A Claude Code skill for consolidating multiple deepresearch results into a single comprehensive report using best practices for iterative processing.
The DeepResearch Integrator skill helps you efficiently merge information from multiple research files into one well-organized document. It follows a structured, iterative approach that processes one file at a time, ensuring accuracy and completeness while maintaining user control throughout the process.
- Iterative Processing: Handles one file at a time to ensure proper context understanding
- Structured Workflow: Creates and follows a detailed TODO list for all integration tasks
- User Supervision: Proposes changes and waits for approval before making modifications
- Organized Output: Creates a well-structured consolidated report with proper sections
- File Management: Automatically moves processed files to keep workspace clean
- Quality Control: Verifies completeness, consistency, and accuracy
- Claude Code installed and configured
- Git (for cloning this repository)
- Clone this repository:
git clone https://github.com/dakesan/deepresearch-integrator.git
cd deepresearch-integrator- Copy the skill to your Claude Code skills directory:
# For global installation
cp -r .claude/skills/deepresearch-integrator.md ~/.claude/skills/
# Or for project-specific installation
cp -r .claude/skills/deepresearch-integrator.md your-project/.claude/skills/- Verify the skill is available:
# In Claude Code, the skill should now appear in the skills list-
Prepare your research files: Place all deepresearch result files in a directory (e.g.,
research/) -
Invoke the skill in Claude Code:
/skill deepresearch-integrator
- Provide the task: Tell Claude what you want to consolidate:
Consolidate all the deepresearch results in the research/ directory into consolidated-report.md
-
Review the plan: Claude will:
- List all source files found
- Create a TODO list
- Present the integration strategy
-
Approve and monitor:
- Review and approve the plan
- Claude will process each file one by one
- Approve each integration step
-
Final review: After all files are processed, review the consolidated report
You: /skill deepresearch-integrator
Claude: DeepResearch Integrator skill loaded. Ready to consolidate research files.
You: Consolidate all files in examples/deepresearch-results/ into examples/final-report.md
Claude: I found 3 deepresearch files to process:
1. research-001-ai-trends.md
2. research-002-quantum-computing.md
3. research-003-cybersecurity.md
I'll create a consolidated report with the following structure:
- Executive Summary
- AI Trends (from research-001)
- Quantum Computing (from research-002)
- Cybersecurity (from research-003)
- Summary and References
Shall I proceed?
You: Yes, go ahead
Claude: [Processes files one by one, updating TODO list, and creating the consolidated report]
deepresearch-integrator/
├── .claude/
│ └── skills/
│ └── deepresearch-integrator.md # The skill file
├── examples/
│ ├── deepresearch-results/ # Sample research files
│ │ ├── research-001-ai-trends.md
│ │ ├── research-002-quantum-computing.md
│ │ └── research-003-cybersecurity.md
│ └── processed/ # Processed files moved here
├── AI_workflow_instructions.md # Workflow principles
└── README.md # This file
The skill follows a systematic workflow:
- Lists all source files to be processed
- Identifies or creates the target consolidated report
- Creates a detailed TODO list
- Presents the plan for user approval
For each file:
- Reads the content completely
- Analyzes key information and findings
- Proposes specific changes to the main report
- Waits for user approval
- Updates the consolidated report
- Moves the file to
processed/directory - Updates the TODO list
- Creates
processed/directory for completed files - Keeps workspace clean and organized
- Prevents accidental re-processing
- Reviews entire consolidated report
- Checks completeness, consistency, and accuracy
- Generates summary statistics
- Provides final recommendations
- ✅ Merge similar topics under common headings
- ✅ Preserve attribution from sources
- ✅ Handle conflicts by presenting multiple perspectives
- ✅ Maintain proper heading hierarchy
- ✅ Add metadata (dates, sources, context)
- ✅ Verify no files are skipped
- ✅ Avoid duplicate information
- ✅ Use consistent markdown formatting
- ✅ Preserve links and references
- ✅ Add cross-references between sections
You can customize the skill behavior by editing .claude/skills/deepresearch-integrator.md:
- Modify the report structure template
- Adjust the processing order logic
- Customize the file organization strategy
- Change the consolidation rules
- Ensure the skill file is in the correct directory
- Check that the file name matches:
deepresearch-integrator.md - Restart Claude Code if necessary
- Verify file paths are correct
- Check file permissions
- Ensure files are valid markdown format
- Review the TODO list to see current progress
- Check the
processed/directory for completed files - Restart from the last completed file if needed
See the examples/ directory for:
- Sample deepresearch result files
- Expected output structure
- Common use cases
Contributions are welcome! Please:
- Fork the repository
- Create a feature branch
- Make your changes
- Submit a pull request
[Add your license here]
Created for efficient deepresearch result consolidation using Claude Code.
DeepResearch Integratorスキルは、複数のリサーチファイルを1つの包括的なレポートに統合するためのClaude Codeスキルです。
- 段階的処理: 1つずつファイルを処理し、正確性を確保
- 構造化ワークフロー: TODOリストで進捗を管理
- ユーザー監督: 変更前に承認を求める
- 整理された出力: 適切なセクション分けされたレポート
- ファイル管理: 処理済みファイルを自動的に移動
- 品質管理: 完全性、一貫性、正確性を検証
git clone https://github.com/dakesan/deepresearch-integrator.git
cd deepresearch-integrator
cp -r .claude/skills/deepresearch-integrator.md ~/.claude/skills/- リサーチファイルをディレクトリに配置
- Claude Codeで
/skill deepresearch-integratorを実行 - 統合タスクを指示
- プランを確認・承認
- 各ステップを確認しながら統合を進める
詳細は上記の英語ドキュメントをご参照ください。