Can I control the depth of the research?

Absolutely. You can use the --max-iterations parameter to control how many cycles of search and reflection the agent performs before finalizing the report.

What API keys are required for setup?

You will need an LLM API key (such as OpenAI or Anthropic) and a Search API key (such as Tavily) to power the reasoning and search capabilities.

What search engines does this skill support?

It primarily uses Tavily by default but is compatible with Brave Search, DuckDuckGo, and SerpAPI for flexible web data collection.

Does it provide citations for its findings?

Yes, the skill produces comprehensive reports that include iterative findings and specific source citations for all data points found during the research.

LangChain Deep Research

Name: LangChain Deep Research
Author: closedloop-technologies

byclosedloop-technologies

•

ウェブスクレイピングとデータ収集

Conducts deep, iterative web research to generate comprehensive reports with verified citations and source tracking.

概要

The LangChain Deep Research skill transforms Claude into a sophisticated research assistant capable of performing multi-step web investigations. It goes beyond simple search queries by using an iterative process that reflects on findings, identifies knowledge gaps, and refines searches until a comprehensive understanding is reached. This skill is particularly useful for technical due diligence, market analysis, or deep-dives into complex topics where accuracy, source attribution, and depth are paramount.

主な機能

Support for multiple search engines including Tavily, Brave, and DuckDuckGo
Automated comprehensive report generation with markdown support
Iterative search cycles with reflection and gap analysis
Configurable research depth and iteration limits
3 GitHub stars
Multi-source citation tracking and bibliography creation

ユースケース

Competitive market analysis and trend reporting
Academic and professional literature reviews with cited sources
Technical due diligence for emerging software technologies