Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
v4.0: Added verify_claim.py for multi-source cross-validation. Now tells you how much to trust the answer with VERIFIED/LIKELY_TRUE/UNCERTAIN/LIKELY_FALSE verdicts. Multi-engine parallel search + 4-tier anti-bot cascade + authority-weighted scoring.
What This Skill Does
Free, private web search for OpenClaw agents using a self-hosted SearXNG instance and the Scrapling anti-bot library. Runs multi-engine parallel searches across Bing, DuckDuckGo, Google, Startpage, and Qwant with no API keys required. Queries go to your local instance by default, with an automatic fallback to a public SearXNG endpoint when the local service is down.
Unlike hosted search APIs, it requires no keys or subscriptions and keeps all queries on your own machine by default.
When to use it
- Looking up breaking news without third-party query tracking
- Fetching full content from Cloudflare-protected or JavaScript-heavy pages
- Cross-validating facts across multiple search engines simultaneously
- Researching GitHub repos or technical docs with freshness filters
- Comparing products or frameworks using live, multi-engine results
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: What are the latest changes in the Python 3.13 release?
- 1Runs search_local_web.py with --intent news and --freshness month against local SearXNG
- 2Reviews results, prioritizing entries tagged [cross-validated] across multiple engines
- 3Selects the top-scoring URL and runs browse_page.py to fetch the full page
- 4If confidence is LOW, retries browse_page.py with --mode stealth to bypass anti-bot protection
- 5Returns a summary based only on HIGH-confidence page content
A factual summary of Python 3.13 changes sourced from a fully fetched, cross-validated page
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.