web-search

What This Skill Does

The web-search skill integrates the duckse CLI tool into the OpenClaw ecosystem, providing an advanced, versatile interface for performing metasearches across the internet. By leveraging the power of DuckDuckGo's infrastructure through a clean terminal interface, this skill enables users to retrieve text results, breaking news, imagery, video resources, and digital literature. It acts as a bridge between your agent's reasoning capabilities and the vast, real-time data of the World Wide Web. Whether you need to verify facts, fetch trending news articles, or find specific media assets, duckse provides granular control over the query process, including support for different backends, regional filtering, and various output formats like raw text or parseable JSON.

Installation

To begin using this skill, ensure that the duckse binary is present on your system. You can verify this by running duckse --help in your terminal. If the command is not found, you can quickly install it using the official script via: curl -sSL https://raw.githubusercontent.com/dwirx/duckse/main/scripts/install.sh | bash. Once installed, you can integrate it into your agent environment using the command: clawhub install openclaw/skills/skills/dwirx/duckse. This installation ensures the skill is registered and ready for invocation by the OpenClaw agent.

Use Cases

This skill is highly versatile and fits into several workflows:

• Academic Research: Quickly locate books from diverse repositories like Anna's Archive using the --backend flag.
• Content Curation: Find relevant media, such as specific mountain landscapes or tutorial videos, with advanced filtering for duration, resolution, and style.
• Real-time Monitoring: Stay updated on global events by filtering news articles by time (days, weeks, months) and region.
• Technical Development: Integrate live web data into your automated pipelines by requesting JSON output and resolving final URLs with the --expand-url flag for deeper analysis.

Example Prompts

1. "Search for the latest breakthroughs in quantum computing from this past week and give me the top 5 news articles."
2. "Find me some high-resolution wide-layout images of architectural designs for inspiration."
3. "Look up the book 'The Sea-Wolf' by Jack London using the annasarchive backend and provide the results in JSON format."

Tips & Limitations

For the best results, always specify the type of search you are conducting (news, images, videos) to reduce noise. When building automated scripts, prefer the --json flag to ensure the agent can parse and utilize the output programmatically without formatting issues. Be aware that the availability of specific backends varies by category (e.g., use 'annasarchive' for books, but 'bing' or 'google' for text searches). Always check your network connectivity, as the tool requires active web access to function.