/mstack:r-and-r

Stage: submit (R&R) · Voice: the author addressing editor and reviewers

Decision letter in hand, revision done: write the response that gets the paper across the line, not the one that argues with the reviewer.

$ARGUMENTS is the round (r1 default; r2, r3 for later rounds).

Procedure

1. Decision letter. Read submission/response-to-reviewers/r<N>-decision.md; if absent or empty, ask the user to paste the editor's letter and the reviewer reports there and stop until they do.
2. Manuscript and diff. Read paper/main.tex + paper/sections/, and output/tables/ and output/figures/ (they may have changed). If git is in use, git log --since="<submission date>" --stat -- paper/ shows what actually changed.
3. Parse the comments into a structured list: editor comments (top-level, then specific); each reviewer's comments numbered as the reviewer numbered them; tag each Major, Minor, or Editor, and note where reviewers contradict each other (resolve those by naming the conflict and following the editor's steer, or asking the editor). Triage: comments that change the headline get new analysis; comments about clarity get rewritten text; comments that misread the paper get the misreading corrected in the manuscript, not only in the response, since the next reader will misread it the same way.
4. Respond to each comment in three parts, in this order:
   1. Quote the comment verbatim (blockquote), so response and comment stay aligned.
   2. Respond. Concede where conceding is right, defend where defending is right, never concede the contribution to placate. Keep "we appreciate / we agree" sparse; the structure is enough.
   3. Point to the change — section, page or paragraph, and the new text quoted. If nothing changed, say so and why.
5. Editor opener: thank editor and reviewers once, briefly; the three most consequential changes; the structure of the document.
6. Change log table at the end: Comment ID | Change made | Location.
7. Cross-checks: every comment has a response; every claimed change cites a location; new analyses run through /mstack:results-audit and appear in the appendix with a pointer, not as loose numbers in the response; the contribution sentence in abstract / intro has not weakened relative to the prior version; no new claims beyond the letter (don't pick fights); the tone is the same for the reviewer you agree with and the one you do not.
8. Save submission/response-to-reviewers/r<N>-response.md; append - "<date>: r<N> <decision> — response drafted" to decisions: in .mstack/config.yaml and set paper.status: "r-and-r".

Outputs

• submission/response-to-reviewers/r<N>-decision.md (created if missing) and r<N>-response.md.
• Summary block: response count, most consequential changes, open items (comments without a manuscript change).

Anti-patterns

• Conceding the contribution to placate R2. Eroding the headline is worse than defending it well.
• Combat. Reviewers respond to deferential confidence.
• Hand-waving. "Addressed throughout the manuscript" is not a response; cite the page.
• Rewriting comments. Verbatim; the reviewer's words are the contract.

Next

/mstack:results-audit if any results changed; /mstack:referee-mock editor to stress-test the response; /mstack:coauthor-review on the revised manuscript.