A powerful metasearch CLI skill for AI agents to retrieve clean web, news, image, and video data using the duckse engine.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install duckse
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 duckse 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 Web Search (duckse)?
The Web Search skill, powered by the duckse engine, is a versatile metasearch tool designed for AI coding agents and developers who require real-time information retrieval. This Openclaw Skills component provides a robust interface to fetch data from multiple backends including Google, Bing, Brave, and DuckDuckGo. It enables agents to move beyond static training data by accessing current news, verifying facts, and sourcing multimedia content directly from the live web.
By leveraging the duckse CLI, users can receive results in either human-readable text or structured JSON, making it an ideal choice for both interactive research and automated data pipelines. It supports advanced filtering options such as time limits, regional settings, and safe search levels to ensure the most relevant and secure data is delivered.
Web Search (duckse) Use Cases
- Performing general web searches for technical documentation or tutorials.
- Monitoring current events and breaking news for specific research topics.
- Sourcing specific image or video assets using advanced media filters.
- Verifying facts or claims through multi-backend search results.
- Resolving final destination URLs from redirects to ensure link safety and accuracy.
How Web Search (duckse) Works
- The user or AI agent initiates a search by executing the
ducksecommand followed by a query string. - Optional flags such as
--type(text, news, images, videos, books) and--timelimitare processed to refine the scope. - The skill routes the request to the appropriate backend (e.g., Bing, Google, or AnnasArchive) based on the specified search type.
- The engine retrieves and parses results, optionally expanding URLs to their final destination using the
--expand-urlflag. - The data is returned as a formatted list or a JSON object for immediate use by the agent or developer.
Web Search (duckse) Setup
To integrate this skill into your environment, ensure that the duckse utility is installed. Use the following command to download and install it:
curl -sSL https://raw.githubusercontent.com/dwirx/duckse/main/scripts/install.sh | bash
After installation, verify the setup by running the help command:
duckse --help
If you are working in a local development environment without the global binary, you can run the skill using:
uv run python main.py "<query>"
Web Search (duckse) Data Schema & Taxonomy
This skill utilizes a dynamic request-response schema without maintaining a persistent local database. Results are structured as follows:
| Attribute | Description |
|---|---|
| Title | The headline or name of the search result |
| Link | The source URL (expanded if --expand-url is used) |
| Snippet | A brief summary or description of the content |
| Metadata | Additional data such as image resolution, video duration, or news timestamps |
When the --json flag is enabled, the skill outputs a standard JSON array of objects, facilitating easy parsing for Openclaw Skills automation.
Web Search (duckse) Advanced Features
- Multi-Backend Support: Toggle between engines like Google, Bing, Brave, Wikipedia, and Mojeek.
- Granular Media Filtering: Filter image results by size, color, or layout, and videos by resolution or duration.
- Academic & Literary Research: Access specialized backends like AnnasArchive for book-specific searches.
- Automation Friendly: Native JSON output support for seamless integration with other scripts and AI tools.
- Privacy & Control: Built-in support for proxies, custom timeouts, and regional result targeting.
SKILL.md
Loading
Related Openclaw Skills
AgentClear
A commerce layer for AI agents to discover and pay for APIs via natural language and micro-billing.
No Nonsense Tasks
A streamlined, SQLite-powered task manager designed for efficient CLI-based workflow organization and status tracking.
Wolt Orders 🍔
An automated assistant for Wolt.com that handles restaurant discovery, complex group ordering, and real-time delivery tracking.
Snail Mail - Operator Inbox
A slow-channel communication bridge that allows AI agents to leave non-urgent messages and decision requests for their human operators.
BangunAI Blog Manager
A professional content management skill for React-based digital gardens that automates MDX authoring and Obsidian-like documentation workflows.
DiaryBeast
A blockchain-integrated virtual pet and diary system designed specifically for AI agents to foster identity and creativity.
Featured*
