rebuttal

Drafts grounded, venue-compliant rebuttals responding to external peer reviews.

Updated Jul 9, 2026
One-click install
npx skills add https://github.com/Lingjie-wang/autoRL --skill rebuttal-lingjie-wang
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: rebuttal
Source: https://github.com/Lingjie-wang/autoRL/tree/main/Auto-claude-code-research-in-sleep/skills/rebuttal
Command: npx skills add https://github.com/Lingjie-wang/autoRL --skill rebuttal-lingjie-wang

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Responding to peer reviews after paper submission is high-stakes and error-prone: authors must cover every reviewer concern, stay within strict character limits, avoid fabricating evidence, and never overpromise revisions. This Skill structures the entire rebuttal pipeline so no concern is dropped and no unsupported claim slips through. ## Core Features & Use Cases - Review parsing and issue tracking: Normalizes raw reviews, atomizes every concern into an issue board with severity, reviewer stance, and response mode, then builds a strategy plan with character budgets. - Safety-gated drafting: Enforces provenance, commitment, and coverage gates so every factual statement maps to the paper, reviews, or user-confirmed evidence, and every promised revision appears in a revision checklist. - Stress testing and follow-ups: Runs external reviewer stress-test rounds via Codex or manual review backends, supports per-reviewer thread or single-document venue modes, and manages follow-up rounds with delta replies. - Use Case: After receiving ICML reviews on OpenReview, point the Skill at your paper and the raw reviews to get an issue board, a strategy plan, a character-counted paste-ready response, and a revision plan for the camera-ready. ## Quick Start Run the rebuttal workflow on my paper directory and the pasted reviewer comments to produce a venue-compliant draft response.

Frequently Asked Questions about rebuttal

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write a rebuttal for ICML or OpenReview reviews?

Provide the paper source, raw reviewer text, and venue rules including character limit. The workflow atomizes each concern into an issue board, builds a strategy plan, drafts a grounded response, and outputs a paste-ready text file within the venue limit.

How does the rebuttal skill avoid fabricating evidence?

It enforces three hard gates: provenance requires every factual statement to map to the paper, reviews, or user-confirmed results; commitment blocks unapproved promises; coverage ensures no reviewer concern is dropped. Drafts failing any gate are not finalized.

Can it handle per-reviewer threads on OpenReview?

Yes. Set VENUE_MODE to per_reviewer_thread to generate one self-contained response file per reviewer instead of a single shared document. Each file is validated to be intelligible without cross-references to other reviewer threads.

Does the rebuttal workflow run new experiments automatically?

No, AUTO_EXPERIMENT is false by default, so evidence gaps are presented to the user for manual handling. When enabled, it invokes the experiment-bridge workflow for supplementary experiments and never fabricates results if experiments fail.

What are the limitations of this rebuttal workflow?

It does not run experiments, generate new theorem claims, edit or upload a revised PDF, or submit to OpenReview, CMT, or HotCRP. It produces text-only drafts and revision plans that the author must review and submit manually.