Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› ecta-rebuttal
ecta-rebuttal
GitHub用于处理Econometrica期刊的修改重投（R&R）及审稿意见回复。指导如何解析编辑信、规划重大修订，并撰写逐点回复信。强调先完成理论推导和代码复现再动笔，确保承诺的修复真实有效，避免空泛回应。
触发场景
收到Econometrica的修改重投决定信和审稿报告
需要制定修订计划并起草点对点回复信
审稿人指出证明漏洞、普遍性限制或缺少有限样本证据
达到有条件接受状态，需满足编辑最后条件
安装
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill ecta-rebuttal -g -y
SKILL.md
Frontmatter
{
"name": "ecta-rebuttal",
"description": "Use when a revise-and-resubmit or referee reports arrive for an Econometrica manuscript and you must write the response letter and plan the revision. Structures the rebuttal and revision; it does not re-derive the theorems (use ecta-theory-model) or rebuild the package (use ecta-replication-package)."
}
Rebuttal and Revision (ecta-rebuttal)
When to trigger
- A decision letter with a handling Co-Editor recommendation and referee reports has arrived
- You need to plan a (typically major) revision and draft a point-by-point response
- A referee flagged a proof gap, a generality limit, or missing finite-sample evidence
- You reached conditional acceptance and must satisfy the Co-Editor's last conditions and prepare final files for the ES Data Editor's reproducibility check
Do not draft the response before the revised theorems, simulations, and code are actually done — promised fixes that are not yet real read as evasive to this referee pool.
Read the letter correctly
- The co-editor's letter is the controlling document. Identify which points the co-editor treats as essential vs. which are referee suggestions left to your judgment. Address the co-editor's essential points fully and first.
- An R&R at Econometrica is a genuine opportunity (the acceptance rate is ~6–9%, and R&R is the common path to publication) but signals substantial remaining work and likely another round, followed by conditional acceptance and the Data Editor stage. Plan accordingly; do not do the minimum.
- Sort every comment into: (a) genuine error / gap to fix, (b) request for added generality / evidence, (c) presentation, (d) a point you will respectfully push back on with argument.
Response-letter structure
Dear Editor / [Co-editor],
We thank you and the referees for [specific, sincere]. The main changes are:
(1) [most important fix — e.g., closed the gap in the proof of Theorem 2],
(2) [added the limiting-distribution / uniform result requested],
(3) [added Monte Carlo design X showing finite-sample size control].
All changes are marked in the revision; new proofs are in Supplemental Material §...
— Response to the Co-Editor —
[CE point 1] ... [CE point 2] ...
— Response to Referee 1 —
[R1.1] Comment: "..."
Response: [what we did, where (section / theorem / appendix), why it resolves the point].
[R1.2] ...
— Response to Referee 2 —
...
Writing each response
- Quote the comment, then answer. Make it trivial for the referee to see you addressed it.
- Point to the exact location of each change (theorem number, appendix section, table).
- For a proof gap: state the corrected argument explicitly; do not just assert it is now fixed.
- For a generality request: either deliver the more general result, or argue precisely why the current scope is the right one (and add the boundary example that shows the limit binds).
- For more finite-sample evidence: add the requested design and report it; show the method's behavior even where it is less favorable — selective new evidence is transparent.
- Disagreement: allowed, but back it with a precise mathematical argument or a citation, framed respectfully. "We respectfully disagree, because [argument]; if the referee still prefers X, we are glad to add it."
- Consistency: never claim a fix in the letter that the revised manuscript does not contain.
Revision discipline
- Re-run the full replication package after every substantive change; numbers in the
letter, text, and tables must agree (see
ecta-replication-package). At conditional acceptance the ES Data Editor will re-run this code, so it must reproduce bit-for-bit. - Watch the 45-page body limit: a revision that adds results can push the body over the cap — move the overflow (proofs, extra Monte Carlo) into the Supplemental Material rather than cutting content.
- Re-check theorem numbering and cross-references after reorganizing — added results shift numbers.
- If a fix changes a result, trace it through: identification → asymptotics → simulations → text.
- Keep a marked-changes version; keep a clean version; keep the response letter in sync with both.
Checklist
- Co-editor's essential points identified and addressed first and fully
- Every referee comment quoted and answered with a location
- Proof gaps fixed with the corrected argument written out, not merely asserted
- Generality / evidence requests delivered or precisely rebutted
- New Monte Carlo / results added where requested; reported fairly
- Disagreements backed by argument, framed respectfully
- Full package re-run; letter, text, tables, and code all consistent
- Theorem numbering and cross-references re-verified after reorganization
Anti-patterns
- Promising fixes in the letter that the manuscript does not actually contain
- Treating a referee suggestion the co-editor flagged as essential as optional
- "We have addressed this" with no pointer to where
- Patching a proof gap with another hand-wave instead of a complete argument
- Adding only the flattering new simulation and omitting the adverse design the referee implied
- Dismissive disagreement with no mathematical backing
- Forgetting to re-run the package, leaving stale numbers in the revision
Output format
【Co-editor essential points】[...]
【Comment triage】fix: [...] | generality/evidence: [...] | presentation: [...] | push-back: [...]
【Proof gaps closed】[Theorem N: ...]
【New evidence added】[design/result ...]
【Push-backs (with argument)】[...]
【Package re-run + consistent】yes/no (Data Editor will re-run at conditional acceptance)
【Body ≤45pp after revision】yes/no — overflow moved to Supplemental Material: [...]
【Status】response letter drafted → re-submit via Editorial Express
版本历史
- 1839142 当前 2026-07-05 12:52
