An advanced research skill that autonomously decomposes queries, performs parallel searches, and verifies information across multiple independent sources.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install auto-deep-research
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 auto-deep-research 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 Auto Deep Research?
Auto Deep Research is a professional-grade research orchestration tool designed for the Openclaw Skills ecosystem. It goes beyond basic search queries by implementing a loop-based investigative logic that ensures data accuracy and depth. The skill is specifically engineered to handle complex investigations where verification is critical, ensuring that every conclusion is backed by multiple independent sources.
By leveraging this skill within the Openclaw Skills framework, developers can automate the tedious process of manual research. It handles everything from sub-problem decomposition to final report generation, providing a structured and reliable way to synthesize large amounts of web data into actionable technical insights.
Auto Deep Research Use Cases
- Investigating complex technical concepts or emerging frameworks like React Server Components.
- Performing competitive analysis or tool comparisons with evidence-based conclusions.
- Tracking evolving news stories or analyzing market trends across various domains.
- Verifying technical claims by cross-referencing official documentation with community-driven data.
How Auto Deep Research Works
- Dynamically decomposes the user query into 3 initial sub-problems to ensure comprehensive coverage.
- Executes parallel search operations for all pending sub-problems to maximize efficiency.
- Conducts serial page reading and content extraction to prevent context overlap and maintain data purity.
- Merges and deduplicates findings into a structured JSON memo for easy agent processing.
- Evaluates the current evidence base to determine if new sub-problems need to be added or if the investigation is complete.
- Generates a final, structured Markdown report including conclusions, uncertainties, and references.
Auto Deep Research Setup
To integrate this into your environment, ensure you have the necessary API keys configured. Like many Openclaw Skills, it supports multiple search providers.
# Configure Search API (Tavily recommended)
export TAVILY_API_KEY="your-key-here"
# Configure Reader API (Jina Reader recommended)
export JINA_API_KEY="your-key-here"
Trigger the research process directly from your agent interface:
/auto-deep-research [Your Research Query]
Auto Deep Research Data Schema & Taxonomy
The skill maintains a strict file hierarchy within the output/{topic-slug}/ directory to manage the research state effectively:
| File | Format | Purpose |
|---|---|---|
| state.json | JSON | Tracks iteration rounds, sub-problem status, and research progress. |
| memo.json | JSON | Stores structured findings, key points, and conflict markers. |
| sources.json | JSON | Maintains a registry of all URLs, source priorities, and metadata. |
| report.md | Markdown | The final synthesized report with evidence-backed conclusions. |
Auto Deep Research Advanced Features
- Multi-Source Verification: Enforces a rule where conclusions must be supported by at least two independent sources.
- Conflict Detection: Automatically flags and highlights contradictions between different sources (e.g., Source A vs Source B).
- Dynamic Task Queue: Can autonomously add new sub-problems mid-research if a critical new concept is discovered.
- Priority-Based Sourcing: Categorizes sources by trust levels, such as P1 for official documentation and P4 for general forums.
- Automated De-duplication: Prevents redundant information from bloating the final report through intelligent data merging.
SKILL.md
Loading
Related Openclaw Skills
Binance Onchain Pay Open API
A professional-grade Openclaw Skills integration for the Binance Onchain Pay Open API, facilitating secure fiat-to-crypto transactions and automated on-chain payments.
Meme Rush
A real-time intelligence skill for tracking meme token lifecycles and AI-driven market narratives across Solana and BSC launchpads.
Car Rental Self-Drive Planner
An AI-driven assistant that generates visually structured HTML guides for car rental preparation, local compliance, and optimized self-drive itineraries.
Test-time Compute Guide
A technical framework for implementing test-time compute and chain-of-thought reasoning to improve the accuracy and depth of AI model outputs.
Binance Fiat
A powerful tool for managing fiat currency transactions and history on Binance through authenticated API requests.
Binance P2P Trading Assistant
A specialized tool for querying Binance P2P market data and personal order history using natural language queries.
Featured*
