Tooluck logoTooluck
SKILL.md

science-manuscript-formatting

Agent skill fromyzlin286-wq/Deepscient

Use when a manuscript draft needs to become a submittable file — you have the content (from science_paper_plan or your own draft) and now need it in the target journal's actual layout; also use when a submission was desk-rejected for formatting (word count over, too many figures, uncited references, wrong reference style), when the same paper must be retargeted from one journal to another without rewriting, or when a submission system demands DOCX or JATS rather than PDF.

How to install science-manuscript-formatting

install science-manuscript-formatting
# install the Deepscient pack with the skills CLI
npx skills add yzlin286-wq/Deepscient

# or add only this skill to Claude Code: copy the folder

dsh-science/skills/science/science-manuscript-formatting/

# into

~/.claude/skills/science-manuscript-formatting/ (all projects)

.claude/skills/science-manuscript-formatting/ (one project)

dsh-science/skills/science/science-manuscript-formatting/SKILL.md

MITFull file
1# 排版与出件:一份源,四个产物,换刊只改一行
2
3## 为什么这一层是机械的(也正因如此才能自动化)
4
5写作里有一部分判断只有作者能做:这个 claim 成不成立、这个 gap 真的存在吗、这条审稿意见怎么回。
6这些在 M2/M4 里,靠 `science_paper_audit` 和 `science_gap_matrix`。
7
8剩下那一部分是**纯机械的**:字数有没有超、图有几张、引用的 key 在文献表里有没有、
9参考文献该用 `[1]` 还是 `(Jordan et al., 2024)`、页边距多少毫米、要不要连续行号。
10这些东西完全由**目标期刊**决定,跟内容无关——所以只要你把「刊」告诉机器,它就该自己算完。
11
12这就是 `paper.md` 的设计前提:**正文里不写任何格式**。
13
14```markdown
15---
16title: Orthogonalized Momentum Bounds the Effective Learning Rate
17journal: nature # ← 换刊只改这一行
18keywords: [muon, optimizer, orthogonalization]
19authors:
20 - name: Zhang Wei
21 affiliation: 1
22 corresponding: true
23 email: zhang@example.edu
24affiliations:
25 - id: 1
26 name: "Institute of Computing, Beijing, China"
27abstract: |
28 一句话背景 → 一句话缺口 → 一句话做法 → 一句最硬的结果(带数字)→ 一句话意义。
29references:
30 - key: jordan2024muon
31 type: article-journal
32 title: "Muon: An optimizer for hidden layers in neural networks"
33 authors:
34 - {family: Jordan, given: Keller}
35 year: 2024
36 journal: arXiv preprint
37 doi: 10.1234/example.001
38---
39
40# Introduction
41
42Momentum methods dominate large-scale training [@jordan2024muon]. ...
43
44![Effective learning rate versus step count.](figures/fig1.png){#fig:lre}
45
46$$
47\eta_{\mathrm{eff}} \le \frac{\|\mathbf{m}_t\|_2}{\sqrt{d}}
48\label{eq:bound}
49$$
50
51| Optimizer | Steps |
52| --- | --- |
53| SGD | 12000 |
54: Comparison of optimizers. {#tab:cmp}
55```
56
57`journal: nature` 改成 `journal: cell`,PDF 从 Nature 投稿版式变成 Elsevier 双栏,
58参考文献从上标数字变成 `(Jordan et al., 2024)`,**正文一个字没动**。
59
60## 一、先建稿,别从空文件开始
61…

Excerpt shown under the repository's MIT license.

More skills in Deepscient

Keep exploring

Similar skill repositories

Newsletter

Join the Community

Subscribe to our newsletter for the latest news and updates