Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› restud-workflow
restud-workflow
GitHubREStud论文工作流路由技能，用于决定从选题到反驳的下一步操作。针对《计量经济学研究》期刊，根据当前瓶颈（如选题模糊、实证设计、理论模型等）推荐对应的子技能，不替代具体执行。
Trigger Scenarios
询问下一步该做什么
分析草稿瓶颈
在多阶段工作中迷失方向
收到审稿意见需切换至反驳模式
准备符合REStud标准的求职论文
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill restud-workflow -g -y
SKILL.md
Frontmatter
{
"name": "restud-workflow",
"description": "Use when deciding which restud-* sub-skill to invoke next, or when sequencing manuscript work from topic selection through rebuttal for a The Review of Economic Studies (REStud) manuscript. Routes — does not replace — the specialized skills."
}
REStud Workflow (restud-workflow)
Overview
This is the router. It does not replace any specialized skill — it tells you which one to use next, and in what order.
Default assumption: unless the user names a different venue, the manuscript targets The Review of Economic Studies (REStud) — a top-5 general-interest journal (OUP, for the Review of Economic Studies Ltd; founded 1933 to advance theoretical and applied economics equally, especially by younger economists). REStud prizes a clean, original contribution — a new model, a new identification strategy, or a striking new empirical fact — presented rigorously and economically. It is not a finance-only stack, not a theory-only stack, and not a working-paper repository.
REStud-specific anchors (carry these through every stage)
- Identity: founded 1933, theory and applied weighted equally; home of the REStud Tour (former May Meetings, annual since 1989) for promising young economists — calibrate ambition to that bar.
- Submission: Editorial Express portal, USD 250 fee (reduced USD 150 if every author is a student / within 6 years of PhD / in a low-or-middle-income economy; rate effective 1 June 2026), double-anonymous review, Harvard author-date references, JEL codes, ~45-page limit; editors desk-reject ~20–23% before refereeing.
- Replication: the Data Editor (Miklós Koren) runs an AEA DCAS pre-publication reproducibility check; packages deposit at Zenodo.
- Editors (2026): Joint Managing Editors include Jan De Loecker, Antonio Penta, Jakub Steiner (verify the full current roster).
When to Use
- The user asks "what should I work on next?"
- The user dumps a draft and you must decide where the bottleneck is
- The user is rotating between theory, empirics, writing, and revision and loses track of the stage
- A referee report has arrived and the work mode must switch from drafting to rebuttal
- The user is preparing a job-market paper for the REStud "May Meetings" / Tour pipeline and wants the manuscript at REStud standard
Routing Map
| Current symptom | Next skill |
|---|---|
| Idea is fuzzy; the one-sentence contribution will not compress | restud-topic-selection |
| Related-work positioning is thin; closest papers not confronted | restud-literature-positioning |
| Empirical design is the bottleneck (DID / IV / RDD / RCT / structural) | restud-identification |
| The paper is theory or has a model whose proofs need an online appendix | restud-theory-model |
| Main results exist but referee-anticipating checks are missing or fragile | restud-robustness |
| Tables are oversized / footnote-bloated; figures do not carry the result | restud-tables-figures |
| Prose is bloated; abstract over length; contribution buried | restud-writing-style |
| Preparing the data/code deposit or auditing reproducibility | restud-replication-package |
| Want to anticipate who referees and what they will demand | restud-referee-strategy |
| Final preflight before clicking submit (format, length, anonymity) | restud-submission |
| A referee report / R&R arrived and a response letter is needed | restud-rebuttal |
Default Sequence
For a typical REStud-track manuscript, prefer this order:
restud-topic-selection— fix the one-sentence original contribution before anything elserestud-literature-positioning— locate the contribution against the closest 3–5 papersrestud-identification— for empirical papers, stress-test the design; if it fails here, no later skill saves the paperrestud-theory-model— for theory papers (or empirical papers with a model), build the model and route proofs to the online appendixrestud-robustness— anticipate the checks the demanding median REStud referee will requirerestud-tables-figures— finalize exhibits so each one carries a resultrestud-writing-style— only now polish prose, abstract, and introductionrestud-replication-package— assemble the deposit while results are fresh in coderestud-referee-strategy— map the likely referee pool and pre-empt objectionsrestud-submission— final preflight: length, anonymity, format, filesrestud-rebuttal— after review, revise the manuscript first, then write the letter
restud-identificationandrestud-theory-modelare parallel branches: empirical papers lean on the first, theory papers on the second, and many REStud papers (theory-with-empirics) need both.restud-writing-styleis a late polish — do not run it before the contribution and design are stable.
Decision Cues
If the user says...
- "I have data but no story" →
restud-topic-selection - "My closest paper is in QJE 2022 and I don't know what I add" →
restud-literature-positioning - "My DiD has staggered treatment and TWFE" →
restud-identification - "My model has a proposition but no proof section" →
restud-theory-model - "A referee will say the result is driven by X" →
restud-robustness - "My table 3 has 14 columns" →
restud-tables-figures - "My intro is 5 pages and buries the contribution" →
restud-writing-style - "I got a conditional accept; I need the data deposit" →
restud-replication-package - "Who is going to referee this and what will they want?" →
restud-referee-strategy - "I'm submitting tomorrow" →
restud-submission - "I got an R&R with three reports" →
restud-rebuttal
Differences vs. Other Top-5 Stacks
REStud overlaps heavily with AER / QJE / JPE / Econometrica in standards, but its center of gravity differs:
- vs AER: AER weights cross-subfield interest and policy salience; REStud equally welcomes a clean theoretical or methodological contribution with no policy hook, and is famously friendly to young scholars — the REStud Tour / May Meetings (annual since 1989) showcase graduating PhD job-market work.
- vs Econometrica: Econometrica tolerates very heavy formal machinery as the contribution itself; REStud wants the economics payoff of the theory to be legible.
- vs JPE: JPE has a price-theory / Chicago tilt; REStud is field-agnostic and European-rooted.
If the paper is finance-specific (JF / JFE / RFS) or theory so abstract it is really a math paper, a different stack fits better.
Anti-Patterns
- Polishing prose (
restud-writing-style) before the identification or model is stable - Treating
restud-theory-modelas optional for an empirical paper that has a model — REStud referees check that the model's assumptions map to the empirics - Assembling the replication package only after acceptance — REStud's Data Editor (Miklós Koren) runs an AEA DCAS reproducibility check before publication (Zenodo deposit), so it is part of the production timeline, not an afterthought
- Writing the rebuttal against the old draft instead of the revised manuscript
- Skipping
restud-referee-strategybecause "the design is obviously fine" — REStud referees are demanding and develop strong papers across rounds
Handoff Contract
Whenever this skill is invoked, end with:
NEXT SKILL: <restud-skill-name>
REASON: <one sentence>
INPUTS NEEDED: <artifacts the next skill needs>
Version History
- 1839142 Current 2026-07-05 14:21
