claude-skills-reference/docs/guides/openclaw-skills-guide.md

OpenClaw Skills Guide — Install & Use Agent Skills with OpenClaw

Last updated: March 2026 · Skills count: 192+ · Compatibility: OpenClaw v2024.12+

What Are OpenClaw Skills?

OpenClaw skills are modular instruction packages that extend your OpenClaw agent with domain expertise — from engineering and DevOps to marketing, compliance, and C-level advisory. Each skill drops into your OpenClaw workspace and works immediately with zero configuration.

Unlike generic prompts, OpenClaw skills include structured workflows, decision frameworks, Python tools, and reference materials that your agent follows autonomously.

Why Use Skills with OpenClaw?

Without Skills	With Skills
Generic responses	Domain-expert-level outputs
Manual prompt engineering	Pre-built workflows with slash commands
No tooling	Python scripts for analysis, validation, formatting
Starts from scratch	References, templates, best practices included

OpenClaw's skill system is the most natural fit in the ecosystem — skills live in your workspace directory and are automatically loaded based on task context.

Installation

Quick Install (Recommended)

bash <(curl -s https://raw.githubusercontent.com/alirezarezvani/claude-skills/main/scripts/openclaw-install.sh)

This installs all 192+ skills into your OpenClaw workspace with the correct directory structure.

Manual Install

git clone https://github.com/alirezarezvani/claude-skills.git
cd claude-skills
./scripts/install.sh --tool openclaw

Install Specific Skill Packs

# Engineering (49 skills)
./scripts/install.sh --tool openclaw --pack engineering

# Marketing (43 skills)
./scripts/install.sh --tool openclaw --pack marketing

# Product (12 skills)
./scripts/install.sh --tool openclaw --pack product

# C-Level Advisory (28 skills)
./scripts/install.sh --tool openclaw --pack c-level

# Regulatory & Quality (12 skills)
./scripts/install.sh --tool openclaw --pack regulatory

ClawHub Install

If you have the ClawHub CLI:

clawhub install alirezarezvani/claude-skills

How Skills Work in OpenClaw

OpenClaw has native skill support — it scans <available_skills> in your workspace and auto-selects the right skill for each task.

Automatic selection: When you ask your OpenClaw agent to "optimize this Dockerfile," it reads the docker-development skill's SKILL.md and follows its workflow. No manual activation needed.

Slash commands: Each skill defines slash commands (e.g., /docker:optimize, /research:summarize) that trigger specific workflows.

Python tools: Skills include executable scripts in scripts/ that your agent can run for analysis, validation, and generation tasks.

Top OpenClaw Skills by Category

Engineering

Skill	What It Does
docker-development	Dockerfile optimization, multi-stage builds, security hardening
terraform-patterns	Infrastructure-as-code patterns and module design
github	PR workflows, CI/CD, code review automation
frontend-design	Production-grade UI components with high design quality
mcp-builder	Build MCP servers for external API integrations

Marketing & Content

Skill	What It Does
content-creator	SEO-optimized blog posts, social media, brand voice
copywriting	Landing pages, headlines, CTAs, product copy
email-sequence	Drip campaigns, onboarding flows, lifecycle emails
launch-strategy	Product launches, Product Hunt, feature announcements
competitor-alternatives	Comparison pages, vs pages, alternative pages

Product & Research

Skill	What It Does
research-summarizer	Academic papers, articles, structured briefs with citations
agile-product-owner	User stories, sprint planning, backlog management
ab-test-setup	Experiment design, hypothesis testing, variant analysis

C-Level Advisory

Skill	What It Does
ceo-advisor	Strategy, board prep, investor relations
cto-advisor	Tech debt, team scaling, architecture decisions
cfo-advisor	Financial modeling, fundraising, burn rate analysis

OpenClaw vs Other Platforms

Feature	OpenClaw	Claude Code	Cursor	Codex
Native skill loading	✅ Automatic	✅ Manual	⚠️ Rules only	⚠️ Instructions
Slash commands	✅	✅	❌	❌
Python tool execution	✅	✅	❌	✅
Multi-agent delegation	✅ Built-in	❌	❌	❌
Persistent memory	✅	⚠️ Session	❌	❌
Cron/scheduled tasks	✅	❌	❌	❌

OpenClaw's architecture — persistent agents, memory, cron jobs, and multi-channel messaging — makes it the most capable platform for running agent skills autonomously.

Skill Anatomy

Every skill in the repository follows the same structure:

skill-name/
├── SKILL.md              # Instructions, workflows, slash commands
├── .claude-plugin/
│   └── plugin.json       # Metadata for plugin registries
├── scripts/
│   ├── tool_one.py       # Executable Python tools
│   └── tool_two.py
└── references/
    ├── patterns.md       # Domain knowledge, templates
    └── best-practices.md

Creating Custom OpenClaw Skills

You can create your own skills following the same format:

1. Create a directory under your workspace skills folder
2. Write a SKILL.md with description, slash commands, and workflows
3. Add Python scripts in scripts/ for any automation
4. Add reference materials in references/
5. OpenClaw will auto-discover and use your skill

Use the skill-creator meta-skill for guided skill creation:

/skill:create my-custom-skill

Resources

• GitHub: alirezarezvani/claude-skills
• ClawHub: clawhub.com
• OpenClaw Docs: docs.openclaw.ai
• Community: Discord

---

Part of the Claude Code Skills & Agent Plugins repository — 192+ production-ready skills for 11 AI coding tools.