An autonomous research agent for OpenClaw that searches, synthesizes, and stores findings with full source tracking and vectorization.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install auto-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-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-Research Skill?
The Auto-Research Skill is a powerful addition to the Openclaw Skills library, designed to transform your AI agent into a sophisticated research analyst. It automates the entire lifecycle of information gathering—from initial web queries via the Brave Search API to the generation of structured Markdown briefings. By integrating this skill, users can bypass manual searching and move straight to analysis, leveraging an agent that understands the difference between a quick fact-check and a deep-dive strategic investigation.
This skill is built for developers and researchers who need high-fidelity data stored in accessible formats. It seamlessly bridges the gap between the live web and your personal knowledge management system by outputting results directly to Obsidian. Furthermore, it ensures long-term utility by vectorizing findings into Qdrant, allowing for semantic retrieval across your entire research corpus within the Openclaw Skills ecosystem.
Auto-Research Skill Use Cases
- Generating executive briefings for market trends and emerging technologies.
- Conducting deep-dive technical research on software architecture or new programming frameworks.
- Automating daily or weekly research digests on specific industry topics via cron jobs.
- Building a searchable, vectorized knowledge base of web findings using Openclaw Skills.
- Rapidly verifying facts or exploring new domains with tiered research depths.
How Auto-Research Skill Works
- The skill receives a natural language prompt or CLI command specifying a research topic and a depth level (quick, standard, or deep).
- It orchestrates a search sequence using the Brave Search API, first checking a Redis or local cache to optimize performance and reduce API costs.
- The agent evaluates the credibility of retrieved sources, categorizing them into confidence tiers based on domain authority.
- It synthesizes the raw data into a structured Markdown template including summaries, key findings, and actionable recommendations.
- The final briefing is saved to the configured Obsidian vault and concurrently processed by a vectorization script.
- Semantic chunks are stored in a Qdrant collection, making the research available for future RAG (Retrieval-Augmented Generation) queries within Openclaw Skills.
Auto-Research Skill Setup
To integrate this skill into your environment, ensure you have the prerequisite dependencies and configure your environment variables.
# Navigate to your skills directory and clone the repository
cd clawhub-skills
git clone https://github.com/openclaw/auto-research.git
# Set required environment variables
export BRAVE_API_KEY="your_api_key_here"
export OBSIDIAN_VAULT="/path/to/your/vault"
export QDRANT_URL="http://your-qdrant-server:6333"
# Run a test research task
./auto-research/research.sh "Openclaw Skills architecture" standard
Auto-Research Skill Data Schema & Taxonomy
The Auto-Research Skill maintains a rigorous data schema to ensure all findings are traceable and actionable.
1. Markdown Briefing Structure
| Section | Content Description |
|---|---|
| Executive Summary | A high-level overview of the research topic (3-5 sentences). |
| Key Findings | Bulleted highlights of critical data points discovered. |
| Sources | A list of URLs with access dates and confidence ratings. |
| Confidence Rating | An assessment of source reliability (High, Medium, or Low). |
2. Vector Metadata
Each entry in the Qdrant web_research collection includes:
source_url: The origin of the information.topic_tags: Semantic labels for categorization.confidence_score: Numeric value based on source quality assessment.access_date: Timestamp of the research event.
Auto-Research Skill Advanced Features
- Smart Caching: Leverages Redis for high-speed result caching with a 24-hour TTL to prevent redundant web scraping.
- Automated Quality Assessment: Logic-based filtering that prioritizes .edu, .gov, and industry-leading publications.
- Tiered Depth Control: Adjustable source counts ranging from 5 (quick) to 10+ (deep dive) depending on the user's needs.
- Standalone Automation: The orchestrator scripts can be run independently of the main agent for automated weekly digests.
- Semantic Retrieval: Native integration with vector search tools to query your research history via Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Openclaw YouTube Transcript
A high-performance utility for Openclaw agents to extract text transcripts and subtitles directly from YouTube URLs without the need for local audio processing.
Fasting Tracker Pro
An intelligent health coach for calculating fasting windows and understanding metabolic science within the AI agent ecosystem.
Dropshipping Mentor Nick
An AI-powered business mentor designed to help entrepreneurs launch, audit, and scale dropshipping and e-commerce stores.
Day Trading & Investing Mentor (Pro)
A professional AI trading floor analyst and mentor designed to master price action, technical analysis, and rigorous risk management logic.
Colony Engagement
A comprehensive toolkit for AI agents to interact with thecolony.cc, facilitating collaborative intelligence through authenticated posting, voting, and engagement tracking.
Construction PM
A production-tested construction project management toolkit designed for AI agents to track jobs from lead to final payment.
Featured*
