maestro-flow 0.5.31 → 0.5.33

package/.claude/skills/scholar-rebuttal-pro/phases/04-rebuttal-writing.md

@@ -0,0 +1,384 @@
+# Phase 4: Rebuttal Writing
+
+> **📌 COMPACT SENTINEL [Phase 4: rebuttal-writing]**
+> This phase contains 4 execution steps (Step 4.1 — 4.4).
+> If you can read this sentinel but cannot find the full Step protocol below, context has been compressed.
+> Recovery: `Read("phases/04-rebuttal-writing.md")`
+
+Generate structured rebuttal document using rebuttal-writer agent, apply conference-specific templates, integrate evidence and citations, optimize professional tone.
+
+## Objective
+
+- Load conference-specific rebuttal template
+- Call rebuttal-writer agent with strategy matrix and evidence
+- Write point-by-point responses with evidence integration
+- Optimize professional tone and persuasiveness
+- Generate final rebuttal.md and draft version
+
+## Execution
+
+### Helper Functions
+
+```javascript
+// Unified error handling for JSON file reading
+function safeReadJSON(filePath, phaseName) {
+  try {
+    const content = Read(filePath);
+    const data = JSON.parse(content);
+    return { success: true, data };
+  } catch (error) {
+    console.error(`[${phaseName}] Failed to read ${filePath}:`, error.message);
+    TodoWrite([
+      {"content": `${phaseName}`, "status": "failed"},
+      {"content": `  Error: ${error.message}`, "status": "failed"}
+    ]);
+    return { success: false, error: error.message };
+  }
+}
+```
+
+### Step 4.1: Load Strategy Matrix and Conference Template
+
+Load the strategy matrix from Phase 3 and select conference template:
+
+```javascript
+// Input from Phase 3
+const strategyMatrix = <from Phase 3 output>
+const evidenceMap = <from Phase 3 output>
+
+// Read strategy-matrix and evidence-references with error handling
+const matrixPath = ".workflow/.scratchpad/strategy-matrix.md"
+const evidencePath = ".workflow/.scratchpad/evidence-references.json"
+
+let strategyMatrixContent;
+try {
+  strategyMatrixContent = Read(matrixPath)
+} catch (error) {
+  console.error(`[Phase 4] Failed to read strategy-matrix.md:`, error.message);
+  TodoWrite([
+    {"content": "Phase 4: Rebuttal Writing", "status": "failed"},
+    {"content": `  Error: ${error.message}`, "status": "failed"}
+  ]);
+  throw error;
+}
+
+const evidenceResult = safeReadJSON(evidencePath, "Phase 4");
+if (!evidenceResult.success) {
+  console.error("Cannot proceed without evidence references");
+  return;
+}
+const evidenceData = evidenceResult.data;
+
+// Select conference template based on user preference
+const conferenceType = workflowPreferences.conferenceType
+
+const templateMap = {
+  'ML Conferences': 'neurips-icml-iclr',
+  'CV Conferences': 'cvpr-eccv-iccv',
+  'NLP Conferences': 'acl-emnlp',
+  'Generic': 'generic'
+}
+
+const templateId = templateMap[conferenceType] || 'generic'
+
+console.log(`
+Rebuttal Writing Configuration:
+  Conference Type: ${conferenceType}
+  Template: ${templateId}
+  Total Strategies: ${evidenceData.totalStrategies}
+  Evidence Found: ${evidenceData.evidenceSummary.found}
+  Requires Experiment: ${evidenceData.evidenceSummary.requiresExperiment}
+`)
+```
+
+### Step 4.2: Load Conference-Specific Template
+
+> **⚠️ CHECKPOINT**: Before proceeding, verify:
+> 1. This phase is TodoWrite `in_progress` (active phase protection)
+> 2. Full protocol (Step 4.X — 4.4) is in active memory, not just sentinel
+> If only sentinel remains → `Read("phases/04-rebuttal-writing.md")` now.
+
+Load template from claude-scholar or custom templates:
+
+```javascript
+// Template search paths
+const templatePaths = [
+  `G:/github_lib/claude-scholar/skills/review-response/references/${templateId}-template.md`,
+  `d:/ccws/.workflow/参考文档1/${templateId}-template.md`,
+  `d:/ccws/.workflow/参考文档1/discussion.md` // Fallback to generic discussion template
+]
+
+let templateContent = null
+let templatePath = null
+
+for (const path of templatePaths) {
+  try {
+    templateContent = Read(path)
+    templatePath = path
+    console.log(`Loaded template: ${path}`)
+    break
+  } catch (e) {
+    // Template not found, try next
+    continue
+  }
+}
+
+if (!templateContent) {
+  // Use built-in generic template
+  templateContent = `# Rebuttal to Reviewers
+
+We thank the reviewers for their thoughtful comments and constructive feedback. Below we address each concern point-by-point.
+
+## Response to Reviewer Comments
+
+[Point-by-point responses will be inserted here]
+
+## Summary
+
+We believe these revisions address all major concerns and strengthen the paper significantly. We are committed to incorporating all accepted suggestions in the final version.
+`
+  console.log("Using built-in generic template")
+}
+```
+
+### Step 4.3: Call Rebuttal-Writer Agent
+
+Dispatch rebuttal-writer agent to generate responses:
+
+```javascript
+// Prepare agent prompt
+const agentPrompt = `Generate a professional academic rebuttal document responding to reviewer comments.
+
+CONFERENCE TYPE: ${conferenceType}
+TEMPLATE: ${templateId}
+
+STRATEGY MATRIX:
+${strategyMatrixContent}
+
+EVIDENCE REFERENCES:
+${JSON.stringify(evidenceData, null, 2)}
+
+TEMPLATE STRUCTURE:
+${templateContent}
+
+REQUIREMENTS:
+1. Write point-by-point responses for each reviewer comment
+2. Use appropriate response type (Accept/Defend/Clarify/Experiment)
+3. Integrate evidence references with specific citations (Figure X, Table Y, Section Z, Page N)
+4. Maintain professional, respectful tone throughout
+5. For Accept strategies: acknowledge concern and commit to changes
+6. For Defend strategies: provide strong evidence and clear justification
+7. For Clarify strategies: address misunderstanding with clear explanation
+8. For Experiment strategies: outline planned experiments and expected outcomes
+9. Follow conference-specific conventions:
+   - ML Conferences: Emphasize novelty, theoretical soundness, experimental rigor
+   - CV Conferences: Focus on visual results, one-page limit, concise responses
+   - NLP Conferences: Address linguistic appropriateness, ethical considerations
+   - Generic: Universal academic tone and structure
+10. Optimize persuasiveness: lead with strongest evidence, acknowledge valid concerns
+
+OUTPUT FORMAT:
+- Markdown document following template structure
+- Clear section headers for each reviewer
+- Numbered responses matching comment IDs
+- Evidence citations in [brackets] or footnotes
+- Professional, confident but respectful tone
+
+TONE GUIDELINES:
+- Respectful: "We thank the reviewer for this insightful comment..."
+- Confident: "Our experiments demonstrate...", "The results clearly show..."
+- Collaborative: "We agree that...", "We will incorporate..."
+- Evidence-based: "As shown in Figure 3...", "Table 2 demonstrates..."
+- Avoid defensive language: "We believe the reviewer may have overlooked..."
+- Avoid dismissive language: "This concern is not valid..."
+`
+
+// Call rebuttal-writer agent
+const agentResult = Task({
+  subagent_type: "general-purpose", // Use general-purpose agent with full write capabilities
+  description: "Generate academic rebuttal document",
+  prompt: agentPrompt,
+  run_in_background: false
+})
+
+// Parse agent output
+const rebuttalDraft = agentResult.output || agentResult
+```
+
+### Step 4.4: Post-Process and Generate Final Rebuttal
+
+Post-process the draft and generate final files:
+
+```javascript
+// Post-processing: ensure all comments addressed
+const strategyIds = evidenceData.strategies.map(s => s.issueId)
+let missingResponses = []
+
+for (const issueId of strategyIds) {
+  if (!rebuttalDraft.includes(issueId)) {
+    missingResponses.push(issueId)
+  }
+}
+
+if (missingResponses.length > 0) {
+  console.warn(`Warning: Missing responses for: ${missingResponses.join(', ')}`)
+
+  // Generate placeholder responses for missing items
+  let additionalResponses = `\n\n## Additional Responses\n\n`
+
+  for (const issueId of missingResponses) {
+    const strategy = evidenceData.strategies.find(s => s.issueId === issueId)
+    additionalResponses += `### ${issueId}
+
+**Response**: [TODO: Address this comment based on ${strategy.responseType} strategy]
+
+`
+  }
+
+  rebuttalDraft += additionalResponses
+}
+
+// Add metadata header
+const rebuttalWithMetadata = `---
+title: Rebuttal to Reviewers
+conference: ${conferenceType}
+generated: ${new Date().toISOString()}
+template: ${templateId}
+total_comments: ${evidenceData.totalStrategies}
+---
+
+${rebuttalDraft}
+
+---
+
+## Revision Checklist
+
+- [ ] All reviewer comments addressed
+- [ ] Evidence citations verified
+- [ ] Tone reviewed for professionalism
+- [ ] Planned experiments documented
+- [ ] Accepted changes committed to
+- [ ] Word count within conference limits
+- [ ] Formatting matches conference requirements
+`
+
+// Write final rebuttal.md
+Write(".workflow/.scratchpad/rebuttal.md", rebuttalWithMetadata)
+
+// Write versioned draft
+const timestamp = new Date().toISOString().replace(/[:.]/g, '-').slice(0, 19)
+Write(`.workflow/.scratchpad/rebuttal-draft-v1-${timestamp}.md`, rebuttalWithMetadata)
+
+console.log(`
+Rebuttal Generation Complete:
+  Total Comments: ${evidenceData.totalStrategies}
+  Missing Responses: ${missingResponses.length}
+  Output Files:
+    - .workflow/.scratchpad/rebuttal.md
+    - .workflow/.scratchpad/rebuttal-draft-v1-${timestamp}.md
+`)
+```
+
+**User Review (if not auto mode)**
+
+```javascript
+if (!workflowPreferences.autoYes) {
+  const confirm = AskUserQuestion({
+    questions: [{
+      question: "Review the generated rebuttal. Proceed to quality validation?",
+      header: "Confirm",
+      multiSelect: false,
+      options: [
+        { label: "Proceed", description: "Continue to Phase 5 for quality validation" },
+        { label: "Revise", description: "Edit rebuttal manually before validation" },
+        { label: "Regenerate", description: "Regenerate with different template or tone" }
+      ]
+    }]
+  })
+
+  if (confirm["Confirm"] === "Revise") {
+    console.log("Please edit .workflow/.scratchpad/rebuttal.md and re-run this phase")
+    return
+  } else if (confirm["Confirm"] === "Regenerate") {
+    console.log("Please adjust preferences and re-run Phase 4")
+    return
+  }
+}
+```
+
+## Output
+
+- **Variable**: `rebuttalDraft` (generated rebuttal content)
+- **File**: `.workflow/.scratchpad/rebuttal.md` (final rebuttal document)
+- **File**: `.workflow/.scratchpad/rebuttal-draft-v1-{timestamp}.md` (versioned draft)
+- **TodoWrite**: Mark Phase 4 completed, Phase 5 in_progress
+
+## Next Phase
+
+Return to orchestrator, then auto-continue to [Phase 5: Quality Validation](05-quality-validation.md).
+
+## Conference-Specific Guidelines
+
+### ML Conferences (NeurIPS/ICML/ICLR)
+
+**Emphasis**:
+- Novelty and theoretical contributions
+- Experimental rigor and reproducibility
+- Comparison with state-of-the-art
+- Ablation studies and analysis
+
+**Tone**: Technical, precise, evidence-heavy
+
+**Common Concerns**:
+- Limited novelty → Emphasize unique contributions
+- Insufficient experiments → Commit to additional ablations
+- Missing baselines → Add comparisons or justify exclusions
+- Unclear methodology → Provide detailed clarification
+
+### CV Conferences (CVPR/ECCV/ICCV)
+
+**Emphasis**:
+- Visual results and qualitative analysis
+- One-page limit (concise responses)
+- Comparison with recent methods
+- Real-world applicability
+
+**Tone**: Concise, visual-focused, practical
+
+**Common Concerns**:
+- Limited visual results → Add qualitative comparisons
+- Missing comparisons → Commit to additional experiments
+- Unclear architecture → Provide detailed diagrams
+- Computational cost → Report efficiency metrics
+
+### NLP Conferences (ACL/EMNLP)
+
+**Emphasis**:
+- Linguistic appropriateness
+- Ethical considerations and bias
+- Dataset quality and diversity
+- Human evaluation
+
+**Tone**: Careful, ethically aware, linguistically precise
+
+**Common Concerns**:
+- Dataset bias → Acknowledge and commit to analysis
+- Limited languages → Justify scope or expand
+- Ethical issues → Address thoroughly with mitigation
+- Missing human eval → Commit to human studies
+
+### Generic Template
+
+**Emphasis**:
+- Clear, respectful communication
+- Evidence-based responses
+- Commitment to improvements
+- Professional academic tone
+
+**Tone**: Balanced, respectful, collaborative
+
+**Structure**:
+1. Thank reviewers
+2. Point-by-point responses
+3. Summary of changes
+4. Commitment to revisions