Skip to content

AI Research Feedback Skills

Key Ideas

  • A collection of five Claude Code skills specifically designed for academic research review: paper review, light paper check, paper-code reproducibility review, pre-analysis plan review, and grant proposal review
  • The full /review-paper skill runs six specialized agents in parallel covering style, internal consistency, unsupported claims, mathematics, tables/figures, and an adversarial journal-specific referee persona
  • Supports targeting specific journals (AER, QJE, JPE, Econometrica, REStud, JF, JFE, RFS, JFQA, and macro journals) to simulate that journal's editorial standards
  • /review-paper-code checks reproducibility and paper-code alignment across Stata, R, and Python, verifying that empirical claims in the paper match the analysis code
  • /review-pap reviews pre-analysis plans against registry standards (AEA, EGAP, OSF) or journal standards, evaluating specification completeness, identification strategy, and statistical analysis
  • /review-grant runs a six-agent panel review for grant proposals targeted at specific funders (NSF, NIH, ERC, Horizon Europe)
  • All skills auto-detect main files (.tex, proposal documents) and produce dated, versioned output reports

Summary

This repository provides a set of Claude Code skills purpose-built for the academic research review workflow. The flagship skill, /review-paper, simulates a pre-submission referee report by running six specialized review agents in parallel. Each agent focuses on a distinct dimension: spelling and academic style, internal consistency and cross-references, unsupported claims and identification integrity, mathematics and notation, tables and figures, and a contribution evaluation from an adversarial referee persona calibrated to a specific journal. Users can target journals from top-5 economics (AER, QJE, JPE, Econometrica, REStud), finance (JF, JFE, RFS, JFQA), or macro (AEJMacro, JME, RED), and the skill adjusts its standards accordingly.

The collection also includes a lighter-weight /review-paper-light that runs just two agents for quick iteration on contribution and identification issues, and /review-paper-code which maps a paper's empirical claims to analysis code and checks for reproducibility issues across Stata, R, and Python codebases. Two additional skills address earlier stages of the research lifecycle: /review-pap evaluates pre-analysis plans against trial registry standards (AEA, EGAP, OSF, ClinicalTrials) or journal standards, checking specification completeness, identification strategy, and statistical analysis plans; and /review-grant runs a six-agent panel review of grant proposals targeted at specific funders.

All skills follow consistent design patterns: automatic file discovery, configurable depth and targeting, dated and versioned output files, and the ability to add project-specific context via local CLAUDE.md files. The tools require Claude Code with subagent access and work primarily with LaTeX source files, though some accept PDF and DOCX inputs.

Relevance to Economics Research

This is among the most directly useful tools for economics researchers in this wiki. The skills address exactly the workflow gaps that academics face: getting structured feedback before submission, checking paper-code alignment for reproducibility, reviewing pre-analysis plans before registration, and pressure-testing grant proposals. The journal-specific calibration for top economics and finance journals (including JF, JFE, RFS, and JFQA) means the feedback is tailored to the standards researchers actually face. The /review-paper-code skill is particularly valuable given the growing emphasis on replication in economics -- it systematically checks whether the code produces what the paper claims. For researchers working with coauthors, these skills provide a fast way to get structured feedback between rounds of revision without waiting for colleagues' availability.