A powerful multi-backend web search skill that enables AI agents to perform real-time information retrieval and recursive deep research.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install xyz
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 xyz 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 xyz: Multi-Engine AI Web Search & Research?
xyz is a versatile network search skill designed for AI agents to provide real-time internet access capabilities. By integrating with Openclaw Skills, this tool supports five major search engine backends, including DuckDuckGo, Tavily, Bing, Google, and SearXNG. It is specifically engineered to help agents find the latest data, verify facts, and collect structured information such as titles, links, and snippets directly from the web.
Beyond simple queries, xyz excels at deep research by combining search results with LLM reasoning. This allows for multi-round iterations where the agent analyzes findings from one search to intelligently formulate the next, ensuring comprehensive coverage of complex topics. It is a critical component for developers building autonomous agents that require up-to-the-minute external knowledge.
xyz: Multi-Engine AI Web Search & Research Use Cases
- Complex Subject Research: Conducting deep dives into technical or academic topics through multiple search iterations.
- Real-time Information Analysis: Generating reports based on the latest news, market trends, or product releases.
- Multi-Perspective Exploration: Investigating a single theme from various angles and dimensions to ensure unbiased data collection.
- Systematic Investigations: Gathering evidence and documentation for competitive analysis or academic literature reviews within the Openclaw Skills framework.
How xyz: Multi-Engine AI Web Search & Research Works
- The AI agent receives a user query and determines the optimal search provider and result depth.
- The xyz script executes a search request via the specified backend (defaults to DuckDuckGo for free, keyless access).
- Results are returned as structured data containing titles, URLs, and descriptive snippets.
- For deep research tasks, an LLM (such as GPT-4o or DeepSeek) analyzes the initial findings to identify information gaps.
- The system autonomously generates new search queries to fill those gaps, repeating for a specified number of rounds.
- Finally, all collected data points are synthesized into a comprehensive report or structured JSON output.
xyz: Multi-Engine AI Web Search & Research Setup
To get started with this skill in your environment, ensure you have Python 3.10+ installed. Most functionality relies on standard libraries, making it highly portable for Openclaw Skills users.
# Optional: Install the duckduckgo_search library for enhanced results
pip install duckduckgo_search
# Configure your preferred search provider via environment variables
export SEARCH_PROVIDER="duckduckgo"
export TAVILY_API_KEY="your_api_key_here"
export BING_API_KEY="your_api_key_here"
Run a basic search command:
python scripts/xyz.py "latest AI developments" --max 5
xyz: Multi-Engine AI Web Search & Research Data Schema & Taxonomy
The xyz skill organizes output into structured formats to facilitate machine readability. Below is the schema for JSON output:
| Property | Description |
|---|---|
query |
The original search query processed by the skill |
count |
Total number of results retrieved |
results |
An array of objects containing title, url, and snippet |
results.title |
The headline or title of the webpage |
results.url |
The direct link to the source |
results.snippet |
A brief summary of the content found at the URL |
xyz: Multi-Engine AI Web Search & Research Advanced Features
- Adaptive Search Strategy: Uses LLM analysis after each round to pivot search directions dynamically based on discovered info.
- Parallel Search Capabilities: Supports up to 10 concurrent search threads to maximize throughput and reduce latency.
- Redundancy Filtering: Built-in state management tracks previously searched topics to prevent duplicate efforts.
- Deep Reasoning Integration: Leverages deepseek-reasoner for advanced analytical synthesis of gathered web data.
- Progress Visualization: Real-time tracking of search depth and execution status for long-running research tasks in Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
File Reader
A core utility that enables AI agents to read, view, and process the contents of various text-based files including source code and configuration files.
Audio Editor
A comprehensive audio manipulation toolkit for trimming, volume control, format conversion, and audio extraction from video files.
Plaid
A powerful command-line integration for managing bank accounts and financial data via the Plaid platform.
Claw MBTI Personality Assessment
An automated assessment tool designed to identify the MBTI personality profile of AI agents through a 60-question psychometric evaluation.
Feishu Messaging & Documents
A comprehensive automation skill for sending Feishu messages, managing group chats, and uploading media through the Lark Open Platform API.
Baserow CLI
Interact with Baserow databases to manage rows, tables, and fields directly from your AI agent using Openclaw Skills.
Featured*
