Social Science Research

Hostile-referee stress test for papers. Not balanced critique — actively tries to kill the paper by finding fatal flaws, over-claims, unaddressed confounders, and identification weaknesses. Use before submission to pressure-test the argument against the strongest possible attack.

Substantive domain review for papers and analyses. Checks derivation correctness, assumption sufficiency, citation fidelity, code-theory alignment, logical consistency, and design-specific diagnostic reporting. Use after content is drafted or before presenting or submitting.

Dataset assessor for the data-finder skill. Applies a 5-point critique to each dataset proposed by the Explorer agent — measurement validity, sample selection, external validity, identification compatibility, and known issues. Adjusts feasibility grades and flags deal-breakers.

Dataset finder for the data-finder skill. Searches across public microdata, administrative data, survey panels, international sources, and novel alternatives. Returns a graded list of candidate datasets for a given research question. Dispatched in parallel pairs by data-finder.

First-time reader perspective on a paper. Reads cold with no prior context — not the spec, not the lit review, not the analysis scripts — and reports what lands, what confuses, and what a reader walks away believing. Use to catch clarity problems authors cannot see because they know too much.

End-to-end data analysis in R or Python — exploration through regression to publication-ready tables and figures. Use when running empirical analysis, writing analysis code, or producing output from data.

Find and assess datasets for a research question. Explorer agents search data source categories; Explorer-Critic stress-tests candidates and produces ranked list with feasibility grades. Use when identifying or evaluating data sources, not searching papers or running analysis.

Structured literature review via parallel Librarian agents. Searches top journals, working paper repos (NBER, SSRN, IZA); traces citation chains from key papers. Use when surveying existing research on a topic, not finding datasets or writing a paper.

Start a new research project. Structured interview formalizes the idea; generates research questions with identification strategies and a project spec. Use when developing or documenting a new research idea, not searching literature or data.

Proofread papers and manuscripts. Checks grammar, typos, layout, consistency, writing quality. Produces report without editing files. Use when finding surface-level errors, not substantive academic critique.

3 hooks across 3 events