biz-research-skills

Scaffold a new academic research repository for business/social science fields (economics, finance, accounting, management, marketing, etc.). Use when the user wants to start a new research project, create a paper repo, set up a research directory structure, or says things like "scaffold a research repo", "start a new paper on X", "new research project", "set up a project for X". Detects OS and installed tools first. Asks only what is unclear, then runs the scaffold script.

Literature discovery and bulk PDF download. PRIMARY discovery is a parallel team of 5-7 WebSearch agents (WebSearch tool ONLY — no rate-limited bibliographic APIs) that split the topic's sub-angles. EBSCO is used to RESOLVE found papers to downloadable records and to DOWNLOAD PDFs via CUFE WebVPN (multi-epoch — re-run until all PDFs land), plus a supplementary search. CRITICAL: ANY request to find, search, collect, or download academic papers IS a request to get PDFs on local disk. No "search-only" mode. Output is always PDF files + manifest.csv. Triggers on: 找论文 搜文献 下论文 文献检索 搜论文 下载论文 找文献 论文搜索 文献收集 查文献 学术搜索 期刊论文 英文文献 find papers search literature download papers get papers collect literature literature search academic papers bulk download journal articles.

Rapid literature digest for research topic selection. Reads a folder of PDF papers in parallel (one subagent per paper), each subagent writes its own structured note in Chinese. Main thread synthesizes research gaps and ideas across the full corpus. Use when the user wants to quickly understand a batch of papers, find research ideas, identify gaps, or needs a Chinese digest of downloaded papers. Triggers on: "帮我分析这批文献", "从这些 PDF 找研究 idea", "快速了解这些论文", "literature digest", "找研究空白", "选题", "research ideas from papers", or any request to process a folder of PDFs for research insight.

Assess data availability and acquisition feasibility for a research topic or specific variables. Evaluates commercial databases, public data sources, web scraping options, and alternative data construction strategies. Use when the user wants to know if data exists for their research idea, which databases to use, whether scraping is feasible, or how to construct a key variable. Triggers on: "数据在哪里", "这个变量怎么找", "有没有数据做这个", "data feasibility", "can I get data for X", "how to get data on X", "数据可得性", "能不能爬", "爬虫获取", "有没有现成数据", or after generating a research idea from lit-scout or deep-read.

Deep reading of a single academic paper to the level of full comprehension, replication, and idea generation. Reads the entire paper (not just abstract/intro/conclusion), produces a structured deep note covering research design, identification strategy, data, empirical specs, critical evaluation, and idea seeds. Ends in interactive dialogue mode for Q&A. Use when the user wants to deeply understand a specific paper, discuss a paper together, generate research ideas from a paper, or prepare for a paper presentation. Triggers on: "精读这篇文章", "帮我读这篇文章", "我想深入了解这篇论文", "deep read", "read this paper with me", "help me understand this paper", "generate ideas from this paper", or when user points to a specific PDF after running lit-scout.

Generate a beautifully typeset PDF reading note for a single academic paper using LaTeX. Takes the deep_note.md output from deep-read and compiles it into a clean, minimal academic document suitable for sharing with supervisors or colleagues. Use when the user wants to generate a PDF note from a paper reading, produce a formal reading note, share a paper summary with a supervisor, or says things like "生成论文笔记 PDF", "做一份可以发给导师的笔记", "把读书笔记编译成 PDF", "paper note PDF", "generate reading note".

Socratic research topic selection dialogue. Helps the user evaluate and choose between multiple candidate research ideas through multi-round discussion covering novelty, identification credibility, data feasibility, and publication risk. Only produces a written report when the user explicitly confirms they have chosen a topic. Use when the user wants to think through which topic to pursue, compare research ideas, discuss whether an idea is worth doing, or says things like "帮我想想选题", "我有几个 idea 不知道选哪个", "这个选题值得做吗", "topic selection", "help me decide on a topic", "讨论一下选题".