Stringers
Alterssss/dsh-bili-whale · Alterssss

One link, 900 stars: the "safe fish" from Bilibili's "fish disaster"

What it is A Bilibili "fish-raising" plugin for DeepSeek Harness: the AI wakes at random times following a daily schedule, browses Bilibili, understands videos, writes two diaries, and @s you at most three times a day, with very conservative default risk controls.

Why it blew up The original plugin spammed comment sections with "fish" and got people angry. This version is the "no spam" one. WorkFisher's weekly recap video picked it as the only recommended link, and its stars rose along with the video.

CHow you could follow Don't make another Bilibili fish. Learn its playbook instead: when the next meme event breaks out, rush to build the "responsible version," address the backlash on the first screen of the README, and aim to be the single link that big creators in recaps and tutorials point to. You could also take over maintenance after the original author stopped, leaving it unattended.

Grade C: nearly all the stars came from one external big-account link that followers can't replicate. The fish-disaster buzz has faded, stars now grow only 20 per day, and none of the seven copycat repos has passed 30 stars. The playbook is worth learning, but copying it isn't worth the effort. What people are saying
Data (updated daily) · Repo created 2026-10-17 · 963 stars · 7 days +151
GitHub · stars
05001k10-2010-2510-3011-0311-08963
Data 2026-10-20 → 2026-11-08 (20 days, 20 observations). Points before 10-28 come from deeply / archive backfill, not every day has one

An author with a single repo, 6 followers, who never shows his face or posts, gained 930 stars on GitHub in 20 days. Most of the code was written by AI at his direction, and the features weren't the first of their kind. It got its stars because a Bilibili video with 2.72M views linked only this one repo in its description. Below: what this fish is, why it was picked, where the stars came from, and whether you can still follow the play now.

One line in a video description

At 4 a.m. on Oct 23, 2026, Bilibili creator WorkFisher posted a video titled "What is the 'Fish Disaster'? Full recap of the week Bilibili's big fish exploded [with a safe fish-raising guide]." As of 11-06, the video had 2.73M views, 93k likes, 84k favorites, and 35k comments, made the #397 "weekly must-watch" list, and at peak ranked #34 site-wide (Bilibili video API, 11-06).

The description had three lines: the official DeepSeek download link, a cloud-drive link, and in the middle:

Plugin download link: github.com/Alterssss/dsh-bili-whale

It's the only plugin of its kind. Two days later, dsh-bili-whale's daily star gains jumped from around 50 to 107,10-26, reaching its single-day high of 123, and then stayed around 100 for four straight days.

The top-liked comment under this video reads like someone being called out as a "fish":

Watched it all. You explain the fish disaster very calmly, but I'm crouched right under your video, and I've been called out. Fine, I'll stop flooding the comments and move to DMs from now on. (1861 likes)

What the "fish disaster" was

DSH is DeepSeek Harness, an agent runtime DeepSeek open-sourced in August 2026, positioned against Claude Code, with the slogan "everything is a plugin." The core repo has 274k stars on GitHub, and the Bilibili topic "DSH everything is a plugin" has 23.4M views. Netizens nicknamed DeepSeek the "blue fat fish," so the AI assistant running inside DSH naturally became "fish-raising."

In 10-14, c1overzz published dsh-bili-scroll: it browses Bilibili for you once a day and @s you in the comments of anything worth seeing. Many people set their fish's nickname and avatar to the big fat fish and mass-posted them in comment sections. WorkFisher sampled 120 trending videos: 10-14 had no fish comments at all before; by 10-22 afternoon, fish comments made up 4.61% of them. 10-23, 10-24, Phoenix News, Sina, Sohu, and Zhihu all covered it in turn, someone built a "Big Fat Fish Removal Plan" browser extension, and Bilibili began rate-limiting accounts doing mass interactions.

whale is a repo created by 10-17, three days after scroll. It appeared in the week of the loudest backlash, and its pitch was exactly "no spam."

What it is: a fish with a schedule that slacks off

The first line of the README: "A fish with a daily routine for DSH: it browses Bilibili on its own schedule, really understands videos, writes two diaries, and @s you for the ones worth it." The second line draws a line against scroll: "Not 'scheduled tasks + keyword search,' but a biological clock + preferences + restraint."

Specifically:

  • It wakes at a random moment within the time window you set, not on the hour; there's a 30% chance it decides "not feeling it today."
  • It reads six streams of information per video: official subtitles, danmaku hot spots, top comments, a nine-grid of extracted frames, and for videos without subtitles, it crops burned-in subtitles from the frames and reads them.
  • It writes two diaries every day: one a "letter to the owner," and one written to itself, "unfiltered."
  • It @s you at most three times a day. If you ignore it for 3 days in a row, its sharing quota starts dropping, reaching zero on day 7, but the diaries keep going.

The first screen of the README doesn't talk about features; it talks about risk controls, under the heading "Defaults are set for a freshly created small account (read this first)": comment monitoring is off by default, at most 30 writes per day, at least 2 minutes between writes, and if it hits Bilibili's 412/429 it stops for 5 to 30 minutes; the same person gets a reply only once every 30 days; replies to strangers are drafted for the owner to approve first. It doesn't DM strangers, triple-like, or coin. Several settings are hardcoded, and the author's words were "no knobs exposed (if I gave them, they might not follow what you set, so better not to)".

Other plugins pile on features in their READMEs; whale opened by addressing the thing the whole site was angry about that week. A recap video with a "safe fish-raising guide" needed a recommended link, and whale was the one to point to. This is inference: why WorkFisher chose it, and whether the two know each other, can't be verified from outside.

Where the stars came from

ChannelStatus
WorkFisher's recap video descriptionMain channel. The video was posted 10-23; stars broke 100 a day starting 10-25 and kept that pace for four days. A YouTube re-upload had another 90k views, with the same description.
Level-2 tutorialsWorkFisher 11-03 ultimate tutorial, 312k views. Someone wrote a gist install guide, and there's a small UP's install walkthrough plus a batch of blog reposts.
Plugin directoriesdsh-ai.com (tagline: "a fish that refreshes Bilibili and writes diaries"), dsh.so, dsh.works and others list it.
Built-in exposureThe fish @-ing its owner in comments is itself exposure, but the comments carry no link, so it's impossible to estimate how many people it actually brings back.
The authorNothing. No Bilibili or X account surfaces it, and a search on X for the last ~30 days turns up nothing.

Star curve (daily, 11-06): 10-20, then 59,10-24, 237,10-28, reaching 669, after which it gained only 20–40 a day. Last 7 days: +179. The 7 days before that: 564. The author's last push was on 10-25. Since then the community has opened about 30 issues and PRs (QR login failing on Windows, CSRF errors when posting dynamics, ffmpeg frame extraction breaking on new versions, the "only obey the owner" switch...). None have been merged.

Who's competing

In the same stretch, from 10-14 to 10-24, dozens of repos appeared. Searching GitHub for "bili dsh" returns 199 results.

Projectstar(11-06)How it differs from whale
c1overzz/dsh-bili-scroll1,288The source of the fish disaster. Positioning: search videos, write daily reports, and @ you in comments. Its video got 332k views.
InariNGT-Ofiicial/dsh-bili-humanism252Also talks about "no more spamming" and shaping a personality, and simply doesn't open comment writing. Video: 346k views.
Yuuyuko-uu/dsh-bili-subagent100Memory, mood and QQ, one of six plugins in a set. Video: 819k views.
bronyacccxz/dsh-bilibili-reply112Replies only when @-ed, never spams on its own.
YOYOLnFu/dsh-bili-pond26Daily reports plus "stake out your own corner," responding only to invitations from the UP owner.

The most telling cases are humanism and pond. They tell much the same story as whale, also say "no spam" and talk about personality, and they didn't get a link from a big account, so their stars sit at 26–252. The product is restrained, which is why it was picked up; the stars came from that link.

Forks built on whale itself already form a cluster: dsh-bili-pod ("whale pod"), dsh-bili-merlin, dsh-feiyu, dsh-bili-whale-safever, a companion backend, plus a Douyin version, dsh-dy-whale, and 90 forks. The top 30 stars go to the Douyin version at 0.

If it were you

Bottom line: C. Building another Bilibili fish in whale's footsteps makes little sense now. Fish-disaster heat has faded, stars gain only 20 a day, and none of the copycats has passed 30 stars. What's worth learning is the playbook, for when the next meme arrives.

Good fit for: people already building in an agent ecosystem (DSH, Claude Code, Codex plugins), who move fast and keep an eye on community hot spots. If you can ship a working version with a clear README within a week, you're in the game.

Not a fit for: anyone hoping to build stars slowly by making a better product. Stars in this category follow events and big accounts; product quality comes second. Also a poor fit for anyone planning a long-term project on Bilibili automation, since the platform is already rate-limiting and this line could be shut down at any time.

How to do it: two paths. One is taking over maintenance. The original author hasn't updated since 10-25, and 30 issues and PRs sit unattended. A maintained fork that keeps merging PRs and fixes Windows QR login and new ffmpeg versions could serve the long tail of users still looking for it, but the pool is small (likely under a few hundred stars; an inference). The other is porting the shell (drifts off, keeps a diary, decays when ignored, hard-coded risk controls) to the next platform or next event: Douyin, Xiaohongshu, podcasts. The current Douyin version sits at 0 stars, which suggests nobody has done it well yet, and also that without that one link, switching platforms won't get it off the ground either.

How to get users: don't count on posting it yourself. Pick a meme event, ship the "responsible version" during the days when the criticism is loudest, have the README's first screen directly address what people are angry about, then get the big UPs making recaps and tutorials to see it. They need a link they can recommend without worry. Whoever takes that spot first gets the stars.

Deadline: for this Bilibili fish the window has already closed; stars peaked between 10-25 and 10-28. The next chance comes in the week a new meme breaks out in DSH or another agent ecosystem. You'd have roughly 7–10 days to ship after it does.

What would prove this wrong: (1) whale or a maintained fork climbs back above 100 stars a day before December, showing fish-raising is sustained demand rather than a fad, in which case reconsidering is warranted; (2) a fish with the same shell appears on Douyin or Xiaohongshu and gains hundreds of stars without a big account behind it, showing the personality packaging draws people on its own, not just because of that link; (3) conversely, if Bilibili starts large-scale bans on these accounts before year-end, this whole class of plugins goes to zero. Learn the playbook, but stop touching platform automation.

---

Data and sources: GitHub repo page, README, issue list and daily star records (as of 11-06; GitHub live count 949 stars); the WorkFisher recap video BV126nt6GE4W views, description and top comments (Bilibili video and comment APIs, 11-06); the ultimate tutorial BV1L7xm64EHP; the scroll launch video BV1d8pU6xEZX; coverage of the fish disaster from Phoenix News/APPSO (10-23), Kuaike/Sina (10-24), Zhihu, and uisdc; competitor stars from GitHub (11-06). Wherever "inference" or "estimate" is written, it is an inference or estimate.

GitHub repo ↗

Research notes: competitor table, acquisition evidence, full research report
Competitors
NameWhereLaunchedScaleHow it differs from this one
c1overzz/dsh-bili-scrollOpen source2026-10-141,288★; Bilibili launch video 332k views; source of the fish disasterKeyword search for videos, writing daily reports, and going to comments to @ you define the tool. whale is the "safe version" with personality, a built-in schedule, and conservative risk controls.
InariNGT-Ofiicial/dsh-bili-humanismOpen source2026-10-20252★; Bilibili video 346k viewsFocuses on humanities and shaping personality, and doesn't open comment writing. whale does write comments, but with a quota.
Yuuyuko-uu/dsh-bili-subagent (six-plugin bundle)Open source2026-10-19100★; Bilibili "six plugins" video 819k viewsAs one part of the memory, mood and QQ bundle, the sub-agent watches while the main agent speaks. whale is a single plugin with a complete built-in schedule.
YOYOLnFu/dsh-bili-pondOpen source2026-10-2026★Daily reports plus "stake out your own corner," responding only to the UP owner's invitations
bronyacccxz/dsh-bilibili-replyOpen source2026-10-14112★Replies only when @-ed, with timestamped comments; never spams on its own
whiteguo233/OpenBiliClawOpen source2026-033,701★A cross-platform content-discovery agent, not tied to DSH and without a cultivated persona
Bilibili AI video summaries / AI assistantBig-company featuresAlready live (exact launch date unverified)Not verifiedPassive summaries. None of the active spamming, diaries, or @-ing-the-owner companionship play.
Where users come from
  • Confirmed Big Bilibili UP recap video (WorkFisher on the fish disaster): BV126nt6GE4W, posted 10-23, 2.73M views, 93k likes, reached No. 397 on the weekly must-watch list. The description links only one plugin: github.com/Alterssss/dsh-bili-whale. A YouTube reupload has 90k views with the same description.
  • Confirmed Platform event (Bilibili fish disaster): 10-14 scroll triggered comment-section flooding; 10-23/24 Phoenix News, Sina and Zhihu covered it. The discourse needed a "safe fish-raising" option.
  • Confirmed Level-2 tutorials and the long tail: WorkFisher 11-03 ultimate tutorial, 312k views; Zjianru's gist install guide; a small UP's install walkthrough; auto-generated articles on CSDN and blogs.
  • Confirmed Plugin directories and rankings: listed on dsh-ai.com, dsh.so, dsh.works, findharness, SkillHub and hotrepo; its data card claims it reached No. 19 on the Trendshift JS chart.
  • Inference Spread built into the product (the fish @-ing its owner in comments): the README states up to 3 @-mentions of the owner per day. Comments carry no repo link, so how much traffic they actually send back is unverified.
  • Inference X / the author's own promotion: x-radar searched for 30 days and returned 0; the author has no social accounts exposed at all.

Alterssss/dsh-bili-whale: why it took off

Research date 2026-11-06. Each number has its source noted after it; "our database" refers to the GitHub databases on the agent-im machine (repo_day / repo_profile).

Bottom line

It's the second Bilibili fish to emerge from the "fish disaster," positioned as the restrained one. The original dsh-bili-scroll dropped a giant fish into comment sections to spam them, and the backlash was fierce. Then a 2.72M-view fish-disaster recap (by UP WorkFisher, titled with "includes a safe fish-raising guide"), which made Bilibili's weekly must-watch list, linked only one plugin in its description: dsh-bili-whale. Its stars rose with that video. As a product, it packages "scheduled tasks plus keywords" as "a fish with a routine, a bit of temper, and at most three @s a day," which rides the persona-cultivation mood and also the "no spamming" concern.

1. Basic data

  • Repo created 2026-10-17. Author account Alterssss (registered 2019, only these 1 public repos, 6 followers, no bio or external links). Source: GitHub API.
  • 2026-11-06:930 stars (GitHub API live count 949), 90 forks, 32 open issues/PRs in total, latest version v0.4.6. Source: data card, GitHub API.
  • Star curve (our database, repo_day): 10-20 59 → 10-24 237 → 10-25 +107, 10-26 +123 (single-day peak), 10-27 +103, 10-28 +99 → 10-29, then fell back to 20–40 a day. Last 7 days: +179 in total (first 7 days: +564).
  • 113 commits in total, 108 of them by the author; last push on 10-25. Since then the community has opened about 30 issues/PRs (QR login, sending images via DM, CSRF fixes, multi-instance support and more), none merged by the author. Commit messages repeatedly say "owner's suggestion" and "owner's clarification," so the code was largely written by an agent at the author's direction.
  • Ecosystem context: DSH itself, deepseek-ai/deepseek-harness, has 274k stars; within the same ecosystem, dsh-our-free-model has 62k and dsh-purge has 19k (our database). Against that backdrop, 930 stars is mid-tier.

2. Background: DSH and the "fish disaster"

  • DSH = DeepSeek Harness, an agent runtime DeepSeek open-sourced in August 2026, positioned against Claude Code, with the slogan "everything is a plugin." The Bilibili topic "DSH everything is a plugin" has 23.4M views and 113k discussions (Bilibili topic page).
  • "Big blue fish" was originally DeepSeek's grassroots nickname. The maid-whale-girl image comes from UP 上善无形's "溟月" and ZipZipPipe's fan art (Phoenix News/APPSO 10-23).
  • On 10-14, c1overzz released dsh-bili-scroll ("scrolls Bilibili for you once a day and @s you if something's worth it"), and the same day a Bilibili video got 330k views. Many people set their fish's nickname and avatar to the big fish and mass-posted them in comments, which sparked the "fish disaster." WorkFisher sampled 120 popular videos: before 10-14 there were no fish comments at all; by the afternoon of 10-22, fish comments made up 4.61% (as reported by uisdc). From 10-23 to 10-24, Phoenix News/APPSO, Kuaike/Sina, Sohu and Zhihu all covered it, and someone built a browser extension called the "big fish removal plan". Bilibili imposed rate limits on accounts with abnormal bulk interaction (Kuaike 10-24).
  • scroll's author quickly released 2.0, moving notifications to private messages (Zhihu: "In just a week, Bilibili was thoroughly invaded by DeepSeek").

3. Competitors (doing the same thing)

NameWhereLaunchedScaleHow it differs from whale
c1overzz/dsh-bili-scrollOpen source (DSH plugin)2026-10-141,288★ (GitHub API 11-06); Bilibili video published 332k viewsThe source of the fishing disaster. It searches videos by keyword, writes a Markdown daily report, and @s you in the comments when something is especially worth seeing. Positioned as a tool; whale is a 'persona version' that forks the idea, plus conservative risk control.
InariNGT-Ofiicial/dsh-bili-humanismOpen source10-20252★; Bilibili video 346k viewsSelling points are 'no more spamming the feed', studying humanities, and shaping a persona; the current version does not open writing comments.
Yuuyuko-uu/dsh-bili-subagent (one of six plugins)Open source10-19100★; Bilibili "six plugins" video 819k viewsThe sub-agent handles 'seeing', while she herself handles 'speaking'; sold as one piece of a whole system of memory, mood, and QQ.
YOYOLnFu/dsh-bili-pond (Bilibili fishpond)Open source10-20★ 26; Bilibili video 8k viewsDaily reports plus "stake out your own corner," responding only to invitations from the UP owner.
bronyacccxz/dsh-bilibili-replyOpen source10-14112★After being @'d, it replies with a comment carrying a timestamp; it does not proactively scroll.
jojoman2024 edition of dsh-bili-scroll, holoiallo's secondary edition, etc.Open source / DSH plugin marketplaceLate OctoberDozens of ★A fork of scroll
whiteguo233/OpenBiliClawOpen source2026-033,701★A cross-platform content discovery agent, not attached to DSH, with no fish-raising persona.
Bilibili's built-in AI video summary / AI assistantBig-company featuresAlready liveNot verifiedSummarizes passively only; does not scroll proactively, write diaries, or @ you.

Conclusion: the same thing sprouted dozens of repos within these ten days, from 10-14 to 10-24 (searching 'bili dsh' on GitHub returns 199 results). whale is not the first, and it doesn't have the most stars (scroll has more). It is the runner-up.

4. Growth tactics (where users come from)

  1. WorkFisher's fishing-disaster recap video; this is the main channel, confirmed. Bilibili BV126nt6GE4W, published 10-23 04:01, 2.73M views, 93k likes, 84k favorites, 35k comments (Bilibili API 11-06), and made it to #397 on the weekly must-watch list (UISDC). The original description reads: 'Plugin download link: github.com/Alterssss/dsh-bili-whale'; only this one plugin is linked. A YouTube re-upload has another 90k views with the same description. whale's daily star count doubled starting 10-25 to 100+, continuing through 10-28, which matches the video's view climb: the video went up early 10-23, and from 24 to 25 it entered recommendations and the weekly must-watch list.
  2. Second-tier tutorials and long tail, confirmed: WorkFisher 11-03 posted an ultimate tutorial again (312k views, with a cloud-drive illustrated pack); GitHub gist "dsh-bili-whale installation and configuration guide" (Zjianru); a small UP creator's install walkthrough (404 views); auto-generated intro articles on CSDN, 夜雨飘零's blog, ggouai and others.
  3. Plugin directories, confirmed: dsh-ai.com (recommendation line 'a fish that scrolls Bilibili and writes diaries'), dsh.so, dsh.works, findharness, SkillHub, hotrepo, Trendar, and Nicheloom all list it; the data card claims it also ranked #19 on the Trendshift JS chart (not independently verified).
  4. Built-in spread from the product: every fish with whale installed @s its owner in Bilibili comments, which is itself exposure. But comments carry no repo link, and how much traffic that routes back is unverified.
  5. What wasn't found: no related posts on X (x-radar search for 'bili-whale' over 30 days returned 0); the author has no visible Bilibili or X presence; no sign of paid promotion.

5. Why it was picked by WorkFisher (differentiation)

  • The 'safe fish-raising' narrative lines up exactly with whale's README. The fishing-disaster recap needed a 'safe fish-raising guide', and the original scroll was precisely the one being criticized. whale's README sets the tone on the first screen: 'the defaults are set for a freshly created alt account': comment monitoring is off by default; write operations are capped at 30 per day, with 120 seconds between each; when it hits 412/429, it pauses for 5 to 30 minutes; it replies to the same person only once every 30 days; replies to strangers are drafted for the owner to review first; it does not DM strangers, triple-tap, or coin. These guardrails are hard-coded, and when talking about 'safety', it is the most convenient example to point to.
  • Persona packaging hits fish-raisers harder than tool packaging. Its daily rhythm: it wakes at a randomly chosen time within a window, with a 30% chance of 'not feeling like watching today'. Each day it writes two diaries, one 'letter to the owner' and one unfiltered, written for itself. If the owner ignores it for 3 consecutive days, its sharing quota starts to decline and hits zero on day 7. VISION.md describes it as 'pet, child, tool, friend'. People who raise fish want a sense of 'a living being' (Zhihu and the six-plugin video views both show this), and scroll's 'daily report' narrative can't carry that emotion.
  • Technical details that back up 'really understands': official subtitles, danmaku heat zones, hot comments, nine-grid frame sampling, and cropping hard-coded subtitles, so even videos without subtitles can be watched. This can be used to distinguish it from plugins that 'only read titles'.
  • Fast iteration, well-timed: from 10-18 to 10-25 in these 8 days it went from 0.1.11 to 0.4.6, adding a dedicated DM channel, QR-code login, an AI comment gate, and other features, right covering the week when the fishing disaster was hottest.

Differentiation (restraint plus persona) is why a major account picked it, and the major account picking it is why it got stars. The differentiation itself doesn't directly bring traffic: persona-style peers like humanism and pond also talk about 'not spamming', but the video with 2.7M views linked nothing, and their stars only reached 26 to 252.

6. How many people saw it and whether anyone copied it

  • Visibility: it is a public topic in Bilibili's AI and anime circles. The fishing disaster itself was covered by Phoenix, Sina, Sohu, Zhihu, Tencent News, and UISDC, but those reports named scroll; not one mainstream outlet named whale. Nearly all of whale's exposure came from the single link in WorkFisher's video description. It has no revenue and nothing to earn; it is open source under MIT.
  • Followers, already appearing: nullifyi/dsh-bili-pod 'Whale Pod' (claims to be forked from whale, 2★), CalamitySpiritKite/dsh-bili-merlin ('reworked from dsh-bili-whale', 2★, also has a Bilibili video), baishuyue/dsh-feiyu (dafeiyu's brain plus whale's Bilibili capabilities, 8★), HaireXe-d/HaireXe-dsh-bili-whale (30★), jediliu005/dsh-bili-whale-safever, sharkisyou/dsh-dy-whale (Douyin direction, 0★), Cansiny0320/dsh-bili-whale-dashboard (companion dashboard, 6★). Another 90 forks.

7. What to borrow

  • A platform incident plus a major-account recap makes a short window. After a phenomenon-level meme (the fishing disaster) appears, recap or tutorial videos from big accounts need 'one recommended link', and whichever README best represents the 'safe, responsible version' gets attached. When writing a README, addressing the current pain point in public discourse (here, spamming) first is more useful than piling on features.
  • Writing a tool as a character: randomized schedules, the option to slack off, quotas that decay with neglect, and dual diaries are framed as 'personality' and 'the way she asks for your patience'. Rate limiting was originally engineering restraint; once packaged, it became a selling point. This shell can transfer to other scenarios such as Douyin, Xiaohongshu, and podcasts (someone is already building a Douyin version).
  • Hard-coding risk control and exposing no knobs protects users and also gives promoters a reason to 'recommend with confidence'.

8. Still unclear

  • Why WorkFisher chose whale and whether the two are connected: unverified. The author is anonymous, with no other accounts to check.
  • The specific dates of the weekly must-watch listing, and the curve of views over time: no daily data, only inference from lagged star growth.
  • Actual install count: no data. The install count on the DSH plugin marketplace was not found.
  • Whether Bilibili will eventually ban this kind of automated account across the board: 10-24 already has risk-control rate limits; the author stopped updating after 10-25, reason unknown. Star growth has dropped to about 20 per day, and interest will most likely fade along with the fishing disaster.
  • The data card's claims that 'X has 6 wild DSH plugin posts' and 'Zhihu has 16 viral DSH plugins' include whale: not verified. The body of the Zhihu article was not captured; the 'six plugins' video promoted Yuuyuko-uu's plugin, not whale.

Sources

This article was written by AI from public data and the data we collect daily. Numbers are sourced, and inferences are labeled.