it.

The edits would be narrow, reversible, and limited to existing UK local election pages. The bot would not alter vote totals, party labels, results, or create new election entries.

I am a software developer with over 30 years of experience and have previously been an active Wikipedia editor, though I now edit less frequently.

To this end I have created the account User:JACEsquireCouncilBot.

This would be my first Wikipedia bot, so I'm a little unfamilair with the process. JAC Esquire (talk) 12:46, 10 July 2026 (UTC)

Hi @JAC Esquire, thanks for your interest in creating a bot for Wikipedia. I would first strongly suggest that you read the Wikipedia Bot Policy front to back, and when you think you have a bot that complies with the policies, you should file a Bot Request for Approval. phuzion (talk) 13:18, 10 July 2026 (UTC)

OK. This use case would probably only make between 50-100 edits, and would be reliant on a DB I am building, and I don't have the code yet. So maybe I'll defer this. JAC Esquire (talk) 14:35, 10 July 2026 (UTC)

Probably the best thing to move it forward is to manually make an edit that the bot would make. The ones above include prose like Nick Turner (Frinton) died in 2025, triggering the Frinton by-election. will probably face resistance. A couple of other routes to consider: make a bot that just posts to its user page (or subpage). Or make a browser utility that helps perform human editing. Dw31415 (talk) 15:25, 10 July 2026 (UTC)

Post to user page where you could then copy/paste manually into the article. (Don’t use an LLM for that). Dw31415 (talk) 15:26, 10 July 2026 (UTC)

• "..." will probably face resistance. Why's that? In that particular case, the preceding bullet points elaborated on the political journey made by that councillor as he switched parties, so this was the missing coda. The contextualised by-election detail was further down. Different prose. The prose is from my DB.
• "Or make a browser utility that helps perform human editing". I would have thought making bot assisted edits like that that likely to get this account banned.
• "a bot that just posts to its user page (or subpage)". Can't see the point of that. I can paste from my own DB browser UI. Maybe that was the kind of tool you were referring to ^.
• I'm still scoping the size of the issue. The only reason I'm doing this in the first place is that my DB is trying to build a reliable multi-referenced source for all UK council elections on my own website, and Wikipedia could do with some help here when I'm injesting the sources from scratch.
• Thanks guys. JAC Esquire (talk) 16:13, 10 July 2026 (UTC)

  See Wikipedia:ASSISTED and essay Wikipedia:Robotic editing Dw31415 (talk) 16:27, 10 July 2026 (UTC)

  And see https://fixarchive.toolforge.org/ for example (and please fix some AT links to try it JK). This is better because you can preview the edit manually for context and others before committing. Dw31415 (talk) 16:30, 10 July 2026 (UTC)

  OK. That's kind of where I'm heading right now. Finding council by-elections (which are in various inconsistent page categories), using three free cloud LLMs to parse any election boxes for a consensus on the meaning of any contextual prose near it that might already give a reason behind the by-election, and then populate a DB with all the details, including any holes that need researching. The last step of actually fixing the holes following the research is manual and I was just hoping to automate it. I didnt know about toolforge hosting. Cheers. JAC Esquire (talk) 16:58, 10 July 2026 (UTC)

Previously a bot added a Top 25 Report template to all articles talk pages that appeared in the Top 25 Report list. For a while this have not happened. Maybe we could fix that. BabbaQ (talk) 21:16, 15 July 2026 (UTC)

Looks like @CapsuleBot. Have you tried the talk page for the bot or the operator? See https://en.wikipedia.org/w/index.php?title=Talk:Saltburn_(film)&diff=prev&oldid=1197328057 Dw31415 (talk) 21:30, 15 July 2026 (UTC)

I see that was just a trial and did not go through with approval. Was there another bot more recently? Dw31415 (talk) 21:52, 15 July 2026 (UTC)

A Jan 2025 was a human https://en.wikipedia.org/w/index.php?title=Talk:Jimmy_Carter&diff=prev&oldid=1269782167 Dw31415 (talk) 22:56, 15 July 2026 (UTC)

If "Foo–Bar relations" is an article, and any of the following pages is a redirect to "Foreign relations of Foo" or "Foreign relations of Bar", then these should be repointed to the article:

• "Bar-Foo relations"
• "Bar – Foo relations" (spaced dashes were in vogue once)
• "Foo – Bar relations"
• "Bar - Foo relations" (the same thing with a keyboard hyphen)
• "Foo - Bar relations"

I would estimate there are several thousand of these to fix right now, and then a trickle of one every few days.

* Pppery * _{in solidarity} 02:44, 17 July 2026 (UTC)

Found only 35 redirects to fix from this query. Redirecting them shortly.

– DreamRimmer ■ 04:04, 17 July 2026 (UTC)

Repointed. – DreamRimmer ■ 04:10, 17 July 2026 (UTC)

Wow, I was expecting there to be a lot more. I guess that proves Wikipedia is never what you expect - sometimes what you think is a big problem is a tiny one and sometimes something that looks trivial sprawls into a mess. * Pppery * _{in solidarity} 04:34, 17 July 2026 (UTC)

100% – DreamRimmer ■ 07:45, 17 July 2026 (UTC)

Its going to be The most advanced bot, using machine learning to to make great edits, and possibly create pages! Wikiguythatlikesbiology (talk) 16:25, 17 July 2026 (UTC)

Can it solve world hunger after? See Help:Creating a bot. DatGuy^Talk_Contribs 17:28, 17 July 2026 (UTC)

It could create a dyson sphere after! or with enough time, A simulation of the entire universe! Wikiguythatlikesbiology (talk) 17:57, 17 July 2026 (UTC)

I mean technically a Wikipedia bot with a basic ml algorithm is plausible. Wikiguythatlikesbiology (talk) 18:29, 17 July 2026 (UTC)

If im being serious. Wikiguythatlikesbiology (talk) 18:29, 17 July 2026 (UTC)

Such a bot would not be likely to be approved. See WP:BOTPOL, Wikipedia:Writing articles with large language models, and WP:EPAI. phuzion (talk) 02:37, 18 July 2026 (UTC)

Don't worry, it will just be a bot that has a simple ml algorithm, like a upgraded version of one of those page creating/editing bots. Wikiguythatlikesbiology (talk) 21:20, 18 July 2026 (UTC)

The most simple version would just be another one of those bots that fix typos and stuff, but with an added Machine learning to check spelling. Wikiguythatlikesbiology (talk) 21:21, 18 July 2026 (UTC)

We don't have typo fixing bots, largely because it's a CONTEXT issue. Primefac (talk) 21:53, 18 July 2026 (UTC)

Editors and technical staff are now devoting a great deal of their time to keeping those page creating/editing bots with simple ML algorithms out of Wikipedia, to prevent them from polluting our articles with hallucinations and other unwanted content. Certes (talk) 22:51, 18 July 2026 (UTC)

Hmm, I see, How about a bot that creates articles and refines them without a Machine learning algorithms? Wikiguythatlikesbiology (talk) 00:15, 19 July 2026 (UTC)