Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› jama-peer-review-revision
jama-peer-review-revision
GitHub用于回复JAMA编辑和审稿人意见，特别是统计审查问题。指导如何先修订手稿再撰写逐点回复信，确保结构规范、引用准确、无夸大结论，并处理常见评论类型及准备追踪版本。
Trigger Scenarios
收到重大/轻微修改决定或鼓励性拒稿后
需要结构化地逐条回复编辑或审稿人意见
统计审稿人提出分析方法疑问时
必须在撰写回复前完成手稿修订
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill jama-peer-review-revision -g -y
SKILL.md
Frontmatter
{
"name": "jama-peer-review-revision",
"description": "Use when responding to JAMA editor and reviewer comments after a revise decision, including the statistical-review queries. Structures the point-by-point response and revision; it does NOT fabricate new analyses or overstate changes."
}
Peer Review & Revision (jama-peer-review-revision)
When to trigger
- A decision letter arrives: major/minor revision (or a reject-with-encouragement)
- Reviewer or editor comments need a structured, point-by-point reply
- The statistical reviewer has raised analysis questions
- You must revise the manuscript before writing the response — not after
How JAMA review works (durable shape)
- Editorial screening first; manuscripts that pass go to external peer review.
- JAMA applies dedicated statistical review to manuscripts under serious consideration — expect detailed methods/analysis queries.
- Decisions typically come with one or more revision rounds; address every point.
Response-letter structure
- Open with thanks and a one-paragraph summary of the main changes.
- Reproduce each comment verbatim, then respond beneath it.
- For each point: state the change, quote the new manuscript text, and give the page/line. If you disagree, do so respectfully with evidence — do not ignore a comment.
- Group editor vs reviewer comments; number them (Reviewer 1, Comment 1.1 …).
- Keep the manuscript and the letter perfectly consistent (numbers, claims, wording).
- Provide a tracked-changes manuscript alongside a clean copy.
Handling common JAMA comment types
| Comment type | Response approach |
|---|---|
| "Report effect sizes with CIs, not p-values" | Add estimates + 95% CIs throughout; confirm in the letter |
| "Was this outcome pre-specified?" | Cite the protocol/registry; relabel post hoc as exploratory |
| "Analysis should be intention-to-treat" | Re-run ITT as primary; move per-protocol to sensitivity |
| "Address multiplicity" | Add/justify the correction; downgrade unadjusted subgroups |
| "Conclusions overstate the data (spin)" | Rewrite to the primary outcome; bound the claim |
| "Reporting incomplete (CONSORT/PRISMA item X)" | Add the item; update the checklist locations |
| "Generalizability / limitations" | Strengthen the limitations paragraph honestly |
| "Registration timing" | State the facts transparently; do not obscure |
Checklist
- Manuscript revised first; letter written from the revised version
- Every comment reproduced and answered (none skipped)
- Each response cites new text with page/line
- Statistical-reviewer points fully resolved (CIs, ITT, multiplicity, missing data)
- No spin reintroduced; conclusions still bounded by the primary outcome
- Reporting checklist updated to match the revised text
- Tracked-changes + clean copies prepared
- Letter and manuscript numerically consistent
Anti-patterns
- Writing the response before actually revising the manuscript
- Silently skipping an uncomfortable comment
- Claiming a change was made that the manuscript does not reflect
- Defensive or dismissive tone toward reviewers
- Running new post hoc analyses and presenting them as confirmatory
- Reintroducing spin while "addressing" the spin comment
- Leaving the checklist out of sync with the revised text
Operating pass for JAMA
Run this as a concrete capability pass. First lock the clinical question, patient population, estimand or endpoint, safety/ethics issue, and reporting checklist; then test whether the manuscript addresses clinical reviewers who ask whether the evidence changes patient care, policy, or medical decision-making while satisfying reporting standards.
- Primary move: Return a claim-evidence-risk ledger; every recommendation must point to a manuscript location or missing artifact.
- Decision ledger: return
claim / evidence / blocker / next editrows so the next pass can patch the manuscript directly. - Sibling comparison: compare against NEJM for field-changing clinical medicine, Lancet for global-health breadth, specialty journals for narrower clinical domains; if the neighboring outlet has the stronger audience claim, recommend re-routing before polishing.
- Verification floor: before submission-ready advice, re-open
resources/official-source-map.mdfor volatile rules and name the one unresolved fact that could change the recommendation.
Output format
【Decision】major / minor revision / reject-with-encouragement
【Comments addressed】X of X (none skipped)
【Statistical-review points resolved】yes / remaining: ...
【New text cited with page/line】yes / no
【Spin reintroduced?】no / fixed: ...
【Checklist updated】yes / no
【Files ready】tracked + clean + point-by-point letter
Version History
- 1839142 Current 2026-07-05 13:24
