skill-creator
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.
Why use this skill?
Learn how to build, package, and deploy custom OpenClaw skills. Master the art of creating modular, efficient agent extensions with this comprehensive guide.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/yixinli867/skill-creator-2What This Skill Does
The skill-creator skill serves as a structural framework and pedagogical guide for developers and power users looking to extend the OpenClaw AI agent. It acts as the definitive playbook for packaging domain-specific expertise, specialized workflows, and custom tool integrations into modular, portable units that Claude can execute. By providing a standardized anatomy for skill development, it ensures that your agent extensions remain efficient, readable, and highly effective within the constraints of the AI's context window.
Installation
To install this skill, run the following command in your terminal:
clawhub install openclaw/skills/skills/yixinli867/skill-creator-2
Ensure you have the OpenClaw CLI initialized in your working directory before running the installation command.
Use Cases
- Onboarding for domain-specific tasks: Create an expert guide for proprietary business logic, coding standards, or internal API documentation.
- Automation workflows: Bundle multi-step scripts for data processing or file manipulation that require consistent execution.
- Knowledge retrieval: Package complex reference documentation that Claude can pull into its context window only when specifically requested.
- Interactive tutorials: Design guided experiences where the AI walks a user through a process step-by-step using pre-defined templates.
Example Prompts
- "I want to build a new skill for summarizing legal contracts; how should I structure the SKILL.md file to ensure Claude prioritizes the relevant clauses?"
- "Help me convert this set of internal bash scripts into an OpenClaw skill that allows for modular reuse across our engineering team."
- "Evaluate my current skill definition and suggest how to adjust the 'freedom' level of the instructions to make the output more consistent and less prone to hallucination."
Tips & Limitations
- Conciseness is your highest priority. Every token used for instructions is a token taken away from the agent's reasoning capacity. Avoid fluff; prioritize clear, actionable bullet points.
- Assess the 'Degrees of Freedom' carefully. If your task is highly error-prone, force the agent into a narrow path using strict pseudocode or low-freedom instructions. If the task is creative or highly contextual, allow for high-freedom natural language prompts.
- Keep your bundled resources minimal. Large files should be indexed in a metadata layer rather than loaded into the context window permanently. Test your skill frequently in different context sizes to ensure the performance remains stable.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-yixinli867-skill-creator-2": {
"enabled": true,
"auto_update": true
}
}
}