/scholar-rebuttal-pro

name: scholar-rebuttal-pro
disable-model-invocation: true
description: Enhanced academic paper review response workflow with Agy/CLI collaborative analysis and multi-perspective discussion. Produces structured rebuttal documents with evidence-based strategies. Triggers on "rebuttal", "respond to reviewers", "review response", "审稿回复".
allowed-tools: Task, AskUserQuestion, TodoWrite, Read, Write, Edit, Bash, Glob, Grep, Skill, mcp__ace-tool__search_context, mcp__maestro__read_file, mcp__maestro__edit_file
session-mode: run
contract:
  consumes: []
  produces:
    - { path: outputs/rebuttal-draft-v1.md, kind: rebuttal-draft, alias: latest-rebuttal, role: primary }
    - { path: outputs/review-analysis.json, kind: review-analysis, role: attachment }
    - { path: outputs/strategy-matrix.md, kind: strategy-matrix, role: attachment }
    - { path: outputs/quality-report.md, kind: quality-report, role: evidence }
  gates:
    entry: []
    exit: []

<required_reading> @~/.maestro/workflows/run-mode.md </required_reading>

Scholar Rebuttal Pro

Enhanced academic paper review response workflow combining Agy/CLI collaborative analysis with multi-perspective discussion. Produces structured, evidence-based rebuttal documents optimized for conference-specific requirements.

Pre-load (before execution)

1. **Codebase docs**: If `.workflow/codebase/ARCHITECTURE.md` exists, read for project context 2. **Specs**: `maestro load --type spec --category coding` — load coding conventions 3. **Wiki knowledge**: `maestro search "academic writing research paper" --json` — top 5 entries as prior context 4. All optional — proceed without if unavailable

Architecture Overview

┌─────────────────────────────────────────────────────────────────┐
│  Scholar Rebuttal Pro Orchestrator (SKILL.md)                    │
│  → Pure coordinator: Execute phases, parse outputs, pass context  │
│  → Run lifecycle: create/resume → phases → check → complete       │
└───────────────────────┬─────────────────────────────────────────┘
                        │
    ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐
    │ Phase 1 │ │ Phase 2 │ │ Phase 3 │ │ Phase 4 │ │ Phase 5 │
    │ Review  │ │ Multi-  │ │Strategy │ │Rebuttal │ │ Quality │
    │ Parsing │ │Perspect │ │Formula  │ │ Writing │ │Validat  │
    └─────────┘ └─────────┘ └─────────┘ └─────────┘ └─────────┘
      reviewA    discussion   strategy    rebuttal    quality
      nalysis    Consensus    Matrix      Draft       Score

Key Design Principles

1. **CLI-Assisted Analysis**: Leverage Agy CLI for semantic analysis, evidence gathering, and quality validation 2. **Multi-Perspective Discussion**: Simulate author/reviewer/expert viewpoints to develop robust strategies 3. **Evidence-Based Responses**: Link every response to paper content or experimental evidence 4. **Conference-Agnostic Templates**: Support extensible template system for different venues 5. **Progressive Disclosure**: Load phase documents on-demand to manage context window

Interactive Preference Collection

Collect workflow preferences via AskUserQuestion before dispatching to phases:

const prefResponse = AskUserQuestion({
  questions: [
    {
      question: "是否跳过所有确认步骤（自动模式）？",
      header: "Auto Mode",
      multiSelect: false,
      options: [
        { label: "Interactive (Recommended)", description: "交互模式，每阶段后确认" },
        { label: "Auto", description: "跳过所有确认，自动执行" }
      ]
    },
    {
      question: "论文内容来源？（用于策略制定时查找支撑证据）",
      header: "Paper Source",
      multiSelect: false,
      options: [
        { label: "Provide Path", description: "指定论文 PDF/LaTeX 路径" },
        { label: "Current Directory", description: "自动搜索当前目录" },
        { label: "Review Only", description: "仅基于审稿意见回复" }
      ]
    },
    {
      question: "目标会议类型？（影响模板和策略选择）",
      header: "Conference",
      multiSelect: false,
      options: [
        { label: "ML Conferences", description: "NeurIPS/ICML/ICLR" },
        { label: "CV Conferences", description: "CVPR/ECCV/ICCV" },
        { label: "NLP Conferences", description: "ACL/EMNLP" },
        { label: "Generic", description: "通用模板" }
      ]
    }
  ]
})

// Derive workflowPreferences from user selection
workflowPreferences = {
  autoYes: prefResponse["Auto Mode"] === "Auto",
  paperSource: prefResponse["Paper Source"],
  conferenceType: prefResponse["Conference"]
}

**workflowPreferences** is passed to phase execution as context variable. Phases reference as `workflowPreferences.autoYes`, `workflowPreferences.paperSource`, etc.

Auto Mode Defaults

When `workflowPreferences.autoYes === true`:

• Skip confirmation after each phase
• Use recommended strategies from multi-perspective discussion
• Apply default conference template (Generic)
• Auto-proceed to quality validation

Execution Flow

> **⚠️ COMPACT DIRECTIVE**: Context compression MUST check TodoWrite phase status. > The phase currently marked `in_progress` is the active execution phase — preserve its FULL content. > Only compress phases marked `completed` or `pending`.

Run Setup (see run-mode.md):
   └─ Birth packet injected run_id/run_dir? → use them, skip create.
      Else self-start: maestro run create scholar-rebuttal-pro --session <YYYYMMDD-scholar-rebuttal-pro-{topic}> --intent "..."
      (Optional --resume <run_id> → maestro run brief <run_id> to continue an existing Run.)
   └─ output_base = {run_dir}/outputs

Input Parsing:
   └─ Convert user input to structured format (reviewCommentsPath + paperPath + conferenceType)

Phase 1: Review Parsing & Classification
   └─ Ref: phases/01-review-parsing.md
      ├─ Tasks attached: Parse reviewer comments structure → Classify comments using Agy CLI → Extract sentiment and key concerns → Generate review-analysis.json
      └─ Output: reviewAnalysis, commentCategories, ${output_base}/review-analysis.json, ${output_base}/comment-classification.md

Phase 2: Multi-Perspective Discussion
   └─ Ref: p