# ChinaGodMan UserScripts: 85 collected scripts and a README that is built by PowerShell > This is not a tool with an API, it is a shelf of Tampermonkey scripts in one directory each, grouped into twelve categories by a generated table. The author says plainly that the scripts came from other people and that they made small edits. What is more interesting than the scripts is the machine around them: a README regenerated by a PowerShell build, a Python helper with two flake8 configs, and icons hosted in a second repository. **ChinaGodMan/UserScripts** — 🐒一些修改自网络的油猴脚本 Some Tampermonkey scripts modified from the internet - Repository: https://github.com/ChinaGodMan/UserScripts - Website: https://greasyfork.org/scripts?by=1169082 - Stars: 2,067 · Forks: 147 - Language: JavaScript - License: NOASSERTION - Published: 2026-09-14 · Updated: 2026-09-14 · Language: en - Canonical page: https://hysenlabs.com/projects/chinagodman-userscripts ## The script count and its category table are generated The README does not list the scripts by hand. It carries HTML comment markers around the count, and a matching pair around the category table, which is the tell that something regenerates them. The generated line reads that the repository currently has 85 published scripts, and the twelve categories break down as: ten web page scripts, twelve from Greasy Fork, eleven in an adult category, two Google scripts, twenty GitHub scripts, three translation scripts, nine AI scripts, five blog site scripts, five Twitter scripts, five YouTube scripts, one TikTok script and two Bilibili scripts. Those add up to the stated total, so the table and the count move together. The script that does it is in the root manifest: ``` "bump:scripts": "python utils/update_version.py", ``` So updating the version metadata of every script is a Python script run over a utils directory. The README body below the table, with each script's icon, title, one-line summary and four links, is written in the same generated style. For anyone who wants to count or audit what is here, the generated table is the cheapest entry point, and the HTML comment markers tell you which parts of the document a machine owns. ## The README is a build product, and the build runs under PowerShell The build script in the root manifest is: ``` "build": "pwsh -noprofile -command buildscript" ``` PowerShell, with profiles disabled, running a command named buildscript. So the document you read on the repository page is assembled rather than maintained, and producing it on Linux or macOS means having PowerShell available. The output is multilingual. The language bar at the top of the README links twelve versions: Simplified Chinese at the root, then Traditional Chinese, English, Japanese, Vietnamese, Italian, Russian, German, Spanish, Portuguese, French and a Korean file. The Korean label reads 대한민국, which is the name of the country rather than the language, so that one link is labelled slightly differently from the rest. The document also holds a table of contents with markers, listing sections for browsers and script managers, how to install or use the scripts, acknowledgements, a security policy, a code of conduct, the licence, contributors, a contribution guide, star history and the script list. Linting closes the loop. The dev dependencies include an ESLint markdown plugin and a JSON plugin alongside the Prettier config, so generated markdown is still held to a formatting standard. ## One Python helper, two flake8 configurations The Python tooling is small and it configures itself twice. There is a .flake8 file at the repository root, and there is a pyproject.toml with its own flake8 section: ```toml [tool.isort] multi_line_output = 3 [tool.flake8] exclude = [".git", "__pycache__", "build", "node_modules"] ignore = ["E501", "W504"] ``` And the manifest points flake8 at the second one explicitly: ``` "flake8": "flake8 --config=pyproject.toml", ``` So the root file is inert unless something invokes flake8 without that flag. Two copies of one config is a small thing, but it is the kind of duplication that makes a lint failure or a lint silence hard to explain. The rest of the Python side follows the same tidy pattern: isort over the utils directory, an autopep8 diff mode and an in-place mode, both invoked with python -X utf8, which is what lets the scripts process non-ASCII paths and text on Windows. The commit side is Chinese-conventional-commits rather than plain commitlint, with cz-git and its runner czg in the dev dependencies, next to a commitlint config and husky hooks running lint-staged. ## Every script is a directory with four links and a borrowed icon One script, one directory, and inside each the same set of files. Looking at the first entry, a script that puts two scroll buttons on the right of a page, the directory holds the readme, the source file ending in .user.js, and links to four places: the Greasy Fork script page, the readme, the GitHub discussions page for the project, and the source. Two of those four are the same file by different routes. The source link points into the tree at the .user.js file, and the install link points at the raw version of that same file, so the difference between reading and installing is a URL. The icon is the part that surprises. Every script's icon is loaded from a different repository, a raw URL pointing at a scriptsIcon folder in a history repository. That keeps binaries out of the source tree, and it also means a page's icons are one repository's availability away from rendering. The top-level listing shows the shape of the collection at a glance: more than fifty script directories named for their target, including a dozen for GitHub, a set for ChatGPT, several for Greasy Fork itself, and a handful for Bilibili, Google, TikTok and Twitter. Utility scripts sit alongside the site-specific ones, a scroll button, a clipboard manager that blocks unauthorised clipboard writes and disables media preload, a table of contents cleaner for blogs and a file list beautifier. ## Two sections announce a list and then show none The README has a recommended-use section and a tip, and in both the promise and the list are separated. The tip says the script managers and browsers already listed above do not represent all of them, and that the scripts run normally in the environments listed below. Then the environment list is empty. The recommendation block above it opens with a bold heading for recommended use and contains nothing after it either, other than the dark-mode icon image above. The table of contents promises a section on browsers and script managers, and the individual entries do carry platform names: the scroll button entry links to Greasy Fork under one engine name, and the icon row above it references browser extensions on a third-party icon host. So the compatibility information exists in fragments, in per-script links and in icon references, but the two places designed to collect it are blank. That is the single most useful thing missing from this repository, and it is a documentation fix rather than a code one. ## The licence is described three ways and the provenance is stated up front The repository's licence field in the metadata reads as not asserted, while the root manifest declares MIT and a LICENSE.md file sits in the tree, and the README has its own licence section. Any tool reading the API rather than the files gets nothing. The provenance statement is the more important thing and the author puts it near the top: the scripts in this repository are for study and exchange only and should not be used commercially, the author accepts no responsibility, the scripts come from the wider internet community with the author having made only small changes, and if there is an infringement they ask to be contacted so it can be removed. The commit history and funding follow the same personal shape. The manifest lists a single author with a name, a Gmail address, a time zone, and links to TikTok, Telegram, Instagram and a Greasy Fork profile, with funding through GitHub Sponsors and Ko-fi. There are separate CONTRIBUTING and CONTRIBUTORS files, a code of conduct and a security policy in the table of contents, and a deepwiki link for generated documentation. One category needs naming without elaborating: eleven of the 85 scripts are grouped as adult content. They are counted and linked in the same table as everything else, with no separate policy or warning attached to that category. ## Conclusion Use this repository as a catalogue rather than as a dependency. Each script is a single .user.js file with its own readme, its own Greasy Fork page and its own discussion, so adopting one means adopting one script and auditing it yourself, which matters because the author's own position is that the code arrived from elsewhere. Do not treat the repository as maintained software: there are no releases, the metadata licence field says nothing while the manifest says MIT, and the README is regenerated by a build script that needs PowerShell. Before you install anything, read the individual script's own readme, check the category you are pulling from, and keep the browser permissions it requests in view, since a userscript can rewrite the page and read the clipboard. ## FAQ ### How many userscripts are in the ChinaGodMan UserScripts repository? The README states 85 published scripts across twelve categories, with twenty for GitHub, twelve from Greasy Fork, ten web page scripts, nine AI scripts, five each for blog sites, Twitter and YouTube, three translation scripts, two each for Google and Bilibili, one for TikTok and eleven in an adult category. ### How do I install one script from this repository? Each script is its own directory with a readme, a .user.js source file and four links: the Greasy Fork page, the readme, the GitHub discussions and the source, plus an install link to the raw version of the same file. A script manager such as Greasemonkey or Tampermonkey is what loads them. ### Are the scripts in this repository original? The author says the scripts come from the wider internet community with only small changes made by them, that they are shared for study and exchange and not for commercial use, and that anything infringing will be removed on contact. The repository description says the same thing in Chinese and English. ### What licence does ChinaGodMan UserScripts use? The root package.json declares MIT and a LICENSE.md file is in the tree, while the repository's own licence metadata field reads as not asserted. The README also has its own licence section. ### How is the README of UserScripts maintained? It is generated. The root manifest builds it with a PowerShell command, and the script count and category table sit inside HTML comment markers that mark the generated regions. Script version metadata is bumped by a Python script run over the utils directory. ## Sources - [ChinaGodMan/UserScripts on GitHub](https://github.com/ChinaGodMan/UserScripts) - [Issues](https://github.com/ChinaGodMan/UserScripts/issues) - [Project website](https://greasyfork.org/scripts?by=1169082) - [README](https://github.com/ChinaGodMan/UserScripts/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/chinagodman-userscripts