Research workflow Skill
科研工作流 Claude 技能:覆盖选题、文献、方法、实验、写作、自审、Rebuttal、投稿九大阶段,支持 mentor / reviewer / mixed 三种模式(触发词含 reviewer mode、审查模式、mixed mode、混合模式),充当贴身导师与挑刺审稿人。
Install
git clone https://github.com/HaoYangJin/Research-workflow-SkillInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →Research Workflow Skill
一个面向学术研究全生命周期的开源 Skill:让 Claude 像资深导师一样提出问题、给出路径,也能像严格同行评审一样审稿、挑错、逼近薄弱环节。
This is an open-source research workflow Skill for the full academic lifecycle. It helps Claude act as a senior research mentor and a critical peer reviewer.
本 README 合并并扩展了仓库原始愿景:覆盖科研环境搭建、课题初始化、文献调研、技术方案设计、实验与数据处理、论文写作、投稿与修回等完整科研流程。
Table of Contents
中文说明
使用图示
下面两张图帮助你快速理解 research-mentor 的安装路径、模式选择和一次典型输入到输出的过程。
适用场景
使用 research-mentor 帮你完成:
- 选题、缩小研究范围、制定文献调研计划。
- 形成研究问题、假设、贡献点和可验证命题。
- 设计方法、实验、数据流程、复现方案和消融实验。
- 起草或审阅摘要、引言、相关工作、方法、结果、讨论。
- 进行自审、模拟同行评议、修改论文和回应 reviewer comments。
- 选择投稿 venue、制定投稿节奏和风险预案。
快速开始
- 确认仓库根目录包含
SKILL.md。 - 将整个仓库文件夹作为一个 Skill 文件夹使用,或按 USAGE.md 生成带
research-mentor/顶层目录的 zip 后上传到 Claude.ai。 - 在对话中直接说明阶段和模式,例如:
使用 research-mentor,mentor 模式。我的方向是基于图神经网络的配电网负荷预测,请帮我做选题收敛和文献调研计划。
更多安装、调用、模式切换和定制细节见 USAGE.md。
How It Works: 9 个科研阶段
- Topic scoping & literature review / 选题与文献调研:从兴趣域收敛到可研究问题,建立文献地图。
- Research question & hypothesis formulation / 问题与假设:把动机转化为可检验研究问题和假设。
- Methodology and experimental design / 方法与实验设计:选择方法、基线、实验矩阵和评价指标。
- Data collection, analysis, and reproducibility / 数据与复现性:检查数据来源、处理链路、可复现性和泄漏风险。
- Results interpretation and ablation thinking / 结果分析与消融:解释结果、设计消融、识别过度结论。
- Paper drafting / 论文写作:分章节打磨 abstract、intro、related work、methods、results、discussion。
- Self-review and simulated peer review / 自审与模拟同行评议:用 reviewer 标准提前发现拒稿风险。
- Revision and rebuttal / 修改与 rebuttal:制定修改计划并回应审稿意见。
- Submission strategy and venue selection / 投稿策略:匹配期刊/会议、范围、贡献、周期和风险。
每个阶段都有对应的 references/ 文件,包含 checklist、mentor prompts 和 reviewer red flags。
示例
输入:
导师模式。我的研究方向是 ML for energy systems,想做配电网短期负荷预测,但不知道如何把题目收敛到能投稿的研究问题。
预期输出:
- 先追问数据粒度、预测时间尺度、可用 baseline、目标 venue。
- 给出 3 个可行选题方向,并指出每个方向的 novelty 风险。
- 建议优先阅读综述、经典 baseline、最新 GNN/Transformer 负荷预测论文和能源系统约束建模论文。
- 给出下一步一周文献调研清单。
English
Use Cases
Use research-mentor when you want Claude to help you:
- Scope a research topic and plan a literature review.
- Form research questions, hypotheses, and claims.
- Design methods, experiments, baselines, metrics, and ablations.
- Audit data collection, analysis, and reproducibility.
- Draft or review academic paper sections.
- Run a simulated peer review before submission.
- Revise a paper and write rebuttals to reviewers.
- Choose a submission venue and publication strategy.
Quick Start
- Make sure the repository root contains
SKILL.md. - Use the full folder as a Skill folder, or create a zip with a top-level
research-mentor/directory as shown in USAGE.md. - Ask for a phase and a mode:
Use research-mentor in reviewer mode. Please critique this methodology section and identify the strongest rejection risks.
See USAGE.md for installation, invocation, mode switching, and customization details.
How It Works: 9 Research Phases
Similar plugs
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local…
Why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman.
Production-grade engineering skills for AI coding agents.
The 1st Collaborative AI Design Agent Workspace. Local-first and open source. Create prototypes, websites, apps, dashboards, presentations, images, and videos with your team…