Deep Research Agent
Research Search & ProductivityInstall Command
npx clawhub@latest install deepresearchagent
Installation Guide
Make sure Node.js 22+ and OpenClaw are installed. Run openclaw --version in your terminal to verify.
Run the install command above in your terminal. ClawHub will automatically download and install Deep Research Agent to the ~/.openclaw/skills/ directory.
Run openclaw skills list to check your installed skills and confirm Deep Research Agent appears in the list.
Follow the configuration instructions in the description below to add skill settings to ~/.config/openclaw/openclaw.json5.
~/.openclaw/skills/ or the skills/ directory in your project root. Make sure the folder contains a SKILL.md file.
Detailed Description
Deep Research Agent is an advanced research skill that simulates a human researcher's workflow: first defining the research question, then iteratively searching, reading, analyzing, and cross-validating, ultimately producing a comprehensive research report with cited sources.
Core Features
- Multi-Round Iterative Search: Rather than returning after a single search, it automatically generates follow-up queries based on initial results, progressively going deeper
- Cross-Source Validation: Compares information across multiple sources, flagging the credibility of each piece of information
- Structured Report Output: The final report includes a summary, body analysis, key findings, and a complete list of citations
- Transparent Research Process: Shows you the current progress, what has been searched, and what has been discovered during the research process
Configuration
{
skills: {
deepresearchagent: {
maxIterations: 5, // Maximum search iteration rounds
sourcesPerQuery: 3, // Sources read per search
outputFormat: "markdown" // markdown or html
}
}
}
Use Cases
- Conducting comprehensive research and comparative analysis of a technical solution
- Understanding the latest developments and trends in a field
- Writing research articles or reports that require cited sources
- In-depth analysis of a product or competitor's market performance
Notes
Due to multi-round searching and webpage reading, a single research task consumes a significant number of tokens and time (typically 2-5 minutes). It is recommended to use this alongside the Web Search skill for best results. Research depth and breadth can be controlled via the maxIterations parameter.