brycewang-stanford/Awesome-Agent-Skills-for-Empirical-Research

Imagine having access to a structured taxonomy of more than 23,000 discrete agent skills mapped across eight social science disciplines — economics, political science, sociology, education, public health, finance, management, and psychology. That is what this repository delivers. Each skill represents a modular capability an AI agent can call on during empirical research workflows: data sourcing, variable construction, regression specification, robustness checks, and more. The companion product, CoPaper.AI, ties these skills together into an end-to-end pipeline that reportedly produces a reproducible, properly structured empirical paper in roughly 20 minutes. For builders, the real value is the ontology itself. If you are constructing research-automation software, academic copilots, or AI-assisted grant-writing tools, this gives you a ready-made skill architecture rather than forcing you to design one from scratch. The user-uploadable skills feature also hints at a community extension model worth watching. The honest reservation is that quality and peer validation of individual skills vary considerably at this scale, and the 20-minute paper claim will draw justified skepticism from serious researchers. -> Best for: founders building AI-powered academic research or scholarly workflow tools targeting social scientists.