@QingQ77: 为学术写作提供审稿式体检和 rebuttal 策略的 Codex skills suite https://github.com/M1n-n9/paper-lifecycle… Paper Lifecycle 是写论文用的两个 Codex…
摘要
Paper Lifecycle 是一套为学术写作设计的 Codex 技能套件,包含审稿式修改(Review Revision)和反驳回复(Rebuttal Response)两个技能,帮助研究者提升论文质量和审稿应对能力。
查看缓存全文
缓存时间: 2026/06/26 16:14
为学术写作提供审稿式体检和 rebuttal 策略的 Codex skills suite
https://t.co/5ieX197Yhm
Paper Lifecycle 是写论文用的两个 Codex skill。Review Revision 像审稿人一样看你的初稿:问题真不真、insight 站不站得住、novelty 有没有、方法跟机制对不对、实验够不够支撑、写作有没有把审稿人带偏。 https://t.co/WZAaBWmE99
M1n-n9/paper-lifecycle
Source: https://github.com/M1n-n9/paper-lifecycle
Paper Lifecycle Skills
中文
paper-lifecycle 现在是一个科研写作 skills suite。根目录只做项目说明,真正可安装、可调用的 skill 是两个平级目录:
review-revision/
rebuttal-response/
两个 Skill
Review Revision
路径:review-revision/
用途:对已有论文、初稿、被拒稿件或已有明确方向的 idea 做审稿式体检和修改规划。它会检查问题是否真实,insight 是否成立,novelty 是否站得住,method 是否服务于机制,experiments 是否支撑 claim,writing 是否把审稿人带到正确理解。
适合:
- 论文 PDF、LaTeX、Word 或 Markdown 初稿诊断
- A 会、期刊或 workshop 投稿前 checklist
- introduction、related work、method、experiments 等逐节修改
- theory、statistics、reproducibility、figures/tables 审计
调用示例:
Use $review-revision to audit this manuscript, plan revisions, and produce a final checklist.
Rebuttal Response
路径:rebuttal-response/
用途:专门处理审稿意见和 rebuttal。它不是泛泛润色回复,而是把 reviews 拆成 AC 能判断的证据包:哪些 concern 会影响接收,哪些 reviewer 可能被说服,优先补哪些证据,怎样写 AC-facing summary,怎样逐 reviewer 回复,怎样纠正事实错误但不显得攻击 reviewer。
适合:
- 收到 reviews 后做问题分类和优先级排序
- 写 AC-facing executive summary
- 回复 novelty、baseline、实验设置、方法动机、scope、事实误解
- 检查 rebuttal 语气,删掉过度道歉、硬刚、空泛承诺
- 生成 rebuttal checklist 和 camera-ready promise list
调用示例:
Use $rebuttal-response to triage these reviews, plan new evidence, and draft a concise rebuttal.
Codex 安装
建议把整个仓库 clone 到一个普通目录,再把两个 skill 分别链接到 Codex skills 目录。
Windows PowerShell:
git clone https://github.com/M1n-n9/paper-lifecycle.git "$env:USERPROFILE\.codex\research-skills\paper-lifecycle"
New-Item -ItemType Junction -Path "$env:USERPROFILE\.codex\skills\review-revision" -Target "$env:USERPROFILE\.codex\research-skills\paper-lifecycle\review-revision"
New-Item -ItemType Junction -Path "$env:USERPROFILE\.codex\skills\rebuttal-response" -Target "$env:USERPROFILE\.codex\research-skills\paper-lifecycle\rebuttal-response"
macOS / Linux:
git clone https://github.com/M1n-n9/paper-lifecycle.git ~/.codex/research-skills/paper-lifecycle
ln -s ~/.codex/research-skills/paper-lifecycle/review-revision ~/.codex/skills/review-revision
ln -s ~/.codex/research-skills/paper-lifecycle/rebuttal-response ~/.codex/skills/rebuttal-response
Claude / Claude Code / 其他 Agent
不依赖 Codex skill 机制时,直接让 agent 读取对应目录即可。
Review Revision:
review-revision/SKILL.md
review-revision/references/review_revision_skill.md
推荐提示词:
Read review-revision/SKILL.md and review-revision/references/review_revision_skill.md, then use the review-revision workflow to analyze this manuscript.
Rebuttal Response:
rebuttal-response/SKILL.md
rebuttal-response/references/rebuttal_playbook.md
推荐提示词:
Read rebuttal-response/SKILL.md and rebuttal-response/references/rebuttal_playbook.md, then triage these reviews and draft an evidence-based rebuttal.
目录结构
paper-lifecycle/
├── README.md
├── CLAUDE.md
├── assets/
│ └── paper_lifecycle_three_conference_results_7200x9600.png
├── review-revision/
│ ├── SKILL.md
│ ├── agents/
│ │ └── openai.yaml
│ └── references/
│ └── review_revision_skill.md
└── rebuttal-response/
├── SKILL.md
├── agents/
│ └── openai.yaml
└── references/
└── rebuttal_playbook.md
English
paper-lifecycle is now a research-writing skills suite. The repository root is documentation only. The actual installable skills are two sibling directories:
review-revision/
rebuttal-response/
Skills
Review Revision
Path: review-revision/
Purpose: reviewer-style diagnosis and revision planning for an existing manuscript, draft, rejected paper, or paper idea with a concrete direction. It checks whether the problem is real, the insight is defensible, the novelty holds, the method serves the mechanism, the experiments support the claims, and the writing guides reviewers through the argument.
Use it for:
- PDF, LaTeX, Word, or Markdown manuscript diagnosis
- A-tier, journal, or workshop submission checklists
- section-level revision for introduction, related work, method, and experiments
- theory, statistics, reproducibility, figures, and table audits
Prompt:
Use $review-revision to audit this manuscript, plan revisions, and produce a final checklist.
Rebuttal Response
Path: rebuttal-response/
Purpose: review-response and rebuttal planning. It does not merely polish replies. It turns reviews into a decision-ready evidence package for the AC: which concerns can change the decision, which reviewers are persuadable, what evidence to add, how to write an AC-facing summary, how to draft per-reviewer responses, and how to correct factual misunderstandings without sounding hostile.
Use it for:
- triaging reviews and ranking concerns
- writing an AC-facing executive summary
- responding to novelty, baseline, experimental setup, method motivation, scope, and factual-error objections
- removing excessive apology, defensiveness, vague promises, and hostile wording
- producing a rebuttal checklist and camera-ready promise list
Prompt:
Use $rebuttal-response to triage these reviews, plan new evidence, and draft a concise rebuttal.
Codex Installation
Clone the suite into a normal directory, then link each skill into the Codex skills directory.
Windows PowerShell:
git clone https://github.com/M1n-n9/paper-lifecycle.git "$env:USERPROFILE\.codex\research-skills\paper-lifecycle"
New-Item -ItemType Junction -Path "$env:USERPROFILE\.codex\skills\review-revision" -Target "$env:USERPROFILE\.codex\research-skills\paper-lifecycle\review-revision"
New-Item -ItemType Junction -Path "$env:USERPROFILE\.codex\skills\rebuttal-response" -Target "$env:USERPROFILE\.codex\research-skills\paper-lifecycle\rebuttal-response"
macOS / Linux:
git clone https://github.com/M1n-n9/paper-lifecycle.git ~/.codex/research-skills/paper-lifecycle
ln -s ~/.codex/research-skills/paper-lifecycle/review-revision ~/.codex/skills/review-revision
ln -s ~/.codex/research-skills/paper-lifecycle/rebuttal-response ~/.codex/skills/rebuttal-response
Claude / Claude Code / Other Agents
If you are not using the Codex skill loader, ask the agent to read the relevant skill directory directly.
Review Revision:
review-revision/SKILL.md
review-revision/references/review_revision_skill.md
Recommended prompt:
Read review-revision/SKILL.md and review-revision/references/review_revision_skill.md, then use the review-revision workflow to analyze this manuscript.
Rebuttal Response:
rebuttal-response/SKILL.md
rebuttal-response/references/rebuttal_playbook.md
Recommended prompt:
Read rebuttal-response/SKILL.md and rebuttal-response/references/rebuttal_playbook.md, then triage these reviews and draft an evidence-based rebuttal.
Repository Structure
paper-lifecycle/
├── README.md
├── CLAUDE.md
├── assets/
│ └── paper_lifecycle_three_conference_results_7200x9600.png
├── review-revision/
│ ├── SKILL.md
│ ├── agents/
│ │ └── openai.yaml
│ └── references/
│ └── review_revision_skill.md
└── rebuttal-response/
├── SKILL.md
├── agents/
│ └── openai.yaml
└── references/
└── rebuttal_playbook.md
相似文章
@yunxi0623: https://x.com/yunxi0623/status/2069054269332889793
介绍5个Codex Skills用于科研提效,包括论文框架搭建、图片转PPT、科研图示编辑、学术写作辅助以及高水平论文结构学习,强调将重复流程固定为可复用技能。
@wsl8297: 用 AI 写论文最怕的不是写不出来,而是它看起来很完整,实际上研究空白、文献支撑、论证结构都站不住。Academic Paper Skills 解决的就是这个问题。 GitHub:https://github.com/lishix520/…
Academic Paper Skills 是一套用于 Claude Code 的论文写作技能框架,将写作流程分为策略和写作两个阶段,并内置文献支撑、审稿人模拟和质量检查等特性,帮助用户从研究想法生成初稿。
@ModengSir: 马上就到了毕业季 各位大学生/研究生 赶紧看过来 写论文有帮手了 把现在最火的 Codex 真正接进学术研究流程, 这个仓库很值得看。 academic-research-skills-codex 把一套研究技能 打包成单一 skill,…
A GitHub repository packages academic research skills into a single Codex skill, covering literature review, paper writing, peer review, and experiment planning to streamline the research workflow using OpenAI's Codex.
@geekbb: 覆盖 200+ 主流期刊的 Claude Code/Codex 期刊论文写作技能包,从选题到审稿回复全流程都有。 https://github.com/brycewang-stanford/Awesome-Journal-Skills…
一份覆盖200+主流期刊的Claude Code/Codex期刊论文写作技能包,包含从选题到审稿回复的全流程,开源在GitHub上。
@nini_incrypto_: 学术圈和审稿人据理力争的神器来了! 第一次投顶会不知道怎么写 Rebuttal 的看过来! 这个叫 Awesome Rebuttal 的开源项目,是专为 AI、NLP、机器人等方向论文打造的反驳辅助外挂。它能直接化身你的学术顾问,帮你优雅…
Awesome Rebuttal 是一个开源项目,专为 AI、NLP、机器人等方向论文打造,可帮助作者分析审稿意见、规划补充实验并生成有理有据的回复,支持嵌入 Cursor 或 Claude Code。