Token导航 LogoToken导航

Agent Skill · 研究检索

scientific-writing(科学写作)

用于辅助文档、README、Markdown、说明文和内容稿件的整理与改写。它适合让 Agent 提炼结构、补齐章节、统一术语、检查链接或把零散材料整理成可读文档。使用时应保留项目已有事实、命令和路径,不要把未确认的信息写成确定结论;涉及对外文案时,还需要控制语气,避免过度营销或夸大能力。

scientific-writing 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 61,604。

总安装 349周安装 15GitHub Stars 3下载 122核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:scientific-writing(科学写作)
来源仓库:https://github.com/braselog/researchassistant
仓库路径:skills/scientific-writing
安装命令:
npx skills add https://github.com/braselog/researchassistant --skill scientific-writing
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/braselog/researchassistant --skill scientific-writing

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于撰写符合出版标准的科学论文,支持 IMRAD 结构与引用格式化。适用于摘要、方法、图表制作及报告规范(如 CONSORT)的应用指导。可协助准备期刊投稿材料,提升稿件接受率。需用户提供原始数据与研究背景,避免凭空补充实验结果。

scientific-writing 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/scientific-writing

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/braselog/researchassistant --skill scientific-…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

scientific-writing 怎么安装?

在终端执行 npx skills add https://github.com/braselog/researchassistant --skill scientific-writing。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

scientific-writing 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

scientific-writing 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

scientific-writing 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

scientific-writing 用的人多吗?

公开数据是累计安装 349 次,近一周安装 15 次,下载 122 次,GitHub 3 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

scientific-writing 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
scientific-writing

科学写作 · 本页

3493需要联网未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Scientific Writing

Write clear, precise, and publication-ready scientific manuscripts.

When to Use

  • Drafting manuscript sections (abstract, intro, methods, results, discussion)
  • Structuring a research paper using IMRAD format
  • Formatting citations and references
  • Creating or improving figures and tables
  • Applying reporting guidelines (CONSORT, STROBE, PRISMA)
  • Preparing manuscripts for journal submission
  • During the WRITING or REVIEW phases

Manuscript Structure (IMRAD)

┌─────────────────────────────────────────────────────────────┐
│ TITLE                                                       │
│ Concise, specific, informative (12-15 words)               │
├─────────────────────────────────────────────────────────────┤
│ ABSTRACT (150-250 words)                                    │
│ Background → Objective → Methods → Results → Conclusion     │
├─────────────────────────────────────────────────────────────┤
│ INTRODUCTION                                                │
│ Context → Gap → Objective → Approach                        │
│ Funnel: Broad → Narrow → Your question                     │
├─────────────────────────────────────────────────────────────┤
│ METHODS                                                     │
│ Study design → Participants → Procedures → Analysis         │
│ Enough detail for replication                              │
├─────────────────────────────────────────────────────────────┤
│ RESULTS                                                     │
│ Objective findings, no interpretation                       │
│ Text + Figures + Tables                                     │
├─────────────────────────────────────────────────────────────┤
│ DISCUSSION                                                  │
│ Key findings → Context → Limitations → Implications         │
│ Reverse funnel: Specific → Broad                           │
├─────────────────────────────────────────────────────────────┤
│ REFERENCES                                                  │
│ Consistent style, verified DOIs                            │
└─────────────────────────────────────────────────────────────┘

Section-by-Section Guidance

Abstract

Purpose: Standalone summary of the entire paper

Structure (for structured abstracts):

  • Background: Why this matters (1-2 sentences)
  • Objective: What you did (1 sentence)
  • Methods: How you did it (2-3 sentences)
  • Results: Key findings with numbers (3-4 sentences)
  • Conclusion: Main takeaway (1-2 sentences)

Tips:

  • Write LAST after all other sections
  • Include specific numbers/results
  • Avoid abbreviations (or define them)
  • Stay within word limit (usually 150-250)

Introduction

Purpose: Establish context, gap, and rationale

Structure (Funnel):

  1. Broad context (1-2 paragraphs): Why does this field matter?
  2. Current knowledge (2-3 paragraphs): What's known? What approaches exist?
  3. Gap/Problem (1 paragraph): What's missing? What's the controversy?
  4. Your study (1 paragraph): What did you do? Why this approach?

Tips:

  • End with clear objectives or hypotheses
  • Cite 20-40 references typically
  • Use present tense for established facts
  • Be specific about what you're studying

Methods

Purpose: Enable replication

Key Sections:

  1. Study Design: Type of study, setting, dates
  2. Participants/Samples: Selection, criteria, sample size, ethics
  3. Procedures: What was done, in order
  4. Measurements: What and how measured
  5. Statistical Analysis: Tests, software, significance criteria

Common Mistakes:

MistakeProblemFix
Vague methods"Standard methods"Specify exact protocol
Missing stats"Data were analyzed"Name specific tests
No software versionsNot reproducibleInclude version numbers
Missing sample size justificationWhy this n?Add power analysis

Tips:

  • Use past tense
  • Be specific: model numbers, concentrations, durations
  • Reference published protocols if applicable
  • Include ethical approvals

Results

Purpose: Present findings objectively

Organization:

  1. Order by importance or by methods flow
  2. Each paragraph: finding + evidence (figure/table reference)
  3. Stats: test, statistic, df, p-value, effect size

Structure Pattern:

[What was found] (Figure X).
[Statistical support] (t(df) = X.XX, p = .XXX, d = X.XX).
[Additional detail or subgroup analysis].

Tips:

  • NO interpretation (save for Discussion)
  • Include negative/null results
  • Reference every figure and table
  • Use past tense
  • Include exact p-values (not just p < 0.05)

Discussion

Purpose: Interpret findings in context

Structure (Reverse Funnel):

  1. Key findings (1-2 paragraphs): Main results, directly address objectives
  2. Comparison to literature (2-3 paragraphs): How do findings fit with prior work?
  3. Mechanisms (1-2 paragraphs): Why might this happen?
  4. Limitations (1 paragraph): Be honest and specific
  5. Implications (1-2 paragraphs): Clinical, practical, theoretical significance
  6. Future directions (optional): What next?
  7. Conclusion (1 paragraph): Main takeaway

Tips:

  • Start with your results, not literature
  • Acknowledge limitations honestly
  • Don't overstate conclusions
  • Distinguish correlation from causation

Citation Styles

APA (7th Edition)

In-text: (Author, Year) or Author (Year)

Previous research found significant effects (Smith, 2023).
Smith (2023) reported significant effects.

Reference list:

Smith, J. D., Johnson, M. L., & Williams, K. R. (2023). Title of
    article. Journal Name, 22(4), 301-318. https://doi.org/10.xxx/yyy

Vancouver/ICMJE

In-text: Superscript or bracketed numbers¹ or [1]

Previous research found significant effects.¹
Multiple studies support this finding.¹⁻³

Reference list (numbered):

1. Smith JD, Johnson ML, Williams KR. Title of article. J Name.
   2023;22(4):301-18.

Nature

In-text: Superscript numbers¹

Previous research found significant effects¹.

Reference list:

1. Smith, J. D., Johnson, M. L. & Williams, K. R. Title of article.
   Nat. Rev. Drug Discov. 22, 301-318 (2023).

Figures and Tables

When to Use Which

Use Tables ForUse Figures For
Exact values neededTrends and patterns
Many variablesComparisons
Summary statisticsRelationships
Participant characteristicsProcesses

Figure Checklist

  • Self-explanatory with caption
  • Axes labeled with units
  • Error bars defined (SEM, SD, CI)
  • Significance markers explained
  • Colorblind-safe
  • Resolution ≥300 DPI

Table Checklist

  • Clear, descriptive title
  • Column headers with units
  • Appropriate precision (not too many decimals)
  • Notes for abbreviations
  • n values included

Caption Template

**Figure 1. Brief descriptive title.**

(A) Description of panel A. (B) Description of panel B.
Data shown as mean ± SEM (n = X per group). Statistical
comparisons by [test name]. *p < 0.05, **p < 0.01, ***p < 0.001.

Reporting Guidelines

Which Guideline to Use

Study TypeGuidelineURL
Randomized trialCONSORTconsort-statement.org
Observational (cohort, case-control)STROBEstrobe-statement.org
Systematic reviewPRISMAprisma-statement.org
Diagnostic accuracySTARDstard-statement.org
Prediction modelsTRIPODtripod-statement.org
Animal researchARRIVEarriveguidelines.org
Case reportsCAREcare-statement.org
Quality improvementSQUIREsquire-statement.org

Using Checklists

  1. Download checklist from guideline website
  2. Complete each item during writing
  3. Include page/line numbers
  4. Submit with manuscript (often required)

Writing Principles

Clarity

  • Use precise, unambiguous language
  • One idea per sentence
  • Define technical terms at first use
  • Use active voice when possible

Example:

❌ "The samples were subjected to analysis"
✓ "We analyzed the samples using..."

❌ "It has been shown that..."
✓ "Smith et al. (2023) showed that..."

Conciseness

WordyConcise
"Due to the fact that""Because"
"In order to""To"
"A large number of""Many"
"At the present time""Now" / "Currently"
"In the event that""If"
"Has the ability to""Can"

Accuracy

  • Report exact values with appropriate precision
  • Use consistent terminology
  • Distinguish observation from interpretation
  • Acknowledge uncertainty

Objectivity

  • Present results without bias
  • Don't overstate findings
  • Acknowledge contradictory evidence
  • Maintain professional, neutral tone

Field-Specific Terminology

General Principles

  • Match terminology to the target journal
  • Use established nomenclature systems
  • Define abbreviations at first use
  • Be consistent throughout

Quick Reference

FieldConvention
GenesItalics (*BRCA1*)
ProteinsRoman (BRCA1)
SpeciesItalics, full at first (*Escherichia coli*), then abbreviated (*E. coli*)
StatisticsItalics (p, n, t, F, r)
DrugsGeneric name first, brand in parentheses

Common Mistakes to Avoid

Top Rejection Reasons

  1. Incomplete or inappropriate statistics
  2. Over-interpretation of results
  3. Poor methods description
  4. Inadequate sample size
  5. Poor writing quality
  6. Inadequate literature review
  7. Unclear figures
  8. Failure to follow guidelines

Writing Issues

IssueExampleFix
Tense mixing"We collected... and analyze"Past for methods/results
Excessive jargonToo many undefined termsDefine or simplify
Paragraph breaksRandom breaksOne topic per paragraph
Missing transitionsAbrupt section changesAdd linking sentences

Manuscript Development Workflow

Recommended Order

  1. Figures/Tables first (core data story)
  2. Methods (often easiest to draft)
  3. Results (describe figures/tables)
  4. Discussion (interpret findings)
  5. Introduction (set up the question)
  6. Abstract (synthesize everything)
  7. Title (last refinement)

Revision Checklist

  • Logical flow throughout
  • Consistent terminology
  • All figures/tables referenced
  • All citations verified
  • Word counts met
  • Reporting checklist complete
  • Journal format requirements met

Integration with RA Workflow

WRITING Phase Files

FilePurpose
manuscript/background.mdIntroduction content
manuscript/methods.mdMethods section
manuscript/results.mdResults + figure refs
manuscript/discussion.mdDiscussion section
manuscript/figures/figN/caption.mdFigure captions

Connected Skills

  • ← /write_background: Drafts introduction
  • ← /write_methods: Generates methods from scripts
  • ← /write_results: Drafts results from figures
  • → /peer_review: Self-review before submission

平台分布

Codex

32.62%
按下载量换算40

Claude

31.67%
按下载量换算39

Cursor

17.31%
按下载量换算21

Gemini CLI

9.74%
按下载量换算12

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills