grapeot

Deep Research Agent — Development skill for Claude Code

Development community

An Agentic Deep Research Assistant similar to Gemini and OpenAI Deep Research.

How to install Deep Research Agent

This entry records only its repository, not the path inside it, so there is no exact command to give. Open grapeot/deep_research_agent and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Deep Research Agent does

An Agentic Deep Research Assistant similar to Gemini and OpenAI Deep Research.

Alternatives in Development

  • OpenAI Codex CLI — (Rust implementation) 67.8k ★
  • Sub2api — Sub2API-CRS2 一站式开源中转服务,让 Claude、Openai 、Gemini、Antigravity订阅统一接入,支持拼车共享,更高效分摊成本,原生工具无缝使用 13.6k ★
  • Agy2api — 将 Antigravity 转换为 OpenAI 、GEMINI 和 Claude API 接口 5.3k ★

README

Deep Research Agent

A document-driven agentic AI research system that helps conduct comprehensive analysis through persistent context management and tool integration.

Core Philosophy

  1. Document-Centric Memory: Uses persistent documents to maintain context and track progress, solving the fundamental context window limitation of language models.
  2. Structured Communication: Documents all information in a shared scratchpad, ensuring no critical information is lost.
  3. Tool Augmentation: Leverages specialized tools for up-to-date information gathering and analysis.
  4. User Agency: Acts as a collaborative partner, keeping users in control through clear documentation and decision points.

Getting Started

# Install dependencies
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

# Install Playwright browsers
playwright install chromium

# Set API keys
export OPENAI_API_KEY=your_openai_api_key
export ANTHROPIC_API_KEY=your_anthropic_api_key

# Run a research query
python3 deep_research_agent.py "your research query"

Available Tools

  1. Web Search & Analysis:
    • Semantic search using DuckDuckGo
    • Dynamic web content scraping with Playwright for JavaScript-heavy sites
    • Content analysis with source attribution
  2. File Operations: Create/edit files and execute scripts with user confirmation
  3. Package Management: Install and manage Python packages with user confirmation
  4. Terminal Commands: Execute system commands with user confirmation

Example Usage

Analyzing NVIDIA's recent stock performance:

python3 deep_research_agent.py "Perform a detailed analysis on the recent trend of NVDA stock. How did the stock price change? What might have caused it? How about the market sentiment?"

The agent:

  1. Creates a research plan in scratchpad.md
  2. Generates and executes analysis script (with user confirmation):
    Confirm execution of command: python3 nvda_analysis.