🔎 Find Skills

Ökosystem-Erkundung · Auto-Routing · Paketmanager

Search the global skill registry to help users discover and install ClawHub skills. Using natural language understanding to find matching results with one-click install commands, giving your AI the ability to "find its own tools" and evolve.

OpenClaw Team

🚀 Schnellinstallation

Führen Sie den folgenden Befehl in Ihrem Terminal aus, um zu installieren:

npx clawhub install find-skills

📊 Statistikübersicht

⭐ Sterne	☁️ Gesamtdownloads	👥 Aktive Nutzer	🎯 Stabile Version
54	12k	456	v1.0.2

🎛️ So funktioniert es

This skill serves as the OpenClaw ecosystem's "package manager LLM assistant" — not just a static query interface, but a discovery engine with semantic-level association capabilities. It grants agents the right to automatically search the central registry for technical solutions when facing unknown tasks:

🧠 Heuristic Skill Discovery: Deeply understands user intent scenarios. When AI determines that currently available skills can't complete the goal, it automatically searches the entire ClawHub skill library to unearth highly relevant installable components.
🎯 Multi-dimensional Semantic Matching: Breaks through traditional keyword matching limitations. Supports direct parsing of extremely colloquial natural language queries like "How do I automatically scrape data from Twitter on a schedule?" or "Is there a tool skill that can help me clean up my disk?", returning optimal results instantly.
📦 Guided Installation: Intuitively displays matched skill listings (with detailed descriptions and global installation popularity indices), and generates ready-to-use installation CLI commands at the end — truly "copy and run."

🧭 Typische Anwendungsfälle

🗣️ Szenario 1: IDE-immersive Intent Conversion

As a developer, you don't need to constantly switch to web browsers to search. While debugging workflows, simply tell the underlying agent: "I need a headless web automation executor for testing," and it will instantly connect to the cloud hub, precisely discovering agent-browser or playwright components and making strong recommendations.

🔄 Szenario 2: Dynamic Multi-Agent Capability Injection

In large multi-agent collaboration networks, if one agent node encounters an unforeseen task mid-execution (e.g., after data analysis, suddenly needing to send an email report to management), the agent can use this pathway to automatically search for gmail or smtp related skills, and with your permission, achieve real-time pipeline completion.

💻 Befehlsreferenz

After installation, you can let AI call these autonomously via conversation, or manually trigger queries from the CLI:

Execute a natural language fuzzy search:

clawhub search "web scraping tool with multi-modal support"

Search by category tags with popularity sorting:

clawhub search --category automation
clawhub list --sort popular

Once you've identified the desired plugin, execute a one-click install:

clawhub install <skill-slug>

🛡️ Voraussetzungen & Sicherheit

🌐 Network Access: This skill requires internet connection to query clawhub.ai's cloud registry index. Ensure your environment doesn't block HTTPS traffic.
🛡️ Ecosystem Security: All recommended skill listings have been verified through OpenClaw's official virus database and trojan scanning to isolate potentially malicious nodes.

🔗 Quellcode auf GitHub ansehen