An advanced multi-agent orchestration workflow that automates deep, systematic research by splitting complex objectives into parallel sub-tasks.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install deep-research-skill
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-skill 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 Multi-Agent Workflow?
The Deep Research skill transforms the concept of depth-first investigation into a repeatable, parallelized production pipeline. By utilizing a master controller to clarify goals and dispatch independent child processes via Claude Code in non-interactive mode, it ensures high-density information gathering without manual intervention. This skill is optimized for Openclaw Skills environments where reliability, traceability, and comprehensive data synthesis are paramount, delivering structured Markdown reports as standalone files rather than simple chat fragments.
Deep Research Multi-Agent Workflow Use Cases
- Systematic web exploration and data collection across multiple industry sources.
- Competitive analysis and market research requiring deep dives into various document sets.
- Long-form writing and evidence integration for technical whitepapers or academic reports.
- Batch link processing and dataset shard retrieval for large-scale AI analysis.
How Deep Research Multi-Agent Workflow Works
- The main agent analyzes high-level research goals and breaks them down into parallel sub-objectives like specific links or topic clusters.
- Independent child processes are initialized using non-interactive CLI commands with strictly defined tool permissions.
- Sub-agents perform targeted web scraping and data extraction using prioritized search skills and MCP tools.
- Raw data and logs are saved to a dedicated local directory structure to ensure persistence and traceability.
- A script-driven aggregation phase merges sub-reports into a unified draft while maintaining citation integrity.
- The main agent performs a multi-stage refinement process to polish the final report, ensuring insight-driven conclusions.
Deep Research Multi-Agent Workflow Setup
To utilize this workflow within Openclaw Skills, ensure your environment supports parallel execution and has search MCPs enabled.
# Ensure MCP tools like Firecrawl or Exa are configured in your agent settings
# The skill will automatically create and manage the .research/ directory
# Trigger the workflow by asking for "Deep Research" on a specific topic
Deep Research Multi-Agent Workflow Data Schema & Taxonomy
This skill organizes all research artifacts within a semantic .research/<task-name>/ directory to prevent data overlap:
| Directory | Purpose |
|---|---|
raw/ |
Stores original search results and cached data fetching responses |
prompts/ |
Contains the generated instructions sent to child agents |
logs/ |
Real-time execution logs for the dispatcher and individual sub-tasks |
child_outputs/ |
Markdown fragments produced by individual sub-agents |
polished_report.md |
The final synthesized and high-quality research deliverable |
Deep Research Multi-Agent Workflow Advanced Features
- Multi-process parallelization using GNU Parallel or shell background jobs for scaling up to 15+ concurrent tasks.
- Automated timeout handling (up to 900s) and failure isolation to ensure overall workflow resilience.
- Tiered tool priority logic: Internal Skills > MCP (Firecrawl/Exa) > WebFetch/WebSearch.
- Section-by-section iterative polishing to maintain narrative consistency and depth in long-form outputs.
SKILL.md
Loading
Related Openclaw Skills
Claude Code Agent Skill
An automated coding agent for Claude Code that handles the end-to-end lifecycle of features, bugfixes, and refactors within isolated worktrees.
Xiaohongshu Login
A specialized workflow for managing Xiaohongshu authentication and QR code login via Openclaw Skills.
Mail Parser
Automatically transform plain text emails into clean, actionable JSON objects for seamless automation.
Genviral Partner API Skill
A powerful CLI wrapper and automation suite for the Genviral Partner API that handles the entire content lifecycle from AI generation to performance tracking.
Nanobanana Image Generation
A versatile image generation and editing tool powered by Google Gemini API for seamless AI-driven visual creation.
xfetch
A cookie-based command-line interface for scraping and retrieving comprehensive data from X/Twitter without requiring API keys.
Featured*
