Talk:Community Wishlist/Wishes/Blocking proxy IP address

This page is for discussions related to the Community Wishlist/Wishes/Blocking proxy IP address page.   Please remember to: Sign posts using the four tildes (~~~~) Remain civil and polite during discussions. Place new text under old text (start a new post).

untitled

• FWIW, proxy blocking is still happening very frequently as can be seen in Special:Log/gblblock, benefiting more than one project. If more automation can be brought in, keeping these at the global blocking level instead of project level may be ideal. @EPIC: for some feedback as he is very active in this area. — xaosflux ^Talk 10:36, 6 August 2024 (UTC)

  I'd also be in favor of potentially automating this further in the future. How to enforce it I'm not too sure about, but I have also kept seeing many proxy block requests on my home wiki when it can, and in many cases should, be handled on the global level. EPIC (talk) 10:49, 6 August 2024 (UTC)

IP Proxy Blocking

Hey @Robertsky, @Xaosflux, and @EPIC, thanks for the wish and commentary on it. I wanted to clarify the problem and situation: When editors edit from an IP Proxy, wikis are more prone to abuse. Admins have built bots such as https://en.wikipedia.org/wiki/User:ST47ProxyBot to manage the flow of edits, however they might not be maintained, and extending AbuseFilter to support IP Proxies would help manage the flow of blocks.

To summarize the dilemma I see: How can we maximize good faith edits (and editors) to pass through, whilst also labeling or flagging edits from IP Proxies?

I'd like to understand which "moment" in the editing journey should trigger a block. Is it better to throw up a block immediately, prohibiting the edit from being attempted, or after the user tries to submit the edit? For example, today, users on IP Relay on Safari see a message their IP is blocked. This might inhibit good faith edits from being attempted.

As to the specific proposal: I'd be concerned if an Abusefilter rejected an edit just because of the IP address, but perhaps Abusefilters can be set to only "flag" edits as being from a proxy.. JWheeler-WMF (talk) 17:20, 11 March 2025 (UTC)

I'd be concerned if an Abusefilter rejected an edit just because of the IP address

Why so? It's pretty much a corollary of the NOP policy. Aaron Liu (talk) 18:06, 11 March 2025 (UTC)

If an AbuseFilter rejects 100 edits due to a proxy IP, it's likely that a % of these edits were made in good faith edits, which could lead to poor editor retention. As @Robertsky says, the construction of the filter could be gradual so that edits could be monitored / flagged first. Of course, an Abusefilter can already reject edits from IP addresses, which can prevent vandalism. JWheeler-WMF (talk) 19:45, 11 March 2025 (UTC)

Good faith editors can follow directions - such as the one that says to stop using a proxy. — xaosflux ^Talk 22:02, 11 March 2025 (UTC)

I rather have the option to reject/block be open and let the project admins decide how to construct the filters to flag or reject. From my observation on English Wikipedia there, admins and AF managers takes gradual steps before enacting direct rejections of edits, i.e. filter and monitor/flag first, etc, and then decide on a case by case basis that the "moment" to trigger block/rejection of edits as each case of long term abuses is unique. Robertsky (talk) 19:07, 11 March 2025 (UTC)

That would also be fine. AFAIK filter options usually work this way—the action left to community configuration. Aaron Liu (talk) 22:51, 11 March 2025 (UTC)

Update on wish

Hi folks - this wish is in development by the Trust and Safety team, and you can track along here. https://phabricator.wikimedia.org/T354599 JWheeler-WMF (talk) 19:09, 13 March 2025 (UTC)