Ccf Rebuttal Writer

CCF Rebuttal Writer

Core Rule

Handle post-review communication and revision accountability. Responses must be calm, factual, evidence-grounded, and promise only feasible changes. Resubmission adaptation is conservative by default: no new experiments and no bibliography changes unless the user explicitly authorizes them. Follow the user's requested response format: plain text, TeX, reviewer-by-reviewer, issue-grouped, table-first, or short response.

Modes

• rebuttal: reviewer/AC response under a word or time budget.
• revision-ledger: reviewer comment -> action -> manuscript location -> owner -> status.
• response-letter: revision summary or camera-ready response letter.
• resubmission: adapt an already written paper to a new venue with conservative defaults.

Workflow

1. Identify venue, response format, word budget, deadline, review scores/confidence, and whether this is rebuttal, revision, or resubmission. 2. Parse comments into issue groups by reviewer, concern type, severity, available evidence, response strategy, and promised paper change. 3. Load references/response-strategy.md and answer high-impact concerns first: soundness, novelty, missing evidence, incorrect assumptions, and shared concerns. 4. Load references/revision-ledger.md whenever promised edits, manuscript locations, or resubmission actions must be tracked. 5. For full rebuttals, load references/tex-templates.md and use the TeX templates in assets/templates/ when useful. 6. For resubmission, map old reviewer concerns to the new venue's constraints through ccf-submission-checker; do not silently add experiments or bibliography changes. 7. Hand off to ccf-paper-writer for manuscript revisions, ccf-experiment-designer for authorized new evidence, and ccf-submission-checker for venue/package checks.

Adaptive Output Contract

Put the requested response artifact first. For "write rebuttal", output the rebuttal text or TeX first. For "make a ledger", output the ledger first. Use the full structure below for standard multi-reviewer response planning or when the user asks for strategy plus draft:

Mode:
Venue and constraints:
Issue table:
Response strategy:
Draft response or response file:
Promised paper changes:
Revision ledger:
Resubmission adaptation notes:
Claims/promises to avoid:
Next CCFA owner:
Checklist status:

References

• references/response-strategy.md: response tactics.
• references/response-checklists.md: tone, evidence, promises, and word-budget checks.
• references/tex-templates.md: reusable TeX response templates.
• references/revision-ledger.md: tracking reviewer comments and manuscript actions.

Response Checklists

Use this file to prevent omissions when drafting conference paper rebuttals, author responses, response letters, revision summaries, or resubmission responses.

Intake Checklist

• Venue and response type are known.
• Word limit, formatting constraints, and deadline are known or marked unknown.
• Review scores, confidences, and reviewer identities are preserved if provided.
• AC/meta-review comments are separated from individual reviewer comments.
• Manuscript, appendix, and available extra results are listed.
• Missing information that affects confidence is named.

Issue Table Checklist

For each issue:

Reviewer:
Comment:
Quoted core concern:
Deeper intent:
Concern type:
Severity:
Decision relevance:
Evidence available:
Response tactic:
Promised change:
Risk:

Response Draft Checklist

Each response should include:

• brief appreciation,
• quoted core concern,
• direct answer before explanation,
• direct answer,
• evidence or location,
• deeper-intent answer when relevant,
• exact revision promise if applicable,
• scoped limitation when needed,
• self-contained acronyms, settings, or dataset details,
• no unsupported new claim,
• no defensive wording.

Positive Opening Checklist

• Start from strengths, reviewer agreement, or acknowledged value when available.
• Mention the paper's central contribution in one concise sentence.
• Avoid ignoring criticisms; use the opening to frame the rest of the response.
• Keep the tone appreciative and grounded.

Common Concern Checklist

• Merge duplicated concerns across reviewers.
• Name involved reviewers using color labels.
• Answer the shared underlying concern once.
• Add paper locations, datasets, numbers, tables, or figures when available.
• Refer back to the common response in reviewer-specific clarifications.

Promise Safety Checklist

Before promising a change, verify:

• The result, analysis, citation, or text change is available or feasible.
• The promise is specific enough to be credible.
• The promised change addresses the actual concern.
• The response does not imply a stronger claim than the paper can support.
• New experiments are not promised unless available or explicitly authorized by the user.
• If the answer can be delivered now, include the result or explanation in the response rather than only promising a future edit.
• Infeasible experiments are disclosed transparently.

AC / Meta-Review Response Checklist

• Identify the decisive issue.
• Summarize how multiple reviewer concerns connect.
• Prioritize changes that affect acceptance discussion.
• Avoid spending budget on minor edits before fatal risks.
• State the strongest paper-level clarification or revision.

Final Readiness Checklist

Do not call the response ready until:

• The default TeX structure or selected template is clear for full rebuttals.
• Positive signals and common ground are stated when available.
• High-confidence, decision-relevant concerns are answered first.
• High-severity concerns are answered first.
• Every answer has evidence or a clear limitation.
• Literal comments and deeper intent are both addressed for major issues.
• Reviewer quotations are concise and followed by direct answers.
• Reviewer color labels are used for multi-reviewer TeX responses.
• Related concerns are merged when doing so saves space.
• Existing paper coverage is cited by line, section, table, figure, or appendix when available.
• Acronyms, experimental settings, and critical context are self-contained.
• Data and completed analyses are preferred over argumentative claims.
• Tone is calm and respectful.
• Promises are feasible and specific.
• Constructive reviewers are thanked specifically where appropriate.
• Hostile or inappropriate review behavior is handled calmly and, if needed, separated into confidential escalation.
• Word budget is respected or overflow is flagged.
• Remaining risks are named.
• The final answer includes checklist status.

Minimal Checklist Status

Checklist status:
- Constraints:
- Issue grouping:
- Positive opening:
- Quote/direct answer:
- Deeper intent:
- Evidence support:
- Promise safety:
- Reviewer colors:
- Tone:
- Word budget:
- Remaining risks:

Response Strategy

Use this file to classify reviewer concerns and choose a response tactic.

Concern Types

• Misunderstanding: the paper already contains the answer but it was easy to miss.
• Missing clarity: the work is sound, but the explanation is incomplete or buried.
• Valid weakness: the reviewer is right and the response should concede, bound, and revise.
• Missing evidence: existing evidence is insufficient or not visible enough.
• Requested new result: the reviewer asks for a baseline, ablation, theorem, dataset, user study, or analysis.
• Scope or limitation: the issue is outside the paper's claim or should be bounded honestly.
• Venue-fit or significance concern: the reviewer is unsure why the work matters to the venue.
• Policy or ethics concern: the response must be factual, cautious, and compliance-oriented.
• Hostile or guideline-violating review: the review appears dismissive, inappropriate, or inconsistent with venue rules.

Global Response Pattern

For a full rebuttal, use this high-level order:

1. Positive opening: begin with strengths, agreement, or reviewer-recognized value. 2. Common concerns: merge repeated high-impact issues across reviewers. 3. Reviewer-specific clarifications: answer individual comments with reviewer color labels. 4. Closing signal: emphasize shared goals, concrete clarifications, and remaining transparency.

Within each issue, use:

Quote -> Direct answer -> Evidence -> Deeper intent -> Exact change or transparent limitation

Prefer answers that give concrete data, paper locations, or already-completed analyses over argumentative prose.

Response Tactics

Positive opening

Use when reviewers acknowledge strengths, novelty, clarity, usefulness, strong experiments, or practical value.

Rules:

• Start by thanking reviewers and restating the strongest positive signals.
• Do not sound self-congratulatory.
• Tie the positive signal to the paper's central contribution.
• Use this to keep the committee aware of the work's upside while reading criticisms.

Quote then direct answer

Use for every material concern.

Structure:

\reviewerquote{R2}{"...why was GLORP3 not evaluated?"}
\directanswer{Direct answer: GLORP3 is not in the task scope because [...]. We additionally report results on [datasets], which test the same completeness concern.}

Rules:

• Quote only the core concern, not the full review.
• Give a direct answer before background.
• Keep the reply conversational and evidence-grounded.

Deeper intent

Use when a literal question points to a broader concern.

Examples:

• "Why no GLORP3?" may mean "Is the evaluation complete?"
• "Why this baseline?" may mean "Are comparisons fair?"
• "What is the mechanism?" may mean "Is the method sound?"

Answer the literal question first, then address the deeper concern so other committee members do not inherit the doubt.

Merge common concerns

Use when multiple reviewers raise the same underlying concern.

Rules:

• Create one common-concern response before reviewer-specific details.
• Name the reviewers with color labels.
• Explain the shared concern once, then refer back in individual clarifications.
• This saves budget and reduces repeated defensive language.

Clarify

Use when the evidence exists but was hard to find.

Structure:

Thank you for pointing this out. The paper addresses this in [location], where we show [evidence]. To make this clearer, we will revise [section/figure/table] to state [specific change].

Correct misunderstanding

Use when the reviewer inferred something incorrect.

Rules:

• Do not say the reviewer is wrong.
• State the correct interpretation and why the manuscript may have caused confusion.
• Promise a precise wording or figure change.

Concede and revise

Use when the reviewer identified a real weakness that can be fixed in text or analysis.

Structure:

We agree this was insufficiently clear. We will revise [section] to [specific action], and we will add [evidence/detail] to support [claim].

Add existing evidence

Use only when the result already exists.

Rules:

• Name the exact result, metric, dataset, table, figure, appendix, or experiment.
• Say where it will be moved or highlighted.
• Do not claim a new experiment is complete unless it is.

Deliver now, not vague promise

Use when the response can include a new analysis, number, table summary, or explanation immediately.

Rules:

• Prefer "We tested ... and found ..." over "We will test ...".
• Include the key result in the response itself.
• If manuscript changes are promised, still give the substantive answer now.

Use emphasis

Use sparingly to guide busy committee readers.

Rules:

• Bold only the decisive answer, dataset, result, or limitation.
• Avoid all-caps except for short labels such as DIRECT ANSWER.
• Do not use emphasis to compensate for weak evidence.

Narrow claim

Use when the paper overclaims.

Structure:

We agree the original wording was too broad. We will revise the claim from [broad claim] to [scoped claim], which is supported by [evidence].

Bound limitation

Use when the concern is real but not fatal.

Rules:

• Admit the limitation plainly.
• Explain why it does not invalidate the central contribution.
• Promise a limitation or discussion update.

Decline infeasible request

Use sparingly when requested work cannot be completed before the response deadline.

Rules:

• Be respectful.
• Explain why the existing evidence still supports the central claim.
• State whether the issue will be added as a limitation or future work.

Handle hostile or inappropriate review

Use only when the review appears to violate reviewing norms or contains unsupported hostile claims.

Rules:

• Do not retaliate in the public response.
• Calmly correct factual errors and point to evidence.
• Ask the committee to evaluate the concern against the evidence.
• If necessary, suggest a confidential note to the program chairs rather than escalating in the rebuttal text.

Thank constructive reviewers

Use when a reviewer provides careful suggestions or helpful analysis.

Rules:

• Thank them specifically.
• If their suggestion was tried, report what happened.
• If it cannot be done, explain transparently and preserve goodwill.

Priority Rules

Answer in this order:

1. Positive opening and shared strengths. 2. AC/meta-review decisive concerns. 3. High-confidence answers to fatal soundness, novelty, evidence, or completeness concerns. 4. Shared concerns raised by multiple reviewers. 5. Misunderstandings that can change score or confidence. 6. Missing baselines, ablations, or reproducibility details. 7. Harder secondary issues and transparent limitations. 8. Local clarity, formatting, minor suggestions, and future-work items.

Tone Rules

• Thank the reviewer without over-apologizing.
• Be direct and evidence-specific.
• Avoid emotional language and defensiveness.
• Avoid "we believe" when a factual result can be cited.
• Avoid promising more than can be done.
• Prefer "we will revise" only for concrete manuscript changes.
• Re-explain acronyms, datasets, settings, and key assumptions because committee members may not remember the paper.
• Emphasize shared goals: fair comparison, reproducibility, complete evaluation, or clearer claims.

Revision Ledger

Use this reference when rebuttal, response-letter, resubmission, or camera-ready work needs a traceable ledger. The ledger is now part of ccf-rebuttal-writer; it is not a separate runtime skill.

Ledger Columns

Rules

1. Do not mark an action done unless the manuscript location or artifact exists. 2. Do not let response text promise an action that is absent from the ledger. 3. Use accepted_limit when the right response is to acknowledge a limitation rather than hide it. 4. Use blocked only when the missing artifact, result, permission, or policy cannot be resolved inside this skill. 5. Keep wording factual; persuasive tone belongs in the response draft, not in the ledger.

Compact Template

| ID | Concern | Response claim | Manuscript action | Location | Status | Evidence | Risk |
| --- | --- | --- | --- | --- | --- | --- | --- |
| R1-C1 |  |  |  |  | open |  |  |

TeX Templates

Use this file when the user asks for a rebuttal TeX file, author-response TeX, response-letter template, or a full multi-reviewer response.

Default Selection

Use assets/templates/default-general-common-reviewer.tex by default for full rebuttals. It contains:

• General Response
• Common Concerns
• Reviewer-Specific Clarifications

This default is best when the user provides multiple reviews and wants a balanced response that starts from strengths, merges shared issues, and then gives reviewer-specific clarifications.

Template Choices

• assets/templates/default-general-common-reviewer.tex: Default. Use for most full rebuttals.
• assets/templates/reviewer-specific-response.tex: Use when the venue expects responses organized by reviewer or each reviewer has distinct concerns.
• assets/templates/compact-response.tex: Use when there is a strict word or character budget.
• assets/templates/ac-focused-response.tex: Use when AC/meta-review concerns, global misunderstanding, or one decisive issue dominates.

Required TeX Features

Templates should include:

• reviewer color macros,
• concise reviewer quote blocks,
• direct answer blocks,
• evidence blocks with line/table/figure references,
• common-concern merging,
• self-contained acronyms and experimental settings,
• transparent limitation language.

Filling Rules

1. Replace reviewer labels with the actual reviewer IDs used by the venue. 2. Quote only the core concern. 3. Put the direct answer before background. 4. Use bold emphasis only for decisive answers, evidence, and limitations. 5. Cite paper locations such as Section, Table, Figure, Appendix, or line number when available. 6. Include newly completed analyses or results in the rebuttal text instead of only promising them. 7. Keep reviewer-specific answers short when a common concern already answers the main issue.