liangdabiao/Claude-Code-Deep-Research-main

Registry

Summary

A project demonstrating how to implement deep research step-by-step using the Claude Skill code agent framework. It serves as a practical guide and comparison tool to evaluate the effectiveness of different agent frameworks by examining their approach to deep research, a primary application scenario for agents.

Key features

• Step-by-step implementation of deep research
• Uses the Claude Skill code agent framework
• Serves as a framework comparison tool
• Focuses on a primary agent application scenario
• Aims to be both powerful and simple to use

Use cases

• Learning agent framework implementation
• Comparing different AI agent frameworks
• Conducting deep research tasks with AI
• Evaluating framework capabilities for research
• Practical guide for building research agents

README excerpt

# Claude Code Deep Research Agent A sophisticated multi-agent research framework that implements OpenAI's Deep Research and Google Gemini's Deep Research capabilities using Claude Code's native features. ## Overview This project leverages Claude Code's Skills and Commands system to conduct comprehensive, citation-backed research through: - **Graph of Thoughts (GoT) Framework** - Intelligent research path management with graph-based reasoning - **7-Phase Deep Research Process** - Structured methodology for quality research - **Multi-Agent Architecture** - Parallel research agents with specialized roles - **Citation Validation System** - A-E source quality ratings with chain-of-verification ## Quick Start ### Prerequisites - Claude Code CLI installed - Active Claude Code account with API access ### Installation 1. Clone this repository: ```bash git clone <repository-url> cd Claude-Code-Deep-Research-main ``` 2. The Skills and Commands are already configured in `.claude/` directory ### Basic Usage The simplest way to conduct deep research: ```bash /deep-research [your research topic] ``` **Example:** ```bash /deep-research AI applications in clinical diagnosis ``` This single command will: 1. Ask clarifying questions to refine your research needs 2. Create a structured research plan 3. Deploy multiple parallel research agents 4. Synthesize findings into a comprehensive report 5. Validate all citations 6. Output results to `RESEARCH/[topic]/` directory ## Advanced Usage ### Step-by-Step Research Workflow For more control over the research process: #### 1. Refine Your Question ```bash /refine-question Should I use WebAssembly for my project? ``` The Question Refiner will ask 5-6 clarifying questions about: - Specific focus areas - Output format requiremen

Topics