# One open-source piece a week from one person: a skill built for AI short dramas, 6000 stars in three months · GitHub open-source project

> Source: https://stringers.io/en/github/gh_eternityspring__shuohao-skills.html (Stringers / 什么值得做, generated 2026-11-08)
> Stringers answers one question every day: what small product can one or two people build in a few months and actually make money from. Numbers come from 15 databases we collect daily, each with source and date.
> For agents: these verdicts and the underlying databases (verified revenue, App Store / Google Play charts, ad libraries, GitHub, Product Hunt…) are queryable live over MCP — tools today / hot / products / directions / direction / data. Setup: https://github.com/komako-workshop/stringers (the user logs in at stringers.io → "配置 agent" to get a personal URL). Site index: https://stringers.io/llms.txt

> MiniMax H3 and Seedance suddenly made AI short dramas feasible, but people were stuck on character and scene inconsistency. The author on Chinese AI X open-sourced one stage a week with a hook in each post, and made each step an engineered artifact with script checks and polished reports. Good timing and good execution.

[GitHub open-source project](https://stringers.io/en/github/index.md) · written 11-6

eternityspring/shuohao-skills · eternityspring

# One open-source piece a week from one person: a skill built for AI short dramas, 6000 stars in three months

**What it is** A set of skills you install into Claude Code or Codex that turns a novel, step by step, into an outline, character designs, scene and prop designs, a script, staging, storyboards and video prompts.

**Why it took off** MiniMax H3 and Seedance suddenly made AI short dramas feasible, and people got stuck on inconsistent characters and scenes. The author open-sourced one stage a week on X within the Chinese AI community, left a hook in each post, and turned every step into an engineering artifact with script checks and a polished report. The good word of mouth followed.

**B** **How you can follow suit** Stop making another Chinese short-drama skill. Copy the playbook instead: find a place where a generation model has just become usable and no upstream project covers it (or build its English version), output JSON for every step, write rules as scripts, produce one report per step, and run a serial on X with one stage a week.

Grade B: the playbook is clear and can be moved to other categories or languages, but this niche is already crowded (the list collected 83 similar projects, and the repos that copied it only have 190 stars). Serializing also depends on an audience the author already had, which is hard to replicate without one. Being passed around in small circles

Data (updated daily) · Repo created 2026-08-06 · 6,214 stars · 7 days +426

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

shuohao-skills is an "AI short-drama pre-production pipeline" installed into AI assistants like Claude Code and Codex. Feed it a novel, and it produces an outline, character designs, scene and prop designs, a script, staging, storyboards and video prompts in sequence. The author is a front-end developer who was job hunting. The repo was created in Aug 2026 and had 6 stars and 6 forks by Nov 6,075 844. Its scripts are not standout in third-party blind tests. It gained stars for three reasons: video models had just reached a usable threshold, the author serialized one repo like a TV series, and each step was packaged as an inspectable engineering artifact. For anyone who wants stars for their own open-source project, what's worth copying is the playbook, not this niche.

### The Aug 13 post

On Aug 13 the author (called @eternityspring on X, Chinese name Shuohao) posted: "Making AI short dramas, anyone can have AI write a script; writing one good enough to make a good short drama is hard! Today I filled that gap: shuohao-skills/novel-script is open source." Three selling points followed: dialogue is timed at 4.5 characters per second, action is converted to episode length at 2.5 seconds per beat, and each episode has an opening hook and an ending cliffhanger, with a dialogue sheet grouped by character that can be fed directly to voice-over. Then two lines: "1.2k stars in a week, one person plus AI, and the barrier to industrializing short dramas shrinks to just the storyboard step. The storyboard will be open-sourced once I polish it further. Stay tuned!"

The previous post covered art design: "The hardest part of AI short dramas isn't ugly drawings, it's 'looking different.' Same boathouse: this shot has a patched awning, the next shot doesn't. In this episode the suitcase is cowhide, next episode it's plastic." The ending likewise left a hook: "Only the script and storyboard are left."

That's how it was released: one stage per post, each explaining which specific pain point in AI short dramas it solved, with the next episode teased at the end. In the first week after creation, 12–30 people forked it each day (GitHub fork records), the fastest-growing stretch.

### What it actually is

Seven skills, six of them stages in the pipeline and one that outputs character reference sheets separately:

| Stage | What it does |
|---|---|
| novel-outline | Adapting a novel into a short-drama outline: adaptation notes, character list, beats, episode summaries, prop list |
| novel-characters | Character design sheets: personality, look prompts, voice prompts |
| novel-art | Scene and prop art design: 3–5 "consistency anchors" per scene, lighting variants for different times of day, prop scale written into prompts |
| novel-script | Script: duration estimated from dialogue word count, with a hook in the first 3 beats |
| novel-staging | At the start of each scene, where characters stand and which way they face, with a top-down diagram whose characters can be dragged around |
| novel-storyboard | Storyboards: each generation at most 15 seconds, each shot 2–5 seconds, output directly as a production package that MiniMax H3 / Seedance can use |
| character-refs | Generate one front-facing full-body image first as the anchor, and generate all other angles based on it |

Each step outputs JSON the next step reads directly. Every skill ships a zero-dependency Node script that hard-enforces 8–18 rules, plus self-tests that don't call a model. Every step also produces an HTML report with images. Install is git clone plus an install script. No API key needed: it runs on your current AI assistant's quota (repo README, 11-06).

### How the stars came

| Timeline | star | Sources |
|---|---|---|
| 08-06 | Repo created | GitHub |
| About 08-13 | About 1.2k | Author's X post |
| About 08-25 | About 2.0k | a long X post by a writer listing 55 AI-generated video projects |
| 10-26 | 5,068 | Daily star log |
| 10-27 / 10-28 | +273 / +266 in a single day | Daily star log, highest since the repo was created |
| 11-06 | 6,075,fork 844 | Daily star log |

Several traffic sources stacked on top of each other:

- **The author's own X account.** He has posted over 4,000 tweets. He already had a formatting tool repo with 832 stars, and a ready readership in the Chinese AI community.

- **Chinese AI accounts and directory sites passed it along.** In August 19, someone posted a long breakdown. In August 25, a roundup post ranked it #1 among short-drama skills. Late August, a batch of Chinese AI directory sites listed it. In September 8, a Tencent Cloud developer community hands-on article, "Lose a copy of Journey to the West, and AI fully auto-produces a short drama," opened with "saw it on X."

- **A post about receiving sponsorship drove a second wave.** In October 26, the author posted: "Folks, open source is worth it! shuohao-skills just received a sponsorship of 4 figures!... 2 months after launch, 5k stars." The post got 58.3K views, and over the next two days the repo gained over 200 stars. The sponsorship amount is the author's own claim and is unverified.

- No sign of paid acquisition or star farming.

It's cooling off now: +390 over the last 7 days, versus +711 over the prior 7 days, back to 40–70 per day (daily star log, 11-06).

### Why now

Video models have crossed the usable threshold. The Tencent Cloud hands-on article put it this way: "MiniMax H3 can run locally now, and many people feel the barrier to AI short dramas has collapsed." The generation step is no longer the bottleneck. People are stuck upstream: the same character and the same room look different with every new shot. Much of shuohao-skills' effort goes into solving exactly that. On top of that, AI short dramas and comic dramas are hot in China, and SKILL.md has become a universal format that Claude Code, Codex, Qwen, and Doubao can all install, so one package works everywhere.

### Is it the best one? No.

The short-drama skills list awesome-short-drama-skills ran three rounds of blind tests, with two AI judges comparing each skill against a no-skill control group. In the October 18 script test, it lost to the control group in all three rounds across the four rounds it was in. Its earliest rival, zenstory-ai/drama-skills, was among the entries. In the October 31 test of a full 60-episode series skeleton, both judges ranked it behind the control group, with comments like "mostly outlines that stop at the segment or act level."

Its star count tells the opposite story: drama-skills launched three weeks earlier and has 3.8k stars, while this one has 6k. It ranks first among the full-pipeline entries on the list. The lead came from building the entire chain end to end, posting serially, having an existing readership, plus a flowchart that is easy to understand at a glance and report screenshots for each step. The README itself works like a shareable poster.

### Who it's competing with

This category is already crowded. That list was created in October 13 and collects 83 short-drama skills and 16 platforms.

| Project | Form | star(11-06) | Compared with alternatives |
|---|---|---|---|
| Toonflow-app | Open-source platform | 20.8k | Full app, you wire up your own API |
| Huobao Drama huobao-drama | Open-source platform | 17.0k | One sentence in, finished video out |
| zenstory-ai/drama-skills | skill | 3.8k | Three weeks earlier, better script blind-test results, covers review |
| seedance2-skill | skill | 4.4k | Video prompts only |
| screenwriting-skills | skill | 2.0k | Scriptwriting only |
| manju-laoli-skill | skill | 1.9k | Comic drama, nine days later |

One more detail: the list notes that 5 novel-* skills in suihe1/short-drama-production "share the same file structure and examples as shuohao-skills, and the repo does not state where they came from." Someone already copied it directly and got only 190 stars. Copying this niche doesn't get you its stars.

### Beyond stars, what the author got

The first screen of the README isn't a feature list. It says "I'm looking for work": open to remote jobs, and also to semi-partnerships with "a few thousand a month living stipend plus revenue share," with a résumé and WeChat contact attached. Below that are a Ko-fi tip link, a sponsor slot for an AI image and video API platform (with a referral link), and Reelbench Studio, his own paid product: an offline, source-available AI video workbench sold as a one-time purchase at $139 / $259 / $459. How much these have sold isn't visible from outside. For anyone doing open source, the order is worth noting: get the project seen first, then put your reputation in the most prominent spot in the README.

### If it were you

One-line verdict: **Building another Chinese short-drama skill is pointless; that lane is packed. What's worth copying is the playbook: catch the moment a generation model first becomes usable, build the layer upstream that's most lacking, and ship one stage per week with serial posts.**

- **Who it suits** : developers who already have a few hundred to a few thousand readers on X or a tech community, and who turn AI output into things that can be checked. Without an existing readership, serial posts have no audience.

- **Who it doesn't suit** : anyone trying to squeeze into the Chinese AI short-drama lane. There are 83 comparable projects, platform-class repos with tens of thousands of stars, and a copied repo with only 190 stars.

- **How to do it** : two directions. First, change the language or audience. It's Chinese-first, and its character-setting check script rejects English outright, while overseas vertical short dramas are also growing. I don't see a clear leader for an English "novel to storyboard" pipeline (inference; I didn't dig into the English ecosystem). Second, change the category: find another place where generation has just become usable and nobody is covering the upstream work, such as AI comics, AI audiobooks, or AI product images, and copy the structure: JSON output at each step, rules written as scripts, and an attractive report at each step.

- **How to get users** : ship one stage per week, with each post covering one specific pain point ("the suitcase is cowhide in this episode and turns to plastic in the next") and ending with a preview of the next episode. Use the README as the landing page, with the flowchart and per-step screenshots. Once it's built, proactively submit it to similar lists and roundup articles.

- **When to be early** : the first month or two after a new model launches. This one came out in early August and captured the biggest wave in its first week. By the time the list had 83 comparable entries, it was already too late.

- **What would show you misjudged this** : if a short-drama skill with clearly better scripts but a later launch and no existing readership overtakes its star count within a few months, then this category rewards quality over being first and how you ship. The advice above would need to be reversed.

[GitHub repo ↗](https://github.com/eternityspring/shuohao-skills)

**Research notes: competitor table, acquisition evidence, full research report**

Competitors

| Name | Where | Launched | Scale | How it differs from this one |
|---|---|---|---|---|
| zenstory-ai/drama-skills | Open source | 2026-07-16 | 3.8k star | Three weeks earlier, 11–12 skills, covers review and editing. Better script blind-test results, but 37% fewer stars. |
| HBAI-Ltd/Toonflow-app | Open source | 2026-01-29 | 20.8k star | Full platform, you wire up your own API. shuohao is a lightweight skill that only covers the early stages and needs no key. |
| chatfire-AI/huobao-drama | Open source | 2026-01-05 | 17.0k star | A one-sentence-to-video platform, not a skill for agents |
| waooAI/waoowaoo | Open source | 2026-01-22 | 14.8k star | Product form |
| HKUDS/ViMax | Open source | 2025-03-30 | 12.9k star | Academic multi-agent video framework, research-oriented |
| dexhunter/seedance2-skill | Open source | 2026-02-11 | 4.4k star | Only the Seedance video prompt step |
| jtydhr88/screenwriting-skills | Open source | 2026-09-06 | 2.0k star | Scriptwriting only |
| Qwen / Doubao / Jimeng (Seedance) / MiniMax Hailuo | Big-company features | — | Not verified | Big-tech products are agent hosts or generation endpoints. shuohao is an upstream pipeline installed on top of these agents. |

Where users come from

- **Verified** : **the author's X serial launch** : one skill per post, each ending on a hook; the 08-13 post says "1.2k stars in one week... storyboard tool will be open-sourced once I polish it a bit more"; readers built up from the earlier article-tools repo, which already had 832 stars

- **Verified** **Chinese AI accounts reposting and long-form roundups** : @krienknight 08-19 breakdown post; @servasyy_ai 08-25 roundup of 55 product lines, ranking it #1 among short-drama skills; GitHubDaily 10-24 pushed a short-drama skill list, with it in the recommended combo

- **Verified** **"4-figure sponsorship" social proof post** : 10-26 posts, 58.3K views, then +273 / +266 stars over the next two days

- **Verified** **AI nav sites, skill directories, and tech-community reposts** : Sumike, Laidaila, AI Tool Set, AI Shop, batch-listing 08-27~31; Tencent Cloud Community 09-08 hands-on tests (opening with "saw it on X"); skills.sh, skillsmp, YouTube, Bilibili

- Inferred **paid acquisition / star farming** : no evidence found

## Why shuohao-skills (eternityspring) took off

> Channel: GitHub · research day 2026-11-06 · repo [github.com/eternityspring/shuohao-skills](https://github.com/eternityspring/shuohao-skills)

### One line

A front-end developer job-hunting (Wesley Wang / 烁皓, X @eternityspring) wrote an AI short-drama production Agent skill: it breaks a novel down step by step into outline, character bible, scene and prop art specs, episode scripts, staging, storyboards, and video prompts, running in Claude Code and Codex. Three things stacked up to make it take off: **MiniMax H3 / Seedance, the latest wave of video models, made 'novel to AI short drama' suddenly doable** ; **the author open-sourced one stage a week and posted it as a serial thread on Chinese AI X** ; and **it was the first pure-skill project to cover the entire production chain, with a script check at every step and a polished HTML report** . The quality of the scripts is not why it took off: in a third-party blind test, its scripts ranked below the control group that used no skill.

### Data

| Timeline | star | Source |
|---|---|---|
| 2026-08-06 | Repo created | GitHub API |
| About 08-13 (one week) | About 1.2k | Author's X post "1.2k stars in a week" (topicdigg archive 2087862241580453925) |
| About 08-22~25 | 2.0k | servasyy_ai's long X post "I went through 55 AI-generated video pipelines on GitHub" |
| 10-17 | 4,669 | Data card (github library) |
| 10-26 | 5,068 | Data card; the author said that day "2 months after launch, 5k stars" |
| 10-27 / 10-28 | +273 / +266 | Data card, single-day peak |
| 11-06 | 6,075 (live via GitHub API 6,128), forks 844–859 | Data card / GitHub API |

- +7 over the last 390 days, +7 over the prior 711 days (data card). After the 10-27 wave passed, it settled back to 40–70 per day.

- The stargazer list API isn't available (it returns 404 / totalCount 0), so I used fork creation dates for the timeline (GitHub forks API): 08-07~08-13 had 12–30 forks per day, about 150 in a week, the strongest stretch; 09-04~05 had a small peak of 60; 10-26~27 had a peak of 65, which matches the X post.

- Fork-to-star ratio is about 14%, high for a skill-type repo. You have to clone it to use it, and the install method is git clone plus symlink.

### Competitors

Two separate categories: **full production platforms (apps)** and **skill packages installed into agents** . Figures come from our GitHub database repo_profile (2026-11-06) and the awesome-short-drama-skills list.

| Product | Type | Repo created | star | How it differs from shuohao-skills |
|---|---|---|---|---|
| HBAI-Ltd/Toonflow-app | Open-source platform | 2026-01-29 | 20.8k | A full app (infinite canvas + agents) that you wire to APIs yourself; shuohao is a lightweight skill that uses your agent session quota and needs no key |
| chatfire-AI/huobao-drama | Open-source platform | 2026-01-05 | 17.0k | A one-sentence-to-finished-video platform; shuohao only does pre-production and outputs a "production pack" |
| waooAI/waoowaoo | Open-source platform | 2026-01-22 | 14.8k | Same, a platform |
| HKUDS/ViMax | Open source (academic) | 2025-03-30 | 12.9k | A multi-agent video generation framework from HKU, research-oriented |
| dramaclaw, Jellyfish, ArcReel, printfilm | Open-source platform | 2026-03~09 | 6–7k | Platform / workbench format |
| zenstory-ai/drama-skills | Open-source skills | 2026-07-16 | 3.8k | The most direct rival, three weeks earlier. 11–12 skills, covering everything through review and editing. In blind tests its scripts won 3 out of 3 against the control group, writing better than shuohao, yet it has 37% fewer stars |
| dexhunter/seedance2-skill | Open-source skills | 2026-02-11 | 4.4k | Only the Seedance video prompt step |
| jtydhr88/screenwriting-skills | Open-source skills | 2026-09-06 | 2.0k | Scriptwriting only |
| lixiaoxiao9888-create/manju-laoli-skill, short-drama-factory | Open-source skills | 2026-08-15 / 08-09 | 1.9k / 0.5k | Comic-drama / short-drama skills, launched a few days after it |
| UnderleveledBuilder/awesome-short-drama-skills | List | 2026-10-13 | 1.8k | 83 short-drama skills and 16 platforms are listed, so this category is already crowded |
| big tech | Platform features | — | — | Qwen, Doubao and WorkBuddy can all install skills; ByteDance's Jimeng / Seedance and MiniMax Hailuo have their own storyboard and finishing tools (not looked into in detail here) |

**Takeaway** : platforms have far more stars than this; among "pure skill packs", shuohao-skills has the most stars (ranked first among full-pipeline entries in the list, ★5.9k).

### Where users come from

1. **The author posts on X and open-sources in serial form (confirmed)** . The author has posted 4,050 tweets on X and already had a 832-star repo, article-tools (WeChat / X cover layout tool), so they have an audience among Chinese AI readers on X. The skills were released one by one, with a post for each, ending with a hook: "Only the script and storyboard are left" → "All four skills are now complete… the storyboard will be open-sourced once I polish it a bit more!" This is open-source serialized content with weekly updates.

2. **Chinese AI-account reposts and aggregator scraping (confirmed)** . cyberkk @krienknight 08-19 wrote a long post dissecting it; servasyy_ai 08-25 ran a long-form roundup of 55 AI video production pipelines and ranked it #1 among short-drama skills (2.0k); 08-27–31 AI directory sites like 苏米客, 来拆啦, AI工具集 and AI铺子 batch-listed it; the Tencent Cloud Developer Community 09-08 published a hands-on piece, "Lose a copy of Journey to the West, and AI fully auto-produces a short drama," that opens with "saw it on X"; people also introduced it on YouTube ("Github 雷達"), Bilibili, Threads and Facebook groups; and skill directories like skills.sh, skillsmp and typingmind picked it up automatically.

3. **The "4-digit sponsorship" post drove a second wave (confirmed)** . 10-26 The author posted: "Folks, we have to go open source! shuohao-skills just received a 4-digit sponsorship!... Launched 2 months ago, 5k stars", **58.3K views** ; stars rose +273 and +266 on the following two days, the highest single-day jump we've logged since we started tracking. The list author @UnderleveledDev said in the comments it is "one of the best in this short-drama skills collection".

4. **Long tail from lists and reviews (likely)** . GitHubDaily 10-24 promoted awesome-short-drama-skills; the list placed it in a "recommended combo" (narrative-to-screen-reader → shuohao-skills → screenwriter-skill), ranked first in the full-pipeline group.

5. No evidence found of paid acquisition, account farms, or star-buying.

### Why now

- **Video models crossed the usable threshold** : the author's posts and the Tencent Cloud test both mention MiniMax H3 (up to 15 seconds at a time, runs locally) and Seedance 2.0. From the Tencent Cloud article: "Recently MiniMax H3 can run locally, and many people feel the barrier to AI short dramas has collapsed." Generation is no longer the bottleneck; people are stuck on "how to lock down characters and scenes up front."

- **Chinese AI short dramas / comic dramas are hot** : in 2026, GitHub was crowded with short-drama platforms (Toonflow and Huobao, both from January, each with tens of thousands of stars), and there are many Hongguo-related repos; servasyy's summary is that "short drama is the most competitive scene."

- **SKILL.md became a common format** : Claude Code, Codex, Qwen, Doubao and WorkBuddy can all install it, so you write one file that works everywhere.

### Differentiation

1. **Build the whole chain, so every step feeds the next** : novel → outline → characters → art → script → blocking → storyboard → H3 / Seedance production pack. Each stage outputs JSON that the next one reads directly. Competitors either cover just one step (seedance2-skill, screenwriting-skills) or are a full platform.

2. **Quality gates use scripts, not model self-discipline** : each skill ships with a zero-dependency Node script that hard-codes 10–18 rules (dialogue 4.5 characters / seconds converted to duration, 2–5 seconds per shot, the first 3 beats must contain a hook), plus a selftest. This is persuasive to programmer readers and is a selling point the X posts repeatedly stress.

3. **Specifically solves consistency** : each scene has 3–5 "consistency anchors", prop state variants, scale written into prompts, and characters get a front full-body image first that all other angles reference. This hits the core pain of AI short dramas: every shot looks different.

4. **Visual HTML reports** : every step outputs an illustrated interactive report (blocking top-down view where you can drag characters), and the README is full of screenshots, which suits posting and reposting.

5. **No API key needed** : it uses the current agent session's quota and runs on Node ≥18, a very low barrier.

**Is differentiation why it took off?** Partly. Chaining the whole pipeline, adding quality gates, and shipping a good-looking report made it feel the most like a real product among a pile of single-point skills, and the easiest to share. But it isn't the best writer: in a blind test on the awesome list 10-18, zenstory's drama-skills beat the control three times out of three, while shuohao-skills was ranked **below the control** by both judges (the reason: most of it is outline-level, written only to the scene or act level). So its star lead comes mainly from **being first, serialized posts, and an author who already had readers on X** , plus the engineered, visual packaging. Script quality itself isn't the reason.

### How the money comes back (the author's conversion path)

The open-source repo is just the funnel. The README header stacks several monetization paths: job hunting / a "semi-partnership" (a few thousand RMB a month in living costs plus revenue share); Ko-fi tips; a sponsor slot (APIMart, with a referral tracking link go.apimart.ai/tw-shuohao-skills); an early README also linked a **paid AI short-drama community group** (WeChat hao_dev); and its own paid product **Reelbench Studio** , an offline, source-available AI video workbench sold as a one-time purchase at $139 / $259 / $459 (studio.reelbenchai.com). There's also a sister repo, reelbench-skills (1.2k stars). The "4-figure sponsorship" is the author's own claim, unverified.

### How many people have seen it

Word on the Chinese AI circles: X, AI nav sites, skill directories, and the Tencent Cloud community. The biggest hit is 58K views. No English-language coverage found: no big English accounts, no Hacker News, no Product Hunt, no mainstream press. The author has not disclosed any revenue (only a sponsorship in the 4-figure range). The space itself is already crowded: the list has 83 short-drama skills, and new ones keep showing up.

### Copycats

No evidence it was copied. But similar skills have kept appearing since: short-drama-factory (08-09), manju-laoli-skill (08-15,1.9k), screenwriting-skills (09-06,2.0k), xixihhhh/ai-short-drama-skill (Dramake), YvonneMovingon/short-drama-skills, leoyeai's ai-drama-prompt-factory and others, whose naming and "novel → characters → storyboard" flow are very similar.

### What others can borrow

- **Release one stage a week, with a hook at the end of each post** ("only the storyboard is left"), turning one repo into a serialized show, with each release driving a wave of traffic.

- **Make "AI output" into a checkable engineering artifact** : rules written as scripts, selftests, and JSON passed downstream, which is especially persuasive to programmer readers.

- **The README works as a landing page** : one report screenshot per step plus a mermaid flow diagram, in bilingual Chinese and English.

- **Open source as a funnel** : free skills → sponsor slots / tips / paid group → one-time-purchase source product + personal job prospects.

- Target the window where "generation models just became usable" and build the layer most missing upstream (here, pre-production setup and consistency).

### Still unclear

- The specifics of the first wave of 08-07–13: the views and repost chain of the launch post were not obtained (X post text couldn't be fully scraped); the author's X follower count was not found.

- What the fork spike around 09-04–05 corresponds to is unverified (guess: the storyboard skill release, not confirmed).

- Real usage: stars and forks don't mean anyone ran the full chain; there are 0 issues (the issue tracker may be closed or unmaintained).

- Whether the "4-figure sponsorship" and Reelbench Studio's sales are real remains unverified.

### Sources

- GitHub API: repo, user, forks timeline (2026-11-06 crawl)

- Data card /home/ljr/fanout/why/cards/gh_eternityspring__shuohao-skills.json; github repo_profile (2026-11-06)

- Author on X: [x.com/eternityspring/status/2114569043743318492](https://x.com/eternityspring/status/2114569043743318492) (10-26,58.3K views); topicdigg archive [topicdigg.com/x/eternityspring/2087862241580453925](https://topicdigg.com/x/eternityspring/2087862241580453925)

- [x.com/krienknight/status/2089921988940968337](https://x.com/krienknight/status/2089921988940968337)(08-19)

- [x.com/servasyy_ai/status/2092059961656537546](https://x.com/servasyy_ai/status/2092059961656537546)(08-25)

- [x.com/GitHub_Daily/status/2113842948945412240](https://x.com/GitHub_Daily/status/2113842948945412240)(10-24)

- [github.com/UnderleveledBuilder/awesome-short-drama-skills](https://github.com/UnderleveledBuilder/awesome-short-drama-skills) (blind test 10-18 / 10-20)

- [cloud.tencent.com/developer/article/2739539](https://cloud.tencent.com/developer/article/2739539)(09-08)

- [studio.reelbenchai.com](https://studio.reelbenchai.com) (pricing)

- Aggregators: xmsumi.com/detail/4056, laicaila.com, kldh.com, aipuzi.cn, magicnetworld.com, etc.

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