A sophisticated AI research leader that orchestrates specialized agents to decompose, execute, and synthesize complex information-gathering tasks.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install deep-research-prime
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install deep-research-prime using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Deep Research Agent?
The Deep Research Agent is a high-level coordination skill designed for Openclaw Skills that transforms the way users conduct investigations. Unlike standard search tools, this skill acts as a Deep Research Leader Agent, prioritizing strategic planning and task delegation. It treats user input as the absolute source of truth, ensuring that even the most obscure or recent terms are verified rather than dismissed. By managing a team of specialized subordinate searchers, it provides a comprehensive overview of any given topic with academic-level rigor.
This skill is built for users who require more than a surface-level summary. It excels at breaking down complex queries into orthogonal sub-tasks, ensuring that no stone is left unturned across the general internet, social media, financial markets, and academic databases. The agent ensures the integrity of information by maintaining a full audit trail of research logs and reports, culminating in a synthesized document with professional inline citations.
Deep Research Agent Use Cases
- Comprehensive market analysis for emerging technologies and unreleased products.
- Academic literature reviews leveraging authoritative databases like Google Scholar.
- Real-time financial monitoring and news aggregation for specific tickers or sectors.
- Social media sentiment tracking and news gathering via specialized Twitter searchers.
- Competitive intelligence reports that expand beyond user-provided examples using heuristic reasoning.
How Deep Research Agent Works
- The agent identifies core entities in the user request and performs a mandatory shallow search to verify definitions and existence.
- It decomposes the primary goal into 2 to 6 independent, parallel sub-tasks to ensure breadth and depth.
- Tasks are delegated to specialized subordinates: Internet Searcher, Twitter Information Searcher, Financial Information Searcher, or Academic Research Searcher.
- The leader agent performs a comprehensive review of all returned documents and logs to ensure no data gaps exist.
- All verified information is synthesized into a final report using standardized markdown citations and a complete reference list.
Deep Research Agent Setup
To integrate this skill into your workflow, ensure your environment is configured for Openclaw Skills. Install the package using the following command:
openclaw install deep-research
Once installed, verify that the required search agent tools (Internet, Twitter, Finance, Academic) are active in your available tools manifest to allow for seamless delegation.
Deep Research Agent Data Schema & Taxonomy
The Deep Research Agent organizes its output through a structured hierarchy of files and metadata to ensure full transparency:
| File Type | Description | Format |
|---|---|---|
| Summary Report | A synthesized final document with inline [ref] citations. | Markdown |
| Research Logs | Step-by-step logs of every search query and agent interaction. | Text/Log |
| Subordinate Reports | Raw reports generated by specialized domain searchers. | Markdown |
| Reference List | A numbered list of all source URLs cited in the research. | Markdown |
All files are passed through the submit_result tool as a complete package to prevent information loss.
Deep Research Agent Advanced Features
- Heuristic Expansion: Automatically identifies and researches relevant topics related to user examples using independent reasoning.
- Mandatory Parallelism: Ensures maximum efficiency by strictly enforcing a multi-task decomposition strategy for every query.
- Single-Round Logic: Optimized for high-velocity results by completing global research within a single, massive parallel execution round.
- Verification-First Protocol: Prevents hallucinations by mandating entity verification before any planning or execution occurs.
- Multi-Domain Delegation: Routes specific sub-tasks to specialized agents optimized for financial, academic, or social media environments.
SKILL.md
Loading
Related Openclaw Skills
Data Analyst
An automated agent skill designed to execute complex data analysis tasks and process uploaded files directly within a local working directory.
Coding Skill for Dynamic Dashboards
Automate the end-to-end process of database schema design, data insertion, and the generation of real-time dynamic HTML dashboards.
Vultisig
A comprehensive SDK for managing self-custodial MPC crypto vaults and multi-chain transactions across 36+ blockchains without seed phrases.
openclaw-update
A specialized skill for automating workspace backups, system restores, and dynamic environment updates.
Deep Research Strategy
A strategic lead agent that orchestrates specialized subordinates to conduct multi-dimensional, deep-dive research with mandatory verification and parallel task decomposition.
Deep Strategy
A high-level strategic orchestrator that decomposes complex tasks and delegates them to specialized AI agents for efficient, economical execution.
Featured*
