[00:18:58] [1/12] Hi! I’m the administrator of a Miraheze wiki, and I’d like to ask for some help with API access. [00:18:58] [2/12] Wiki: https://valsalem.miraheze.org/ [00:18:59] [3/12] I’m trying to set up an automated editing workflow using GitHub Actions and the MediaWiki API (/w/api.php), authenticated with a Bot Password. [00:18:59] [4/12] However, all requests from GitHub Actions are returning HTTP 403 Forbidden, even before authentication. [00:18:59] [5/12] I checked the response headers and found: [00:18:59] [6/12] * server: cloudflare [00:19:00] [7/12] * cf-mitigated: challenge [00:19:00] [8/12] The same issue occurs when requesting the wiki’s main page or accessing the API without authentication, so it appears that Cloudflare is blocking requests from GitHub Actions. [00:19:01] [9/12] Is there an officially supported way to allow legitimate automated API requests from GitHub Actions? [00:19:01] [10/12] For example, is it possible to request an exception, use a recommended authentication method, or configure the client differently to avoid triggering Cloudflare’s bot protection? [00:19:02] [11/12] I want to make sure my automation complies with Miraheze’s policies and security requirements. [00:19:02] [12/12] Thank you for your help! [01:16:38] [1/2] I used all of the normal templates and css stuff and whatnot, why is the red border and apostrophe there? [01:16:39] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1561548599387758752/Screenshot_2026-10-18_181527.png?ex=6ad6c0f6&is=6ad56f76&hm=80002d21dc02f79324ed55f53901ad6163948696ac1c98543e04dec030cb8f56& [03:29:58] Is there any harm to dumping everything into a single big database? [03:36:54] [1/22] Title: Cloudflare blocks legitimate MediaWiki API requests from GitHub Actions (HTTP 403) [03:36:54] [2/22] Hello, [03:36:55] [3/22] I’m the administrator of a Miraheze-hosted wiki and would like to report an issue with automated API access. [03:36:55] [4/22] Affected wiki: https://valsalem.miraheze.org/ [03:36:55] [5/22] I’m setting up an automated wiki editing workflow using GitHub Actions and the MediaWiki API, with Bot Password authentication. [03:36:56] [6/22] However, requests from GitHub Actions are blocked before authentication can take place. [03:36:56] [7/22] Observed behavior: [03:36:56] [8/22] * HTTP status: 403 Forbidden [03:36:56] [9/22] * Server: cloudflare [03:36:57] [10/22] * Response header: cf-mitigated: challenge [03:36:57] [11/22] * Affected endpoints: /w/api.php, /api.php, and /wiki/メインページ [03:36:58] [12/22] * The issue also occurs with unauthenticated GET requests. [03:36:58] [13/22] * Last confirmed: October 19, 2026, 00:09 UTC [03:36:59] [14/22] This appears to be a Cloudflare challenge affecting GitHub-hosted runners rather than a MediaWiki authentication issue. [03:37:00] [15/22] Expected behavior: [03:37:00] [16/22] Legitimate automated clients should be able to access the MediaWiki API using a supported authentication method, subject to Miraheze’s rate limits and applicable policies. [03:37:01] [17/22] Questions: [03:37:01] [18/22] 1. Is automated MediaWiki API access from GitHub Actions officially supported? [03:37:02] [19/22] 2. Is there a recommended way to prevent legitimate API requests from being challenged by Cloudflare? [03:37:02] [20/22] 3. If an exception is necessary, is there an official process for requesting one? [03:37:03] [21/22] I’m happy to provide additional diagnostic information if needed. [03:37:03] [22/22] Thank you for your assistance! [03:59:38] [1/2] https://logicamirrorworlds.miraheze.org/wiki/Wilhelmina [03:59:39] [2/2] would like to ask someone to take a look at this for me. Even though the clearly works in the infobox, it seems to fail when it put inside `Template:Hangugeo` [04:25:44] [1/2] I think it just breaks [04:25:44] [2/2] you can insert the code for font color inside the template directly using a different parameter though [04:26:55] assuming you need to color the first parameter or the name, replace {{{1}}} with {{{1}}}/font [04:27:33] then when calling the template add "|color = silver" at the end [04:28:55] yeah but its not a per template kind of the thing [04:29:06] its a per instancce of text [04:29:23] Template:Hangugeo is just a template for korean language and stuff [04:29:32] i just need it to not break is what im saying [04:33:05] I didn't quite get it [04:33:37] Do you need to color every text in the template? Or just a specific one [04:35:48] [1/2] if you need to color every text, even the korean letters, that may need a different approach [04:35:48] [2/2] but if you just need to color one, adding the code bit might save you more time [04:38:40] [1/2] here's how it should work/look like when applying my proposed code [04:38:40] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1561599442702966854/image.png?ex=6ad6f050&is=6ad59ed0&hm=8f4af92a71769c2a23d24ddee2012b51ee183418a82d583a56a1be75b1574b25& [04:52:21] What are you trying to do in the template? Color some text? Or is the problem with the infobox? [05:27:18] iirc you can dm a tech member the user-agent you use. [[Tech:User Agents]] might have more info [05:27:18] [05:28:38] github actions was banned bc it runs on azure which is also used by openai's bots [06:23:53] Please email tech@miraheze.org in plain English without any AI fluff [06:24:03] And make sure you are setting a user agent as above [06:25:20] Google does use Chrome browser data to influence SEO in some indirect ways but not enough to have any meaningful impact in realty. [06:59:27] Thank you for your response. I'm Japanese, and to be honest, I'm not very good at English, so I'm asking questions with the help of an AI assistant. I'm sorry if that was a confusing expression. [07:00:28] It mainly just adds a lot of extra useless data. I would rather see broken English than AI fluff. [07:03:36] If it upsets you, I won't let our AI write it. However, since everything was conducted in Japanese, I know almost no vocabulary or grammar at all. So, I'm speaking through a Japanese translation app; is that okay? [07:05:47] You can also write in your own language if that's easier and we can translate it ourselves. [07:06:24] [1/2] ありがとうございます [07:06:24] [2/2] 手厚いサポートに感謝します(´・ω・`) [07:07:59] いえいえ! [07:09:06] [1/2] 仕事の小休憩に返答していたので、そろそろ戻らないといけないんです [07:09:06] [2/2] また終わったら来ます! [07:17:20] oh i forgot tech has a mail address [07:18:03] sometimes i saw similar tasks on phorge with no reply for days. next time when i see i'll tell them to email too [07:19:21] Phorge works as well. Feel free to mention the tasks in #tech though if you see that we haven't replied in a long time, with the backlog we could also end up missing them. [07:20:17] okay i will. i guess such tasks need security project? bc ua is usually a confidential info [07:21:29] Yes, that is probably true. I feel I should clarify though as I was not clear above I see. Email is preferred for that, but if they already make a Phorge task anyway there is no point to duplicate that to email. [07:23:07] many existing tasks aren't set private. e.g. https://issue-tracker.miraheze.org/T16325 [07:23:43] i'm afraid ai crawlers can get these and pretend with those public uas [07:24:20] Yeah that may be a valid concern. We haven't been very consistent with these types of tasks have we... [07:25:52] neither [07:26:04] Fixed [07:26:06] I am only trying to color The Silver Bullet in that particular instance and only that instance [07:26:16] let me give some context so its easier to understand [07:26:28] in this world, mercenaries of the highest rank are given a title involving a Color [07:26:34] and when this Color is listed in the wiki [07:26:42] the font-color is meant to reflect that title [07:27:17] I see, so you want to change the color of the text whenever a particular name is mentioned [07:27:33] it has nothing to do with the template itself [07:27:48] im just trying to understand why is it when i use the font-color descriptor inside of the hangugeo template, it breaks [07:28:06] You can use to set text colour, unless that's the thing that's breaking. [07:28:32] https://logicamirrorworlds.miraheze.org/wiki/Wilhelmina [07:28:35] have a look if you'd like [07:28:37] I tried and both and broke iirc [07:28:55] mmmm [07:28:56] no fun [07:29:27] Those are the only two I know, unfortunately. [08:16:52] Do you have a source for this because I doubt this is true [08:17:37] Are you saying the more people who use Chrome to visit a wiki increases SEO? Or if you as the submitter use Chrome increases SEO? Both sound absurd and I doubt are the case and I can't find anything that says anything remotely similar [08:24:02] https://www.justice.gov/atr/media/1398871/dl page 3 [08:44:31] It uses Chromes browser data to track things like CTR more for example. The more people who use Chrome does not directly affect SEO, but using Chrome indirectly affects SEO. [09:45:29] [1/2] 退勤しました [09:45:30] [2/2] 送ってもらったURLを見て、連絡先情報を入れるようにしました [09:57:21] Finally upgraded from Maps to Kartographer and it doesn't even work [10:05:42] [1/2] 教えてもらったところにもEメールを送りました [10:05:42] [2/2] 日本語ですが、許してもらえると嬉しいです:(( ꒪꒫꒪)): [10:06:51] 確認しました。全然気にしないでください。 [10:07:45] [1/3] ありがとうございます [10:07:45] [2/3] 国が違うゆえに不快にさせてしまわないか心配ですが、全てを通して僕の言葉に悪意はありません [10:07:45] [3/3] 親切な先輩たちに感謝!(ㅅ´꒳` ) [10:09:46] 全然心配しないでください。私たちも十分理解しています。今はもう離れますが、後ほど私か他のメンバーが確認します。 [10:18:57] https://cdn.discordapp.com/attachments/407537962553966603/1561685075979141231/veyr_innocent_audio.mp3?ex=6ad74010&is=6ad5ee90&hm=eef0ec13f86a1fdda18050506db6ff7afddc6d8818ad90702f1d276d45861c95& [10:58:34] ||bro did you check the snails|| [10:58:42] ? [11:12:00] This mentions chrome once in the entire document and the signal it's talking about is redacted so Im not sure that backs up the point [11:12:46] I doubt they're using Chrome data for CTR as opposed to CTR on Google itself as a whole [11:17:49] I'm sure its part of a whole, but even if minor it is a part of that [11:29:27] Does anyone have any extra experience with using interlanguage links for translated content? A primary concern we have is that this setup might pollute the article search by adding the translated articles to the results, when most of the time people won't be interested in searching explicitly for translated versions, just the article for their language. [11:30:59] Also would be affected too [11:31:52] (also we're unsure about using the translate extension due to its quirks, and creating separate wikis might be overkill) [11:32:05] That is going to happen and there isn't much you can do about it. Pages cannot be exempted from search on a per-page basis. The best you can do is probably isolating translations inside a namespace and turn search off for them. [11:34:00] I see, hmm [12:10:33] Would the magic word NOINDEX  not be useable? [12:18:24] [1/3] Since Circussearch is enabled on my wiki, the search bar has stopped showing any search result & only works when I type in the exact and full article title. I keep getting this error when I tap "Search for page containing (any letter i type) [12:18:24] [2/3] "An error has occurred while searching: We could not complete your search due to a temporary problem. Please try again later." [12:18:24] [3/3] Circussearch was enabled yesterday. [12:22:12] That one controls search engines like Google, not the mediawiki search if I understood it correctly [12:45:31] what's the wiki [14:05:05] [1/2] Ah ok, I thought it was just a search engine they were trying not to be in. [14:05:05] [2/2] The only other way I'm aware of is putting it into its own namespace that isn't searched by default but its not really a solution [14:06:09] Yeah, we were discussing it in our own server and handicapping search sucks [14:06:48] Someone has intrusive thoughts of hacking something in with gadgets [15:16:16] What do I do if I matched the abuse rule of linking external wikis on my user page on login.miraheze.org? What I'm trying to do is to link to the wikis I usually edit, but it appears that I can't due to the abuse filter. [15:24:18] the immediate problem should be solved, however internal links should also solve it by proxy and are the preferable choice [15:24:37] instead of [https://industrialist.miraheze.org/wiki/Main_Page Industrialist Wiki], you might consider [[mh:industrialist|Industrialist Wiki]] for example [15:26:21] this is trickier for the custom domain since you'd need to know that, but this can be consistently retrieved from managewiki core, so for the internal link is [[mh:projectsekai]] (database name - wiki = internal url name) [15:26:22] [15:27:21] but in any case the approach is up to you as I have punted the external link problem for 2 years by assigning confirmed and this is mostly for general/future reference [15:36:21] did you just see my revision [15:36:23] what [15:37:15] okay, thanks [15:51:40] did you decline the retro windows thing? [15:51:48] i can withdraw it [15:53:04] I believe abusefilter-blocked revisions are still visible, but in a different place? Maybe? [15:53:10] I don't know much about abusefilter, to be honest. [16:02:37] Special:AbuseLog [16:02:48] though depending on the filter, entries may not be visible without elevated access [16:04:10] [1/2] in your case this is probably advisable. I was not sure how to proceed but know that a) I don't think it would be wise to simply grant you rights from what is essentially a whim and b) in the end I would expect the wiki to end up in the same place, and I need to rethink how the wiki is going to be approached. It's not an active problem but it is a zom [16:04:10] [2/2] bie conservatorship that is unlikely to change. [16:06:41] [1/2] it was never meant to be in that indefinite status so realistically there is someone reasonably trustworthy with a relevant interest to take up intermittent checks and that is that. Possibly that's even me since I have a tangential interest in obsolete windows stuff but all I can really do is stop by once in a blue moon to put out a fire. What's weird is that it is a zombie yet it ha [16:06:41] [2/2] s a community over time that's nibbled on it, so it is a grey spot, and a problem that I've never seen urgency in 'fixing'. And everyone who's expressed interest had some sort of issue or just came and went in a short time. [16:17:49] no wonder [16:20:57] perhaps it would be friendlier to note about internal links? I might look at that [18:26:00] how do i make my own wiki appear when you search [19:10:42] Ensure it you have SEO settings set to "index, follow", then pray [19:11:42] Jokes aside, read [[Help:Improving Wiki Search Ranking with WikiSEO]] [19:11:43] [19:32:11] [1/2] man THIS much? [19:32:11] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1561824303320793159/image.png?ex=6ad7c1bb&is=6ad6703b&hm=4e2041afc63f8989ae06064fa824d3588ab7dc73f3b91f18ed415d664b922b8f& [19:36:49] You don't have to do all of them, I strongly recommend at least doing Google Search Console though [19:36:58] k [19:54:52] yo mediviewer won't work for me for some reason, anything I can do about it? [20:32:34] [1/2] how can i make links display images next to them when searching? i've have SEO things set up for a while but never seen any [20:32:34] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1561839495689666581/image.png?ex=6ad7cfe1&is=6ad67e61&hm=58054504fea2a53fa8fd02f0813f11767540fef591eb3e75f675db66faa2ccb7&