Thing is, "content warnings" are a text field on Mastodon. The exact same text field that Identi.ca introduced as a summary field. The exact same text field that everything that can do long-form blogging still uses as a summary field, including Friendica and its descendants. Which, by the way, is a summary field in the W3C ActivityPub spec as well.

The content warning flag that blurs images is a non-standard Mastodon thing. It's actually half-independent from the content warning field. If you put a Mastodon toot under a content warning, Mastodon sets the content warning flag for all media attachments. But Mastodon does not set the content warning flag for external, non-Mastodon content with a "content warning". The flag is set upon sending, not upon receiving.

I guess this is intentional. If Mastodon did that, each long post with a summary that comes in as a Note-type object (= Mastodon does not hide it behind a link to the original) would have its perfectly SFW images blurred out, just because the post has a summary which Mastodon automatically understands as a CW.

The only other Fediverse software I know about that can set Mastodon's content warning flag are (streams) and Forte. And they set that flag for everything embedded into/attached to a post that contains the hashtag #sensitive and/or #nsfw. I don't think this is documented anywhere.

As for filters on Mastodon, they're cumbersome. You want one keyword filtered, well, go click together a whole new filter. No such thing as keyword lists like on Friendica & Co. Nope, each keyword requires its own filter. In addition, I guess many Mastodon apps don't even offer UI elements for the filters, and next to everyone on Mastodon only ever uses it on a phone with an app. That's why Mastodon users resort to muting, blocking or calling the mods in cases in which everyone else would just add a filter.

Wafrn, in turn, is literally decentralised Tumblr, just like Lemmy is literally decentralised Reddit. With the exact same mentality and the exact same culture as the respective original silo. A culture in which hashtags in the message text body don't exist, in Tumblr's and Wafrn's case. (In Reddit's and Lemmy's case, hashtags don't exist at all, technologically and thus culturally. Just for comparison.)

The Wafrn users, all former Tumblr users, will never add any hashtags to their post texts. That's primitive and un-Tumblr-like to them. For them, it's the same dedicated keyword field as on Tumblr (which Mastodon won't support, which even ActivityPub has no support for) or bust.

I think Wafrn doesn't have and doesn't support the summary field (= Mastodon's CW field) either because Tumblr has no such thing. On Tumblr, the only way to keep NSFW content away is to choose wisely whom to follow. But to my best knowledge, Tumblr has never had a public feed (for very good reasons), and I guess Wafrn doesn't have one either (for the same reasons), whereas Mastodon does have a public, federated timeline in which everything washes up that happens on the server either way. One Mastodon user who follows a Wafrn account that posts gay porn will inevitably taint that Mastodon server's entire federated timeline.

Combining Tumblr's culture (hashtags in post bad) and Wafrn's lack of certain Mastodon features (summary/CW field), Wafrn users can post NSFW content either up front into everyone's faces with no filter or not at all. And Wafrn users (= Tumblr refugees) associate not being allowed to post outright pr0n with Yahoo's takeover of Tumblr and forcing it into being G-rated.