OpenClaw Skills List 2026

What Are OpenClaw Skills?

Skills are plug-in abilities that extend what your OpenClaw AI agent can do. Each skill adds a new capability — web search, image generation, code execution, calendar access, and hundreds more. Skills are built on the MCP (Model Context Protocol) standard, meaning they work reliably across any AI model your agent uses.

Without skills, your agent can only reason and respond. With skills, it can act: search the web, run code, send messages, generate images, manage files, and automate tasks. The OpenClaw skills list on ClawHub now exceeds 3,200 community-built entries — one of the largest AI agent plugin directories available.

Skills are installed with a single command:

openclaw skill install @username/skill-name

Once installed, enable the skill in your openclaw.json config or toggle it on from your OpenClaw Launch dashboard. Your agent gains the new capability immediately — no restart required for most skills.

Where to Find OpenClaw Skills

The primary source for the OpenClaw skills directory is ClawHub — the official skills marketplace at clawhub.io. ClawHub is maintained by the OpenClaw team and hosts skills published by both the core team and the broader community.

On ClawHub you can:

• Browse the full openclaw skills list organized by category
• Search by keyword to find skills matching your use case
• View install counts, ratings, and changelogs for each skill
• Read documentation and configuration examples before installing
• Publish your own skills to share with the community

Skills are also distributed via GitHub — many developers publish their skills as public repos and list them on ClawHub for discovery. You can install directly from a GitHub repo URL if a skill isn't listed on the hub yet.

For a curated view, browse the Skills page or visit the ClawHub overview to learn more about how the marketplace works.

Top Skills by Category

The OpenClaw skills hub organizes its 3,200+ entries into categories. Here are the most popular skills in each major area:

Web Search & Research

Give your agent real-time access to the internet. These are among the most-installed skills on the entire ClawHub directory.

• Brave Search — Privacy-focused web search with clean, summarized results. No tracking, no filter bubbles. One of the top skills in the openclaw skills store by install count.
• Tavily — AI-optimized search built specifically for agents. Returns structured, citation-ready results with high relevance. See the Tavily setup guide for configuration details.
• SearXNG — Self-hostable meta-search engine. Aggregates results from Google, Bing, DuckDuckGo, and dozens more. Great for privacy-conscious deployments.
• Firecrawl — Web scraping and crawling at scale. Extract structured data from any website, handle JavaScript-rendered pages, and crawl entire sites.
• url-reader — Fetch any URL and summarize its contents. Lightweight alternative to full browser automation for reading articles and documentation.

Coding & Development

Turn your agent into a hands-on development partner that writes, reviews, and runs code.

• code-interpreter — Execute Python, JavaScript, and other languages in a sandboxed environment. Your agent can write code and immediately run it to verify results.
• GitHub integration — Manage repositories, open issues, review pull requests, and push commits directly from chat. Covers the full GitHub API surface.
• code-review — Static analysis and best-practice suggestions for any code snippet or file. Catches bugs, security issues, and style violations.
• npm-search — Search the npm registry, compare packages, and view readme documentation without leaving your conversation.

Productivity & Automation

Automate your daily workflow — from inbox management to scheduled tasks and file operations.

• email — Read, draft, and send emails via Gmail or SMTP. Your agent can monitor your inbox and surface what matters.
• calendar — Check availability, create events, and manage meetings across Google Calendar and other providers.
• file-manager — Create, read, move, rename, and delete files. Manage your agent's workspace or connect to cloud storage.
• cron-scheduler — Set up recurring tasks that run automatically on a schedule. Daily reports, periodic checks, timed reminders.
• http-request — Make arbitrary HTTP requests to any REST API or webhook endpoint. The universal connector for external services.

Image & Media

Generate, edit, and analyze visual content directly through your agent.

• image-gen — Generate AI images from text descriptions using Stable Diffusion, DALL-E, or Flux. One of the most-searched skills on the openclaw skills hub.
• OCR — Extract text from images, scanned documents, and screenshots. Supports handwriting recognition and table extraction.
• screenshot — Capture screenshots of websites or local applications. Useful for visual testing, documentation, and monitoring.
• audio-transcribe — Convert voice messages and audio files to text. Supports 50+ languages with speaker diarization.

Communication

Connect your agent to the messaging platforms where you already work.

• Telegram connector — Send and receive Telegram messages through your agent. Native to OpenClaw's channel architecture — no separate bot token needed when using OpenClaw Launch.
• Discord connector — Bring your agent to Discord servers. Respond to mentions, monitor channels, and post automated updates.
• WhatsApp connector — Connect via WhatsApp Business API or QR-based pairing. Useful for customer-facing bots and personal productivity.
• slack-notify — Post messages to Slack channels and respond to Slack commands. Ideal for team bots and DevOps alerting.

Knowledge & Memory

Give your agent long-term memory and access to your personal knowledge base.

• long-term-memory — Persist facts, preferences, and context across sessions. Your agent remembers what you told it last week.
• document-qa — Upload PDFs, DOCX files, and text documents. Ask questions and get cited answers from your own knowledge base.
• vector-search — Semantic search over your stored documents and notes using embeddings. Finds relevant content even when exact keywords differ.
• note-taker — Capture structured notes during conversations and retrieve them by topic, date, or keyword later.

How to Install Skills

Installing a skill from the OpenClaw skills list takes one command:

openclaw skill install @author/skill-name

Replace @author/skill-name with the identifier shown on the ClawHub page for any skill. After installing, enable the skill in your openclaw.json configuration:

{
  "plugins": {
    "entries": {
      "skill-name": { "enabled": true }
    }
  }
}

Most skills activate immediately. Skills that add new channel connectors or modify gateway auth require a container restart.

OpenClaw Launch users: you don't need to touch config files. Browse skills from your dashboard, toggle them on, and they're live. OpenClaw Launch handles installation, configuration, and restarts automatically.

How to Find the Right Skill

With 3,200+ entries in the OpenClaw skills directory, the fastest way to find what you need is to use ClawHub's search and filtering tools:

• Search by keyword — Type what you want your agent to do (e.g. “send email”, “generate image”, “scrape web”) and ClawHub returns matching skills ranked by relevance and popularity.
• Browse categories — The openclaw skills hub sidebar lists every major category. Start broad, then drill down to find niche skills for specialized workflows.
• Check install counts and ratings — High install counts signal a stable, well-maintained skill. Read the rating comments for real-world usage feedback before committing.
• Read the documentation first — Each ClawHub listing includes a readme with configuration examples, required API keys, and known limitations. Five minutes of reading saves hours of debugging.
• Check the “awesome-openclaw-skills” list — The community maintains a curated awesome list on GitHub with hand-picked skills in every category. A great starting point for new users.

Building Custom Skills

If the OpenClaw skills list doesn't have exactly what you need, you can build your own. Skills are standard MCP servers — any language or runtime works as long as it implements the MCP protocol.

The typical workflow for creating a custom skill:

• Scaffold a new skill using the official OpenClaw skill template on GitHub
• Implement your tool functions using the MCP SDK
• Test locally by pointing your openclaw.json at the local skill path
• Publish to npm under an @username/skill-name package name
• Submit to ClawHub to appear in the public openclaw skills store directory

Custom skills can wrap any API, script, or service. Teams commonly build internal skills for proprietary databases, company-specific automation, and private integrations that they don't publish publicly.

Pre-installed Skills on OpenClaw Launch

OpenClaw Launch comes with the most popular skills from the OpenClaw skills list pre-installed and ready to enable — no command line required. When you deploy a new instance, popular skills like web search, code interpreter, and file manager are already available in your dashboard.

For any other skill in the ClawHub directory, one-click install from the UI adds it to your instance automatically. OpenClaw Launch handles:

• Downloading and installing the skill package
• Injecting the correct configuration into your openclaw.json
• Restarting the container if required
• Surfacing the skill's settings in the visual config editor

No SSH, no config files, no restarts to manage yourself. Deploy your agent with the full OpenClaw skills hub available at your fingertips.

Related Guides

• Best OpenClaw Skills in 2026 — Curated picks across every category
• OpenClaw ClawHub Guide — How to browse, install, and manage skills from the marketplace
• OpenClaw Tavily Search Setup — Detailed setup guide for the Tavily web search skill
• OpenClaw Browser Automation — Browser relay and web automation skills
• OpenClaw Memory Guide — How agent memory and context persistence works