paper-reading-zh skill
A skill for reading research papers in Chinese: it first states what was read, then ties every number to a table or figure.
- 110 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill only sets reading rules and takes no actions outside the conversation.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/paper-reading-zhDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill only sets reading rules and takes no actions outside the conversation.
This is third-party code. Review the repository files before installing.
What it does
paper-reading-zh is a rule set for Codex and Claude Code that makes an agent read papers without guessing. Input is a PDF, a link or an abstract. The agent first says which pages and materials it actually saw, then breaks the work down by type: problem, method, evidence, limitations. Experimental numbers come with a table, figure or section reference. Venues and code links without external verification are marked "未核验", and an answer based on the abstract alone says "仅基于摘要". The skill is four text files in `paper-reading-zh/` with no scripts. The rules can also be pasted into a Claude or ChatGPT project.
Who it is for. Graduate students and researchers who read papers in Chinese and want a source for every number.
Good fit when
- A seminar needs a fast breakdown of a paper with table references.
- The agent got only an abstract and the answer must say so.
- A literature review has to separate a claimed result from a supported one.
- You check whether the code link and venue are actually verified.
Not a fit when
- You need to search for papers on a topic: the skill only reads material it is given.
- You need a Russian or English reading: the rules are written for answers in Chinese.
- You need automatic processing of hundreds of papers without a person.
Example request
Read this PDF with the skill rules: say what you saw and give the table for every number.Limitations
The skill does not search for papers or check metadata online. The "未核验" mark only means no external verification ran. Answers are in Chinese; other languages need the rules rewritten. The authors show a comparison on one paper and note that the no-skill version also produced a good answer: the gain is that every number carries a visible source. The skill is added with `npx skills add`, which asks which tools to install into.
How to disable. Remove the `paper-reading-zh` folder from the Codex or Claude Code skills directory it was installed into.
FAQ
How do I install it?
With `npx skills add MrGeDiao/paper-reading-zh`. It installs only the `paper-reading-zh/` folder. By default the skill applies to the current project; with `--global` it works everywhere.
Can I install it by hand?
Yes. Copy the folder into `~/.codex/skills/` for Codex or `~/.claude/skills/` for Claude Code.
What does "仅基于摘要" mean?
The agent received only an abstract and states plainly that the answer does not rest on the full text.
Does it work outside Codex and Claude Code?
Yes, the rules can be pasted into a Claude project or a ChatGPT project as instructions.
Which license?
MIT, free to use and modify.
Discussion
No questions or reviews yet.
Related
SkillsTwo skills for Claude Code and Codex: the agent translates papers in the EasyRead library and answers your margin notes
paper-reading-skill turns one arXiv AI or robotics paper into a report with each claim tied to evidence in the source.
A set of three research skills: paper writing, academic Word and PowerPoint documents, scientific computing and figures
scientific-writing-zh is a Claude Code skill, a Chinese-language writing coach for papers, revision, reviews and talks.
A skill for deep-reading one scientific paper: six sections, key figures in the text and each claim checked against the source.
A skill for researchers who draft in Chinese: drafts and results become an English manuscript, with citations and claims checked
README in short
The README compares the same question about one paper with and without the skill. The difference shows in number sources.