
A meme earned 6000 stars: ASu-skills turned "Su-god résumé packaging" into a serious job-hunting plugin
What it is A Chinese job-hunting plugin pack installed into 8 AI coding tools such as Codex, Claude Code, Trae and Qoder. 9 entry points cover open-source contributions, résumé editing, résumé creation, job matching, filling application forms, interview follow-up questions and campus recruitment progress.
Why it took off It launched on the day the ASu résumé-packaging gossip peaked, so gossip traffic brought the first wave of stars. The author quickly turned the joke requests into real features and added a "no fabrication" factual boundary, which fit the autumn campus recruiting season. Dozens of job-seeking students then submitted PRs to get their work into résumés and spread it through their own circles.
CHow you could follow Don't build another Chinese job-hunting skill; learn its playbook instead: when the next hot figure or event emerges, turn it into an installable skill within 48 hours to cold-start with memes, then legitimize it as a proper workflow the same day and flood Chinese agent install channels.
- Data (updated daily) · Repo created 2026-08-12 · 6,275 stars · 7 days +151
GitHub · stars Data 2026-10-09 → 2026-11-08 (31 days, 31 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
On the afternoon of Aug 12, 2026, Nanjing developer Hisn00w, who had at most 12 stars before, created a GitHub repo called ASu-skills. The first version's invocation phrase was "I want to Su-ify". 44 minutes later the first issue arrived: "Wow, it's blowing up!" Three days later it had 749 stars; by Nov 6 it had 6,257 stars, 119 PRs and 62 contributors. Functionally it was about the same as a dozen Chinese job-hunting skills that already existed, which topped out at 724 stars. Its extra order of magnitude came from a meme, a timely pivot and campus recruiting season.
What it is
A Chinese job-hunting plugin pack installed into AI coding tools, with 9 entry points: /contributor finds real open-source issues for you to submit PRs to; /evidence-recap reconstructs your AI coding conversations into an evidence chain showing "what you did"; /project-guide organizes a project into a study guide and interview Q&A; /great-resume edits your résumé; /make-resume produces an editable HTML résumé (18 templates, exportable to PDF and LaTeX); /job-match checks fit against a job description; /job-apply fills recruiting-site application forms in your already-logged-in browser via Kimi WebBridge and stops before submitting; /interview predicts interview questions and follow-ups; /offer tracks campus recruitment progress.
There are 8 places to install it: Codex, Claude Code, DeepSeek Harness Desktop, TraeWork, Qoder, Cursor, OpenCode, WorkBuddy. The Chinese tools are all covered, so students in China can use it without a VPN.
The meme
"ASu" is "阿酥". The Xiaohongshu AI tech blogger "阿酥在coding" was exposed by netizens in summer 2026 for résumé "packaging": a degree from a US bachelor's program, a frontend internship rewritten as AI algorithm and Agent core contributor. In August it was also reported he had joined Meituan's Beidou program at a package of about ¥2M a year. Tencent News published "Understanding ASu, Becoming ASu" on Aug 16, Zhihu had a question "How do you evaluate AI blogger ASu-in-coding joining Meituan?", and there were long posts in X's Chinese-language community.
ASu-skills launched right when the gossip was at its peak. The first README's title was "Chinese Résumé Packaging and HR Opening Self-Introduction Skill", and the acknowledgments said the content was compiled from public shares by ASu and another blogger, "Hi Mr Lonely". /make-resume's default template is still called "ASu-style high-density technical résumé" to this day, as the README says: if you want "ASu's résumé", use this one.
Almost everyone who showed up on day one came for the gossip. In chronological order, the first issues were: "Wow, it's blowing up!", "May I ask if using this gets me into Meituan's Beidou program?", "I'll study it line by line 🫡", "Can't keep a straight face", "Even keeping a straight face is impressive". About 2.5 hours after the repo was created, @MaxForAI on X (a LobeHub member who says he has 40k followers on Xiaohongshu) posted: "Amazing, someone made the Su-god into a Skill 🤣 Go use it, folks, and you'll get a million-yuan package (I'm serious)". On Aug 15, @QingQ77 posted a feature explainer, and around Aug 16 CSDN and Zhihu also published analysis articles.
How the meme became a product
A meme alone usually stops drawing eyes within a few days. ASu-skills got about 750 stars in its first 3 days; the bulk of its growth, from about 750 to 5,456 (10-09), came over the following seven weeks. The author got two things right during that stretch.
First, play along with netizens and turn the joke into real features. On the second day, someone opened an issue titled "Suggestion: add /contributor, automatically turning your GitHub contribution graph into a résumé project entry": automatically find documentation projects to fix typos in, automatically submit PRs, and after merge generate a sentence like "Contributed to multiple open-source projects…" to put on a résumé, ending with "The above was generated by ChatGPT". This was clearly satire. But the same day, the person who opened the issue submitted a PR building /contributor, and the author merged it that day. It is now the first of the 9 entry points, except the rules were changed: show the diff first, submit only after the user confirms each item, and unmerged PRs can only be described as "submitted".
Another user submitted two parody PRs, titled "Led the development of the ASu-skills open-source license governance system", with the body "As Owner leading the compliance refactor... a single-file change achieved full-chain License coverage of 100%, a milestone achievement". The author merged them too. The README still ends with the line "The open-source governance system was built under community Owner leadership, achieving full-chain License coverage of 100%".
Second, they scrubbed the smell of fakery. On the afternoon the repo was created, one skill was split into three, later expanding to 9 entry points. The README added a "Fact boundaries" section: keep real job titles, companies, dates and degrees; use "led", "lead" or "Owner" only with evidence; don't pass off AI-written code as your own skill. There's also a bolded warning banning bulk applications and harassing HR. That way, people who came for the joke find it fun, and people who really job-hunt feel safe using and sharing it.
Where users come from
| Timeline | Stars (source) | What happened |
|---|---|---|
| 08-12 | Repo created | Gossip issue; @MaxForAI reposted it 2.5 hours later |
| 08-15 | 749 (Wayback snapshot) | @QingQ77 feature-intro post; Zhihu and CSDN explainer articles |
| 09-07 | Single-day +183 (Trendshift) | Peak, right as the autumn campus recruiting window opened |
| 10-09 | 5,456 | 9–October, dozens per day |
| 10-28 | 6,059 (56 that day) | From mid-October, 15–45 per day |
| 11-06 | 6,257 / 326 fork(GitHub) | Last 7 days +149, first 7 days +177 |
The meme only carried the first week. What kept it going after that was these:
- Campus recruiting season. 9, October is when China's campus hiring peaks. I couldn't find the specific post that drove the 09-07 peak; my inference is that the season itself was the main driver.
- Contributors became the distribution channel. The product itself teaches people to "make real open-source contributions and put them on your resume," and it is a Chinese-language repo with several thousand stars, so students submitting PRs can add a line to their CVs. 119 PRs, 103 merged, 62 contributors; Qoder support, OpenCode adaptation and the macOS browser path were all contributed by outsiders. Each contributor takes it back to their circle and spreads it again. Issues include people saying outright "I want to get on the Contributors list" and asking "Can ASu be added as a contributor?"
- Aggregator sites listed it automatically. Skill directories like skills.sh, skillsmp, claudepluginhub, DeepWiki and dsh.so (with a security badge) all indexed it, bringing long-tail traffic.
The fork-to-star ratio is about 5.2%, not high. But skills install through plugins, so forking isn't needed, and there are 100 merged PRs, so the stars don't look manufactured. I didn't check how far it spread within Xiaohongshu.
Who it's competing with
| Project | Scale (11-06) | How it differs from ASu-skills |
|---|---|---|
| career-ops-hq/career-ops | 80,423 stars | English, aimed at overseas markets; scans job boards and batch-scores them into application packs |
| ai-job-search-dev/ai-job-search | 69,831 stars | An English job-search framework built on Claude Code |
| srbhr/Resume-Matcher | 29,153 stars | A standalone app for 2020 years, only doing resume and JD matching |
| loks666/get_jobs | 9,586 stars | Batch auto-apply scripts for platforms like Boss Zhipin; ASu explicitly opposes bulk applications |
| agentenatalie/get-job.skill | 724 stars | Nearly the same positioning, 2.5 months earlier, but no meme and no multi-platform support |
| GresonKwan/JobOK | 438 stars | Also a "evidence-driven Chinese job-search Skill" with similar features |
| liyupi/yupi-skill (a distilled persona of programmer Yupi) | 456 stars | The "turn celebrities into skills" route; even its author's main account has only 456 |
| Claycui828/ASu-resume-skills | 425 stars | A copycat repo released the next day, "ASu-style high-density tech resume" |
Before it, a dozen or so Chinese job-search skills already existed, the top one with 724 stars. /job-match, interview follow-ups and "evidence-driven" were all done by others too. ASu pulled ahead by an order of magnitude, and the difference was all distribution: riding a viral meme, landing in campus season, being installable on domestic agents, and having contributors spread it, all four at once. The copycat repo appeared the next day and reached at most 425 stars: the meme only rewards the first one.
Is there room left in this wave?
This slot for Chinese job-search skills is now taken, campus season is winding down, and ASu's own growth rate is falling. The ASu meme won't go viral a second time. What's still open is applying the same playbook to other Chinese audiences around their seasonal milestones: postgraduate entrance exams (December exams), civil service exams (national exam from late November to December), and study-abroad applications (autumn-winter deadlines). I haven't seen skills with several thousand stars in these areas (I didn't check in detail), but there's no ready-made meme to ride either. Also, the motivation of "contributing PRs to strengthen your resume" is unique to job-hunting and wouldn't carry over to exam prep or civil service.
If it were you
One-line verdict: stop building Chinese job-search skills, that slot has an owner; what's worth learning is the playbook of "cold-start on a hot meme, then rebrand as a legitimate tool," ready for the next meme.
- Best for: people who watch the Chinese AI scene (X's Chinese community, Xiaohongshu, Zhihu) and can ship an installable skill in an afternoon, ideally knowing one or two accounts willing to repost memes.
- Not for: people who want to slowly polish a fully featured job-search tool and wait to be discovered. get-job.skill and JobOK did it that way and stalled at a few hundred stars.
- How to do it: when a public figure or incident goes viral, turn its "methodology" into a skill you can install directly, with the name and trigger phrase riding the meme; on day one, split it into several proper entry points and add a "fact boundaries" and anti-abuse note; take playful issue requests and do the feasible ones the same day, and merge the parody PRs when they're acceptable. Make installation available across Codex, Claude Code, and domestic tools like Trae, Qoder, WorkBuddy and DeepSeek all at once.
- How to acquire users: on the day the repo is created, find AI accounts on X's Chinese community and Xiaohongshu and post in a meme-friendly tone; in the second week, submit to directories like skills.sh and claudepluginhub; write a clear contributing guide so that "submitting a PR" itself becomes something users want.
- When to be early: a meme's window is measured in 48 hours. ASu-skills launched on the day the controversy was fermenting, and the copycat that came out the next day only reached 1/15 of its stars. For seasonal topics, launch three to four weeks before the milestone: the civil service exam is late November, postgraduate exams are late December, and right now that's getting late.
- Signs you misread it: if within 48 hours no major account reposts it and no one in the issues plays along with the meme, the meme isn't big enough or you aren't the first. If it doesn't reach a few dozen stars in a week after the core features are complete, these people aren't on GitHub; stop adding features and change the topic.
Research notes: competitor table, acquisition evidence, full research report
- Competitors
Name Where Launched Scale How it differs from this one career-ops-hq/career-ops Open source 2026-04-04 80,423 star English-language overseas job hunting, scanning job boards and batch-scoring; ASu does Chinese campus recruiting and emphasizes evidence and fact boundaries ai-job-search-dev/ai-job-search Open source 2026-03-18 69,831 star An English job-search framework built on Claude Code; ASu is in Chinese and covers domestic agents srbhr/Resume-Matcher Open source 2020-04 29,153 star A standalone app that only does resume and JD matching; ASu is a full-workflow skill inside the agent loks666/get_jobs Open source 2024-03 9,586 star Batch auto-apply scripts for platforms like Boss Zhipin; ASu explicitly opposes bulk applications and stops before submission agentenatalie/get-job.skill (Intern.skill) Open source 2026-05-29 724 star Nearly the same positioning and launched earlier, but without the meme, the multi-platform plugins or the community; its scale is about 1/9 of ASu's GresonKwan/JobOK Open source 2026-06-17 438 star Also an evidence-driven Chinese job-search skill with similar features, but lacking distribution liyupi/yupi-skill Open source 2026-04-09 456 star A celebrity-distillation skill (programmer Yupi); its author has a main account but only 456. ASu rode a negative meme and built a workflow, not a persona AI resume editing and AI interviews on Boss Zhipin, Niuke and Super Resume Big-company features Already there Large user bases Built into the platforms; ASu runs inside the developer's own Codex or Claude Code workflow - Where users come from
- Confirmed Piggybacking on a trending incident (the "Ah Su at coding" resume-packaging controversy): the first README's trigger phrase was "我要酥化" and it credited Ah Su at coding; 08-12 the first batch of issues included meme-driven requests like "Will using this get me into Meituan's Beidou program?"; Tencent News 08-16 and Zhihu discussed it at the same time
- Confirmed Reposts by large Chinese AI accounts on X: @MaxForAI (LobeHub, 40k followers on Xiaohongshu) posted 2.5 hours after the repo was created, "Someone turned the Ah Su god into a Skill 🤣", with about 5.2K views; @QingQ77 08-15 posted a feature-intro thread
- Confirmed Zhihu / CSDN explainer articles: Zhihu's "I Figured Out the 'Ah Su Resume-Packaging Skill'"; a CSDN campus-recruiting resume tutorial 08-16
- Inferred fall recruiting seasonal demand: peak 09-07 +183/day, 9–October dozens per day, cooling off after fall recruiting ends
- Confirmed Contributor flywheel (job seekers submit PRs to add to their resumes): 119 PRs, 103 merged, 62 contributors; the product's /contributor itself teaches people to make open-source contributions for their resumes
- Confirmed Skill/plugin aggregator sites and Trendshift: skills.sh, skillsmp, claudepluginhub, awesomeclaudeplugins, agentflowshub, DeepWiki, dsh.so, SourceForge mirror; Trendshift HTML chart top 20–25 for 6 days; someone submitted a listing PR to hermes-agent
- Inferred Xiaohongshu: README credits Xiaohongshu bloggers, and YouTube has a re-upload titled "Ah Su helps you fix your resume? Ah Su skill sharing"; the scale of spread within Xiaohongshu was not checked
Hisn00w/ How ASu-skills took off
In short: a Nanjing developer, during the peak of the "Ah Su at coding" resume-packaging controversy, turned the "Su-god-style resume packaging" into a Codex skill, riding gossip traffic to the first wave of stars; over the following month they reworked the meme into a proper Chinese job-search plugin pack (9 entry points, 8 agent platforms), timed with campus recruiting, and with 60 job seekers submitting PRs, growth held up until November.
1. Basic data
| items | Value | Source |
|---|---|---|
| Repo created | 2026-08-12 07:12 UTC (first commit 07:35) | GitHub API |
| star / fork | 6242 / 324 (11-06), live GraphQL reading 6257 | Data card; GitHub API |
| Early curve | 08-15 had 749 stars (Wayback snapshot); around 09-06 about 3700; 10-09 had 5456; 10-28 had 6059; 11-06 had 6242 | web.archive.org 20260815 snapshot; Trendshift "61 days +2526" back-calculated; our GitHub repo repo_snap (from deeply source 10-09 onward) |
| Peak | Single-day peak recorded by Trendshift +183 (09-07); from mid-October, 15–45 per day | Data card recall.found; GitHub repo repo_day |
| Recent momentum | gain_7d 149, first 7 days 177, cooling | Data card |
| Co-built | 119 PRs, 103 merged, 62 contributors, 333 commits | GitHub search / contributors API(11-06) |
| Author | Hisn00w (Hisn0w, Nanjing, 19 public repos, 38 followers); its previous most-starred repo had only 12 stars | GitHub API |
Fork/star ratio is about 5.2%, lower than the median for repos in our database of the same age with ≥5000 stars (about 11%). But the skill is installed as a plugin, so it doesn't need forking to begin with. Combined with 103 merged PRs and 62 contributors, the stars don't look fake.
2. Origin: a meme
- 阿酥在coding (A Su Codes): an AI tech blogger on Xiaohongshu. In summer 2026, netizens dug up his résumé "packaging" (listing a US bachelor's as an elite degree, a frontend internship as core AI algorithm / Agent contributions). Reportedly he got an offer from StepFun, then was talked out of it; in August it was also reported he had joined Meituan's Beidou program, with an annual package of about ¥2M. The story became a major gossip item in AI circles. Tencent News' "理解阿酥,成为阿酥" (2026-08-16), Zhihu's "'阿酥在coding' joins Meituan and gets mocked: packaging résumés isn't scary, what's scary is getting famous," a Zhihu question on AI blogger 阿酥在coding joining Meituan, and long threads on X by @jakevin7 and @yanbo2004, plus debunking posts on Threads, all covered it. Sources: news.qq.com/rain/a/20260816A01VE800, zhuanlan.zhihu.com/p/2071743450492442328, www.zhihu.com/question/2069134194081584733
- The repo's first version was built on the meme: the first README (08-12 commit) was titled "Chinese résumé packaging and HR opening self-introduction skill," with the invocation phrase "我要酥化" (roughly "I want to get Su-ified"); commit history includes "Refine 酥化 method notes and résumé packaging reference" and "Remove résumé packaging reference image." The README credits the content as compiled from public shares by Xiaohongshu bloggers "阿酥在coding" and "Hi Mr Lonely." "ASu" is short for "阿酥"; the default template for
/make-resumeis still called "ASu-style high-density technical résumé," and the README says "I want the same résumé as 阿酥" and routes it through there. - Day one's issues were all gossip (08-12): "Wow, it blew up!" "Does using this get you into Meituan's Beidou program?" "I'll study it line by line 🫡" "Can't stop laughing." Source: GitHub issues sorted chronologically.
- A big account reposted it the same day: about 2.5 hours after the repo was created (08-12 09:43 UTC), @MaxForAI (a LobeHub member who says he has a Xiaohongshu account with 40K followers) posted "LOL, someone turned 酥神 into a Skill 🤣 Go use it, folks, and you'll get a million-yuan package (I'm sure)" which got about 5.2K views (a number from the search snippet, not verified). 08-15 @QingQ77 (Geek Lite) posted a feature rundown. CSDN 08-16 has a tutorial article, and Zhihu has "I Figured Out the '酥神 résumé packaging Skill'." Sources: x.com/MaxForAI/status/2087474969781715094, x.com/QingQ77/status/2088417888705167539, blog.csdn.net/qq_19703985/article/details/163800342, zhuanlan.zhihu.com/p/2071908256692409078
3. Second growth phase: turning the meme into a real product, timed with autumn recruiting
A meme alone usually fizzles within a few days. In the first 3 days it picked up about 750 stars. The bulk (about 750 to 5500) came from mid-August to early October, for these reasons:
- Rebranding to wash off the "fake" vibe. On day 08-12 that afternoon, one skill was split into three; later it expanded to 9 entry points:
/contributor(find real open-source issues and submit PRs),/evidence-recap(reconstruct AI coding conversations into an evidence chain),/project-guide(project study guide and interview prep),/great-resume,/make-resume(18 editable HTML templates, exports to PDF and LaTeX),/job-match,/job-apply(fills application forms on recruiting sites via Kimi WebBridge, stopping before submit),/interview(contract-based follow-up questioning), and/offer(campus recruiting progress tracker). The README has a dedicated "factual boundaries" section (no fabricated experience; strong claims require evidence) and a bolded warning against mass applications and HR harassment. "酥化 packaging" thus became an "evidence-driven job-hunting workflow." That keeps the meme's name recognition without dragging the project into the backlash. - The timing is autumn recruiting season. 9–October is when Chinese campus recruiting peaks, and the spike 09-07(+183/day) lands right as autumn recruiting opens. The grade card also notes it's riding the 9-November autumn recruiting window.
- It installs on 8 platforms, and easily: Codex (send it the repo link and it installs), Claude Code marketplace, DeepSeek Harness Desktop, TraeWork, Qoder, Cursor, OpenCode, WorkBuddy. Chinese agents (Trae, Qoder, WorkBuddy, DeepSeek) are all covered, so domestic users can use it with Claude without a VPN.
- Contributing is itself job-hunting material, creating a flywheel. The product tells users to "make real open-source contributions and put them on your résumé"; the repo is a Chinese project with thousands of stars, so students submitting PRs have something concrete to list. The result: 119 PRs and 62 contributors (Qoder support, OpenCode shared resources, macOS browser paths, and others were all made by external contributors). Each contributor then passes it along in their own circles. The closing line of the README, "The open-source governance system is led by community Owners, with full-chain License coverage of 100%," reads like someone stockpiling résumé material.
- Auto-indexed by skill / plugin aggregator sites: skills.sh, skillsmp, claudepluginhub, awesomeclaudeplugins, agentflowshub, hysenlabs, SourceForge mirrors, DeepWiki, dsh.so (with a security badge). Someone also opened a PR to NousResearch/hermes-agent proposing
asu-job-huntas an optional skill (merge status unverified). These sites bring long-tail traffic.
4. Competitors
| Competitors | Where | Launched | Scale | How it differs from ASu-skills |
|---|---|---|---|---|
| career-ops-hq/career-ops | Open source | 2026-04-04 | 80,423 stars, 7 day +1785 | English-first; scans job boards, scores postings, batch-generates application packs, aimed at overseas roles. ASu does Chinese campus recruiting and emphasizes evidence and factual boundaries. |
| ai-job-search-dev/ai-job-search | Open source | 2026-03-18 | 69,831 star | An English job-hunting framework built on Claude Code. ASu is Chinese and covers multiple domestic agent platforms. |
| srbhr/Resume-Matcher | Open source | 2020 | 29,153 star | A standalone app for résumé and JD matching. ASu is a skill inside your agent and covers the whole chain, from résumé to interview to progress tracking. |
| JOYCEQL/magic-resume | Open source / web | 2024-05 | 11,352 star | Online résumé editor, creation only |
| loks666/get_jobs | Open source | 2024-03 | 9,586 star | Batch auto-apply scripts for platforms like Boss Zhipin; ASu explicitly opposes bulk applications |
| agentenatalie/get-job.skill (Intern.skill) | Open-source skills | 2026-05-29 | 724 star | Almost the same positioning (non-elite undergrads landing big-tech offers, résumé editing, interview experiences), 2.5 months earlier, but without the meme or multi-platform plugins. Its scale is 1/9 of ASu's. |
| GresonKwan/JobOK | Open-source skills | 2026-06-17 | 438 star | Also an "evidence-driven Chinese job-hunting Skill" with similar features. Didn't take off. |
| wyh0626/resume-optimizer, wangyafu/resume-skills, wanyichen06/LLMInternSkill, and others | Open-source skills | 2026-02~06 | 200–560 star | Single-purpose résumé skills |
| liyupi/yupi-skill (a distilled persona of programmer Yupi) | Open-source skills | 2026-04 | 456 star | The "distill a celebrity into a skill" approach. The original figure, Yuji, has his own big account, but it only has 456. |
| AI résumé editing and mock interviews on Boss Zhipin / Nowcoder / Super Résumé and similar platforms | Big-company features | Already there | Large user bases | Built into the platforms, not part of developers' agent workflows. ASu's users are tech job-seekers already using Codex / Claude Code. |
Source: our GitHub repo_profile database (2026-11-06).
Takeaway: Chinese job-hunting skills existed by the dozen before ASu, the best of them peaking at 724 stars. Functionally it has nothing unique. job-match, interview, and evidence-driven design were all done by JobOK and get-job.skill. It outperforms its peers by an order of magnitude, and the difference is in distribution: a viral meme, timing with autumn recruiting, installability on every domestic agent platform, and a contributor flywheel, all at once.
5. Copycats
- Claycui828/ASu-resume-skills (08-13, day two, 425 stars): "ASu-style high-density technical résumé generation and evidence-first résumé auditing"
- youngbeauty/GodSU (08-16,74 stars): "Su-God résumé skill, formatting and polishing"
- wnzzer/susu-style on skillsmp; Ch1nfo/ASu-skills and other repos with the same name
- The author later opened Hisn00w/Asu, "ASu job-hunting Harness engineering" (08-13,42 stars), and Offer-skills
- @Charles77xixi replied under MaxForAI's post, "someone already distilled her a few months ago," suggesting an A Su distillation skill existed even earlier and didn't take off (specific repo not found).
6. How many people can see it
Circulating in niche circles: the Chinese AI developer community (X's Chinese side, Zhihu, CSDN, Xiaohongshu) knows about it, and the skill aggregators index it. Mainstream media covered the A Su gossip itself, not this repo. The open-source project has no revenue. Copycat repos appeared the next day, but all had only a few hundred stars.
7. What to borrow
- Trending figure / event + skill: Turn the "methodology" of a controversial Chinese-internet figure (here, "resume packaging") into an installable skill, and big accounts will be sharing it with a meme-style tone the day it's built. The key is speed: launch within the 48 hours of peak heat.
- The meme only handles cold start; the product must clean up its image immediately: split features within hours, add the "factual boundaries," anti-abuse disclaimer, and full workflow within weeks, so serious users would also install and share it.
- Build around seasonal demand by audience (autumn recruiting, postgraduate entrance exams, civil service exams, study-abroad applications): Chinese audiences have fixed time windows.
- Full coverage of domestic agents: Trae, Qoder, WorkBuddy, and DeepSeek Harness all have install entry points, which greatly lowers the barrier for domestic users.
- Make "contributing" itself something users want: job-hunting projects naturally attract people who farm PRs. Write a good contributor guide and you can get dozens of contributors.
8. Still unclear
- The day-by-day curve for mid-August to early September couldn't be retrieved (the stargazers API returns empty for this account), and no direct evidence was found for what triggered the 09-07 peak (a Xiaohongshu post, a weekly magazine, or autumn recruiting itself).
- The spread on Xiaohongshu (TikHub / Xiaohongshu search not done) and whether the author has any connection to A Su himself (appears not) are unverified.
- Whether any stars were artificially inflated: the fork ratio is low, but PRs and contributors are many, which leans toward real.
- Whether it can retain users after autumn recruiting ends (daily star growth has already dropped from 7 to 177 to 149).