scientific-research-plugin

v0.5.0

A full-lifecycle research pipeline for OR & ML, 10 skills + 1 subagent take a project from prior-art survey to the conference talk — Zotero literature intake, structured reading notes, publication-grade figures, LaTeX polishing, noun-term jargon auditing, citation verification against official APIs, three-blind adversarial review, and point-by-point rebuttal. The survey-first discipline applies to any non-trivial task, not just papers.

CodexClaude Code10 Skills

By pengkangzhenLicense: MIT6 GitHub starsUpdated 5 hours ago

Directory evidence

Runtimes: Codex and Claude Code
Parsed components: 10 skill or MCP entries
Source updated: Sep 24, 2026
Manifest status: Canonical path parsed

The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →

Install scientific-research-plugin for Codex and Claude Code

Installs for the current user

codex plugin marketplace add IchenDEV/agent-plugin-mkt
codex plugin marketplace upgrade agent-plugin-marketplace
codex plugin add scientific-research-plugin@agent-plugin-marketplace

Paste and run these commands in a terminal with Codex. They add and refresh the PluginsMP catalog, then install this plugin.

The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.

Get the source manually

$ git clone https://github.com/pengkangzhen/scientific-research-plugin

Clone the source repository, then follow its setup instructions to add the plugin to a compatible client. The repository root is the plugin root.

Plugin files

scientific-research-plugin/

├── .codex-plugin/plugin.json
├── .claude-plugin/plugin.json
├── skills/academic-ppt/SKILL.md
├── skills/figure-plot/SKILL.md
├── skills/paper-polish/SKILL.md
├── skills/paper-review/SKILL.md
├── skills/rebuttal/SKILL.md
├── skills/reference-verify/SKILL.md
├── skills/research-before-build/SKILL.md
├── skills/term-audit/SKILL.md
├── skills/zotero-paper-fetch/SKILL.md
└── skills/zotero-paper-note/SKILL.md

Included Skills10

academic-pptskills/academic-ppt/SKILL.md

学术会议演讲 PPT/slides 制作 — 从论文稿件（LaTeX/PDF）到可现场演讲的完整流程： LaTeX Beamer 路线 + 视觉设计系统（有官方模板则提取背景复用，无模板则自建色板/版式）+ 数学公式与论文原图复用 + 讲稿撰写与时长控制 + PDF→图片→官方模板交付包装 + 演讲者备注与现场预案。 IMHFC 2026 会议实战多轮迭代沉淀。触发词："会议PPT"、"演讲PPT"、"做幻灯片"、"把论文做成PPT"、"presentation"、"slides"、 "Beamer"、"会议模板"、"讲稿"、"演讲稿"、"presentation speech"、"演讲者备注"、"贴回模板"。 Use this skill whenever the user wants to turn a paper/manuscript into conference or seminar presentation slides, build a Beamer deck, fit slides into the official template, write a time-budgeted talk script, or package a slide PDF back into a pptx with speaker notes — even if they never say "PPT" (e.g. "下周要用这篇论文做口头报告", "15 分钟的 talk 怎么准备").

figure-plotskills/figure-plot/SKILL.md

数据可视化与绘图技能。用户要求画图、绘图、出图、作图、重画，或提到 figure、plot、matplotlib、数据图、示意图、算法流程图、框架图、拓扑图、柱状图、折线图、热力图、帕累托前沿、网络图、收敛曲线、中文图、学位论文图、海运航线、地理网络、世界地图、避陆，或问 drawio/math/数学公式/数学符号/LaTeX 渲染等任何图形与公式渲染问题时使用。覆盖数据图（Python/matplotlib）与示意图/流程图（drawio）的完整规范：图契约、Times New Roman 字体（中文宋体回退）、色盲安全配色、图例强制审计（序列缺图例导出即报错）、矢量 PDF 导出与字体嵌入验证、脚本落盘与迭代约定、海运航线 searoute 避陆生成与穿陆检测。

paper-polishskills/paper-polish/SKILL.md

语言润色 (language polishing) — Academic English paper polishing for LaTeX manuscripts. Polishes grammar, word choice, sentence structure, logic flow, and academic tone while preserving all LaTeX markup; does NOT judge scientific content (use paper-review for that). 触发词："润色"、"语言润色"、"论文润色"、"改英语"、"帮我改这段的英文"、"检查语言质量"。 Use this skill whenever the user asks to polish, refine, improve, or proofread a paper, manuscript, or LaTeX file, or mentions language editing, writing quality, or English improvement for academic writing. Also trigger on phrases like "check my English", "improve the writing", "make it more academic", "fix the language", "polish this section".

paper-reviewskills/paper-review/SKILL.md

科学审稿 (scientific review) — three-blind adversarial review of the SCIENTIFIC merit of papers and experimental results in any technical domain: 3 isolated reviewers (methodology rigor / domain contribution / adversarial attack) + author-defense arbitration + cross-review synthesis, with claim-evidence anchoring and journal-profile axes. Domain-conditional gates are detected per manuscript and composed freely (built-in: mathematical programming, stochastic/robust, decomposition, network/combinatorial, application domains, ML+OR, LLM/agents; extensible). NOT prose quality (use paper-polish). 触发词："科学审稿"、"学术审稿"、"审一下这个模型/结果"、"这个结果合理吗"、"投稿前自审"、"review these results"、"pre-submission review"、"simulate review"。

rebuttalskills/rebuttal/SKILL.md

审稿回复 (rebuttal) — Point-by-point rebuttal agent for academic manuscript revision: locate each pending reviewer comment in the response letter, propose manuscript changes (ADD/MODIFY/DELETE) for user confirmation, apply \changed{}/\deleted{} markup, compile both PDFs, and update the response letter with professional tone templates. 触发词："回复审稿意见"、"逐条回应审稿人"、"rebuttal"、 "response letter"、"审稿修改"、"改稿回应审稿人"。Use when the user receives reviewer comments and needs to revise the manuscript and write the response letter.

reference-verifyskills/reference-verify/SKILL.md

论文参考文献真实性与书目准确性核查（投稿前体检）。当用户要求「检查参考文献」「核查引用是否真实存在」「这些引用是不是编造的/幻觉的」「文献体检」「citation check」「verify references」，或提到预印本要不要升级为正式版、arXiv 链接要不要换官方链接时使用。输入支持 LaTeX 手稿（外部 .bib 或内嵌 thebibliography）与纯引用清单；三层方法：官方 API 机核（CrossRef / arXiv / OpenAlex / PMLR / OpenReview / ACL Anthology / NeurIPS，命令取事实、零模型回忆）→ 机核未决条目联网核查（证据必须带可访问 URL）→ 存疑结论独立复核；输出字段级核对表、严重度分级与预印本升级建议。学科不限。

research-before-buildskills/research-before-build/SKILL.md

先调研后构建。在开始任何非琐碎的技术、调研或实现任务之前，先检索人类先例与社区方案（官方文档、成熟开源库、GitHub issues、高票问答、论文、内部知识库）。分级触发、按信任采信、以决策影响为验收。琐碎任务（改错别字、格式化、用户已明确指定做法的直接执行）不适用。

term-auditskills/term-audit/SKILL.md

名词黑话批量审计（term audit）——两段式漏斗审一篇 LaTeX 手稿的名词术语：确定性脚本提取候选名词短语并按五信号排序，分批派 jargon-check terms 模式并行审计，归一化变体分组 + 合并复查术语漂移。当用户说"审一下术语""检查黑话/名词""术语要统一""投稿前术语核查"，或要把 jargon-check 跑在全篇论文上（而不只是一段话）时使用。黑话病理单位是高度凝练的多词复合名词（controller-level execution 类），本技能就是为它们设计的批量漏斗。

zotero-paper-fetchskills/zotero-paper-fetch/SKILL.md

批量检索文献、下载 PDF 并入库 Zotero 的完整管线。当用户提供文献/引用列表（Markdown 编号列表、参考文献节选、DOI 清单、arXiv ID 等），要求"搜索下载入库 Zotero""把这些文献加到 Zotero""下载 PDF 并归类""下载 arXiv 预印本"时使用。覆盖：CrossRef / arXiv / OpenAlex 检索补全元数据 → Zotero 入库归类打标 → 按出版商分层下载 PDF（OA 直链 / 仓库反爬 / 校园 VPN 付费墙）→ 挂载或落盘报告。适用于运筹学、供应链韧性、物流网络、风险建模等领域。

zotero-paper-noteskills/zotero-paper-note/SKILL.md

读取 Zotero 中的文献并生成结构化阅读笔记，把笔记添加回 Zotero 条目下，同时追加一行结构化 JSON 记录到 literature.jsonl 供后续批量分析。当用户提供论文标题/作者/DOI/Zotero 条目，要求"读这篇文献""给这篇文献做笔记""总结这篇 paper 并加到 Zotero""帮我做文献阅读笔记"时使用。适用于运筹学、供应链韧性、物流网络、选址优化、风险建模等领域。务必在用户提到 Zotero + 文献阅读/做笔记的组合，或要求把笔记写回 Zotero 时触发本技能。

Plugin manifests2

.codex-plugin/plugin.json

{
  "name": "scientific-research-plugin",
  "version": "0.5.0",
  "description": "A full-lifecycle research pipeline for OR & ML, 10 skills + 1 subagent take a project from prior-art survey to the conference talk — Zotero literature intake, structured reading notes, publication-grade figures, LaTeX polishing, noun-term jargon auditing, citation verification against official APIs, three-blind adversarial review, and point-by-point rebuttal. The survey-first discipline applies to any non-trivial task, not just papers.",
  "author": {
    "name": "pengkangzhen",
    "url": "https://github.com/pengkangzhen"
  },
  "homepage": "https://github.com/pengkangzhen/scientific-research-plugin#readme",
  "repository": "https://github.com/pengkangzhen/scientific-research-plugin",
  "license": "MIT",
  "keywords": [
    "academic",
    "research",
    "zotero",
    "latex",
    "peer-review",
    "references",
    "rebuttal",
    "figures",
    "presentation",
    "slides"
  ],
  "skills": "./skills/",
  "interface": {
    "displayName": "Scientific Research Plugin",
    "shortDescription": "Every task is research: survey, fetch, read, plot, polish, review, rebut, present",
    "developerName": "pengkangzhen",
    "category": "Research"
  }
}

.claude-plugin/plugin.json

{
  "name": "scientific-research-plugin",
  "version": "0.5.0",
  "description": "A full-lifecycle research pipeline for OR & ML, 10 skills + 1 subagent take a project from prior-art survey to the conference talk — Zotero literature intake, structured reading notes, publication-grade figures, LaTeX polishing, noun-term jargon auditing, citation verification against official APIs, three-blind adversarial review, and point-by-point rebuttal. The survey-first discipline applies to any non-trivial task, not just papers.",
  "author": {
    "name": "pengkangzhen",
    "url": "https://github.com/pengkangzhen"
  },
  "homepage": "https://github.com/pengkangzhen/scientific-research-plugin#readme",
  "repository": "https://github.com/pengkangzhen/scientific-research-plugin",
  "license": "MIT",
  "keywords": [
    "academic",
    "research",
    "zotero",
    "latex",
    "peer-review",
    "references",
    "rebuttal",
    "figures",
    "presentation",
    "slides"
  ],
  "skills": "./skills/"
}