ScienceClaw Deep Research skill
With Deep Research, the agent gathers scientific papers and data, compares sources and prepares a PDF or DOCX report.
- 678 ★ on the repository
- Last commit
Extended access
Why this level
- The agent searches external services, downloads papers and writes reports; the selected model receives research context.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/scienceclaw-deep-researchDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The agent searches external services, downloads papers and writes reports; the selected model receives research context.
This is third-party code. Review the repository files before installing.
What it does
ScienceClaw Deep Research defines a multi-source workflow inside ScienceClaw. The agent breaks a question into research dimensions, selects papers and scientific databases, stores findings and prepares source summaries. It then compares sources around the user’s question and writes a report. PDF is the default output, with DOCX on request. Plans, extracted text, summaries and report sections remain in the session workspace for inspection.
Who it is for. Researchers and educators preparing topic reviews from papers and structured scientific data.
Good fit when
- A review compares several scientific approaches and results.
- The question requires both literature and a specialist database.
- You need a saved search plan before report writing.
- A client needs a sourced research report as PDF or Word.
Not a fit when
- A short answer from one known document is enough.
- The agent lacks ToolUniverse and literature-search tools.
- An unchecked output would determine treatment or another consequential decision.
Example request
Compare data-center cooling methods in ScienceClaw: show the source plan first, then prepare a report with verifiable data.Limitations
The instruction expects ScienceClaw tools and report templates; moving one file to another agent requires environment adaptation. It requires both literature search and ToolUniverse even for a small topic, so define a narrow question first. Papers and databases may be unavailable or incomplete; the agent should expose gaps and the user should inspect key references. A model assembles the long report, so verify figures, chart labels and conclusions before sharing.
How to disable. Remove the deep-research folder from user Skills and stop the active task. Collected material and reports remain in the workspace.
FAQ
What comes before searching?
A plan of research dimensions and sources, available for review before collection.
Can it use only web search?
The instruction requires scientific literature and relevant ToolUniverse tools as well.
How do I get Word instead of PDF?
Request DOCX; the agent uses ScienceClaw’s corresponding document skill.
Where are intermediate files?
In the session workspace, with separate source data, summaries and report sections.
Can example numbers be used in charts?
No. The instruction requires collected evidence; sample values must be replaced with research results.
Discussion
No questions or reviews yet.
Related
SkillsA Claude Code skill for scientific literature reviews: an interview, Scopus and OpenAlex search, DOI checks and a final docx file
A skill set for academic papers: deep method analysis, visual slides and a coherent write-up from a paper PDF
A DeepSeek Harness plugin links research stages to data, experiments and checks before the next stage begins.
The paper-spine skill takes a topic, materials and data through a scientific paper: literature, text, figures, review and Word, LaTeX and PDF files
A set of three research skills: paper writing, academic Word and PowerPoint documents, scientific computing and figures
A library of 400+ academic skills: literature search, reviews, statistics, scientific writing and grant proposals
README in short
The skill lives in the user Skills directory and is matched to the request. It separately describes paper reading, data interpretation and bibliography assembly.
README badge
Original SKILL.md
---
name: deep-research
description: "多源深度调研与专业报告生成。适用场景广泛——只要用户的问题涉及需要深度分析的专业话题，就应使用此技能。包括但不限于：(1) 用户明确要求调研/research/综述/报告/发现；(2) 用户提出一个技术或科学话题，话题复杂度需要多源深度分析；(3) 用户要求对比多种技术方案的优劣；(4) 涉及生物医药、蛋白质、基因、药物靶点等需要专业数据库支撑的问题。核心能力：根据问题性质自动组合 arXiv 论文检索 + ToolUniverse 科学工具（UniProt/OpenTargets/PubMed/FAERS 等 1000+ 工具）+ Web 搜索，多源采集数据，交叉分析，生成专业 PDF/DOCX 研究报告。"
---
# Deep Research — 多源深度调研系统
端到端的深度调研系统：从用户问题出发，智能选择数据源组合（arXiv 论文 + ToolUniverse 科学工具 + Web 搜索），多维度采集数据，逐源深度解读，围绕用户问题进行交叉归纳，最终生成专业研究报告。
## 工作流程总览
```
Phase 1: 问题拆解与数据源规划
→ Phase 2: 多源数据采集
2A: arXiv 论文检索 + 筛选 + PDF 下载
2B: ToolUniverse 科学工具数据采集
2C: Web 搜索补充采集
→ Phase 3: 数据提取与整理
→ Phase 4: 逐源深度解读
→ Phase 5: 围绕用户问题的交叉分析与报告撰写
→ Phase 6: 生成 PDF/DOCX 研究报告
```
**工作空间路径**：所有文件输出到 `/home/scienceclaw/sessionid/` 目录下。
---
## Phase 1: 问题拆解与数据源规划
收到用户问题后，**不要直接执行**。先分析问题，拆解为多个调研维度，并规划需要哪些数据源。
### 1.1 问题拆解
将用户问题按以下维度展开：
| 维度 | 说明 | 示例 |
|------|------|------|
| 核心概念 | 问题的主题词 | BRCA1, data center cooling |
| 技术路线 | 不同的技术方案 | PARP inhibitors, liquid cooling |
| 应用场景 | 特定的应用环境 | triple-negative breast cancer |