Resources

Claude Code is Anthropic’s coding agent, and this is the research workflow I have built on top of it. Twenty-one skills cover most of the empirical cycle, from finding and reading papers to auditing a .bib file to assembling a replication archive. Eight of them handle causal inference and natural experiments (difference-in-differences, regression discontinuity, instrumental variables, and conjoint among them), each built on a literature that has actually been read, so a recommendation arrives with the paper that licenses it.

It also ships my global CLAUDE.md, a Quarto slide system whose quality gates catch an overflowing slide before you find out at the podium, and a closing section, Things you may not know, on the features that changed how I work day to day, like the agent view for running many sessions at once and hold-to-talk voice input.

Everything there, including the two example decks, was generated with Claude Code as a proof of concept, so verify citations, numbers, and claims before relying on any of it.