An autonomous web learning skill that empowers AI agents to search, browse, and extract real-time knowledge from the internet.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install web-learner-1-0-0
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 web-learner-1-0-0 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 Learner?
The Web Learner skill is a powerful extension for AI agents within the Openclaw Skills ecosystem, enabling them to break free from static training data limitations. It provides the logic and tools necessary for an agent to autonomously navigate the web, perform precision keyword searches, and fetch detailed content from specific URLs. By integrating this skill, developers can create agents capable of staying updated with current events, learning complex new subjects on the fly, and providing data-backed responses derived from the latest online sources.
This skill focuses on high-quality information extraction, prioritizing structured Markdown output for better AI comprehension. It bridges the gap between a language model's internal knowledge and the ever-changing landscape of the live web, making it an essential component for any Openclaw Skills implementation requiring research or real-time data.
Web Learner Use Cases
- Retrieving real-time information such as news, weather updates, and stock market trends.
- Learning new technical topics or researching detailed documentation for specific projects.
- Fetching and summarizing content from provided URLs or online documents.
- Verifying facts or supplementing existing training data with up-to-date internet resources.
- Generating comprehensive reports based on multi-source web research and data gathering.
How Web Learner Works
- Requirement Analysis: The AI agent identifies the need for external information based on user intent and identifies the specific knowledge gap.
- Tool Selection: The skill dynamically chooses between web_search (for discovery), web_fetch (for direct content extraction), or browser (for complex JavaScript-heavy sites).
- Information Retrieval: The skill executes search queries or direct page fetches using optimized parameters like region filters or time constraints.
- Processing & Synthesis: The agent extracts key information from the raw Markdown or text outputs, filtering out navigation menus and irrelevant noise.
- User Feedback: Finally, the skill presents a structured, concise summary of the findings in the requested language with clear source attribution.
Web Learner Setup
To get started with this skill in your Openclaw Skills environment, follow these steps:
- Ensure your core environment is configured and updated.
- Obtain a Brave Search API Key and configure it as an environment variable to enable the
web_searchfunctionality. - For the
browsertool, ensure the browser service (such as Playwright) is active and reachable. - Install the skill via your terminal:
openclaw install web-learner
Web Learner Data Schema & Taxonomy
The Web Learner skill organizes retrieved data into several formats to ensure high utility for Openclaw Skills agents:
| Data Type | Description | Primary Format |
|---|---|---|
| Search Results | Metadata including titles, snippets, and source URLs. | JSON |
| Page Content | Main text body or documentation extracted from HTML. | Markdown/Text |
| Visual Data | Screenshots for UI interaction or layout verification. | PNG/JPG |
| Search Metadata | Constraints like country (e.g., 'CN') and freshness (e.g., 'pd'). | Key-Value Pairs |
Web Learner Advanced Features
- Multi-tool fallback logic that automatically switches to the browser tool if standard fetching fails due to JavaScript requirements.
- Localized search filtering using specific country and language parameters for high-relevance regional results.
- Temporal filtering with 'freshness' parameters to limit searches to the last day, week, or month.
- Integrated anti-crawling and login-restriction handling for more robust data retrieval from protected sources.
- Support for UI interaction and screenshot capture to help agents understand complex web layouts.
SKILL.md
Loading
Related Openclaw Skills
Fennec SEO Audit
A specialized skill for conducting quick on-page SEO audits and Generative Engine Optimization assessments using the Fennec SEO Auditor extension.
Gmail Recent Digest
An automated utility for Openclaw Skills that retrieves recent unread Gmail messages and generates a concise, summarized digest.
Todoist CLI
A powerful command-line interface for managing Todoist tasks, allowing for rapid task creation, filtering, and organization from the terminal.
Spotify Control for macOS
A macOS-native skill for controlling Spotify playback, volume, and media selection via CLI and AppleScript.
Crypto Levels Analyzer
A sophisticated technical analysis tool that identifies key support and resistance levels for major cryptocurrency pairs.
Telegram Bot Manager
A specialized tool for registering, configuring, and maintaining Telegram bot integrations within the OpenClaw ecosystem.
Featured*
