IEEE Reviewer Response and Cover Letter
Use this skill to turn a set of reviews into a disciplined response letter, and to write the cover
letter that frames a submission for the editor. The response letter is a contract: each reply
states what changed and where to find it.
Core stance
- Answer every comment, in order. No reviewer point is skipped, merged away, or silently
ignored — even the ones you disagree with.
- Every "we changed X" must be locatable. Quote the revised text or cite the section/line; a
promise with no anchor reads as evasion.
- Cooperative, not submissive. Thank, then engage on the merits. You may disagree — back it
with evidence, not insistence — but concede readily when the reviewer is right.
- No fabrication. Do not claim experiments you did not run, results you did not get, or
citations you did not add. If a request needs work not yet done, mark it
[AUTHOR INPUT NEEDED]
and list it — never paper over it.
- The letter and the manuscript agree. Anything the response says you did must actually exist
in the revised PDF.
When to open extra files
| File |
Open when |
| references/response-structure.md |
Building the point-by-point letter: comment IDs, the triage/action model, reply templates, agree/disagree/defer patterns, traceability, tone |
| references/cover-letter.md |
Writing the cover letter for a new or revised submission: structure, contribution framing, editor-facing claims, required statements |
The triage model (classify every comment first)
ACCEPT_TEXT reviewer is right → make the change, quote it in the reply
SOFTEN_CLAIM reviewer found an overclaim → narrow the wording, point to the new sentence
ADD_EVIDENCE needs an experiment/analysis → run it (ieee-experiments) or mark AUTHOR INPUT NEEDED
CLARIFY a misunderstanding → fix the text that misled, then explain (the fault is usually the writing)
DISAGREE reviewer is mistaken → rebut with evidence, respectfully, and still add a clarifying line
DEFER out of scope / future work → acknowledge, justify briefly, note in limitations
Disagreement is legitimate but expensive: every DISAGREE should still produce a small text change
so the reviewer sees they were heard.
Workflow
- Extract and number every comment. Split each reviewer's text into atomic points
(R1.1, R1.2, ...). One concern per ID, even if buried mid-paragraph.
- Triage each with the model above; decide the action before drafting a word.
- Do the work first, write the reply second. Text edits via ieee-writing/ieee-polishing, new
results via ieee-experiments. The reply describes work that exists.
- Draft each reply as: thank/restate → what we did → where it is (quote or location). Keep
the reviewer's words quoted above your response.
- Open with a summary of the main changes and a thank-you to the editor and reviewers.
- Check coverage: every comment ID has a reply; every "changed" has an anchor; no fabricated
claim; tone consistent.
- Write the cover letter separately (see cover-letter.md) — it addresses the editor, not the
reviewers, and frames fit and contribution.
- Return the response letter, the cover letter, and a coverage checklist.
Output format
Response letter: per-reviewer, numbered, quoted comment + reply (thank → change → location).Cover letter: editor-facing, framing contribution and fit (and, for revisions, the headline
changes).Coverage check: table of Comment ID → action → anchor in manuscript → status.Open items: anything marked [AUTHOR INPUT NEEDED] the user must still supply.
1---2name: ieee-response3description: Draft the two letters an IEEE submission needs — the point-by-point response to reviewers and the cover letter to the editor — so every comment is answered, every change is traceable to the revised manuscript, and the tone stays cooperative without conceding more than the evidence requires. Use this whenever the user is replying to a review or submitting a manuscript: "respond to reviewers", "rebuttal letter", "reviewer 2 says...", "write a cover letter", "我要回复审稿意见", "审稿回复", "投稿信", "怎么回复 reviewer". For changing the manuscript text the response promises, hand to ieee-writing/ieee-polishing; for new simulations a reviewer demands, hand to ieee-experiments.4---56# IEEE Reviewer Response and Cover Letter78Use this skill to turn a set of reviews into a disciplined response letter, and to write the cover9letter that frames a submission for the editor. The response letter is a contract: each reply10states what changed and where to find it.1112## Core stance1314- **Answer every comment, in order.** No reviewer point is skipped, merged away, or silently15 ignored — even the ones you disagree with.16- **Every "we changed X" must be locatable.** Quote the revised text or cite the section/line; a17 promise with no anchor reads as evasion.18- **Cooperative, not submissive.** Thank, then engage on the merits. You may disagree — back it19 with evidence, not insistence — but concede readily when the reviewer is right.20- **No fabrication.** Do not claim experiments you did not run, results you did not get, or21 citations you did not add. If a request needs work not yet done, mark it `[AUTHOR INPUT NEEDED]`22 and list it — never paper over it.23- **The letter and the manuscript agree.** Anything the response says you did must actually exist24 in the revised PDF.2526## When to open extra files2728| File | Open when |29|---|---|30| [references/response-structure.md](references/response-structure.md) | Building the point-by-point letter: comment IDs, the triage/action model, reply templates, agree/disagree/defer patterns, traceability, tone |31| [references/cover-letter.md](references/cover-letter.md) | Writing the cover letter for a new or revised submission: structure, contribution framing, editor-facing claims, required statements |3233## The triage model (classify every comment first)3435```36ACCEPT_TEXT reviewer is right → make the change, quote it in the reply37SOFTEN_CLAIM reviewer found an overclaim → narrow the wording, point to the new sentence38ADD_EVIDENCE needs an experiment/analysis → run it (ieee-experiments) or mark AUTHOR INPUT NEEDED39CLARIFY a misunderstanding → fix the text that misled, then explain (the fault is usually the writing)40DISAGREE reviewer is mistaken → rebut with evidence, respectfully, and still add a clarifying line41DEFER out of scope / future work → acknowledge, justify briefly, note in limitations42```4344Disagreement is legitimate but expensive: every DISAGREE should still produce a small text change45so the reviewer sees they were heard.4647## Workflow48491. **Extract and number every comment.** Split each reviewer's text into atomic points50 (R1.1, R1.2, ...). One concern per ID, even if buried mid-paragraph.512. **Triage each** with the model above; decide the action before drafting a word.523. **Do the work first, write the reply second.** Text edits via ieee-writing/ieee-polishing, new53 results via ieee-experiments. The reply describes work that exists.544. **Draft each reply** as: thank/restate → what we did → where it is (quote or location). Keep55 the reviewer's words quoted above your response.565. **Open with a summary** of the main changes and a thank-you to the editor and reviewers.576. **Check coverage:** every comment ID has a reply; every "changed" has an anchor; no fabricated58 claim; tone consistent.597. **Write the cover letter** separately (see cover-letter.md) — it addresses the editor, not the60 reviewers, and frames fit and contribution.618. **Return** the response letter, the cover letter, and a coverage checklist.6263## Output format64651. `Response letter:` per-reviewer, numbered, quoted comment + reply (thank → change → location).662. `Cover letter:` editor-facing, framing contribution and fit (and, for revisions, the headline67 changes).683. `Coverage check:` table of `Comment ID → action → anchor in manuscript → status`.694. `Open items:` anything marked `[AUTHOR INPUT NEEDED]` the user must still supply.
TenWalk