A comprehensive framework and toolkit for developing specialized AI agent skills that extend agent capabilities through structured workflows and domain-specific knowledge.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install infiniteresearchloop
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 infiniteresearchloop 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 Skill Creator?
Skill Creator is the definitive guide and utility suite for developers looking to extend the capabilities of AI agents. By utilizing Openclaw Skills, developers can transform a general-purpose model into a specialized expert equipped with procedural knowledge, custom scripts, and domain-specific schemas. The skill emphasizes the principle of progressive disclosure, ensuring that agent context windows remain efficient while providing deep access to complex tools when necessary.
This system allows for the creation of modular, self-contained packages that include YAML-based metadata, specialized Markdown instructions, and bundled resources. It bridges the gap between raw LLM intelligence and brittle, hard-coded automation by providing a framework that supports varying degrees of freedom for the AI agent, from strict scripted sequences to flexible, heuristic-based tasks.
Skill Creator Use Cases
- Creating specialized onboarding guides for AI agents to handle company-specific business logic.
- Developing modular tool integrations for proprietary APIs or unique file formats.
- Packaging repetitive DevOps or data processing tasks into reusable scripts for AI execution.
- Scaling AI agent performance by offloading complex documentation into a structured reference system.
How Skill Creator Works
- Discovery: Gather concrete usage examples to define the necessary functionality and specialized knowledge the agent requires.
- Resource Planning: Identify reusable assets such as Python/Bash scripts, documentation references, and output templates.
- Initialization: Use the provided CLI tools to generate a standardized skill directory structure including the required SKILL.md and frontmatter.
- Development: Implement imperative instructions in the SKILL.md and populate the resource directories with functional code and data.
- Validation: Run automated scripts to ensure the skill meets technical requirements and metadata standards.
- Deployment: Package the skill directory into a distributable format for integration into the agent environment.
Skill Creator Setup
To begin creating new Openclaw Skills, use the initialization script to scaffold your project directory:
# Initialize a new skill directory
python /home/ubuntu/skills/skill-creator/scripts/init_skill.py <skill-name>
# Validate your skill after development
python /home/ubuntu/skills/skill-creator/scripts/quick_validate.py <skill-name>
Skill Creator Data Schema & Taxonomy
Each skill created with the Skill Creator follows a strict organizational taxonomy to optimize for agent context loading:
| Component | Purpose | Context Impact |
|---|---|---|
| SKILL.md | Core instructions and metadata | Medium (Loaded on trigger) |
| Frontmatter | Trigger definitions and name | Low (Always in context) |
| scripts/ | Executable Python/Bash code | None (Runs via CLI) |
| references/ | Heavy documentation and schemas | Variable (Loaded as needed) |
| templates/ | Boilerplate code and assets | None (Output only) |
Skill Creator Advanced Features
- Progressive Disclosure: A three-level loading system that protects the context window by only loading deep documentation when explicitly triggered.
- Variable Freedom Constraints: Ability to define high, medium, or low degrees of freedom for agent behavior based on task fragility.
- Automated Validation: Built-in linting and validation scripts to ensure skill integrity before deployment.
- Modular Reference Patterns: Support for multi-domain skills where specific reference files (e.g., finance.md vs. sales.md) are loaded dynamically.
SKILL.md
Loading
Related Openclaw Skills
K241 Word Upload
An automated vocabulary uploader for the K241 platform featuring integrated translation and pinyin fetching.
WeChat Publisher
A professional automation tool that converts Markdown articles into WeChat Official Account drafts with integrated image management.
Node.js Best Practices & Code Auditor
A comprehensive quality enforcement skill that audits Node.js applications for security, performance, and architectural integrity.
bee-products-read
A specialized skill for reading and extracting comprehensive product lists and metadata from Bee-powered websites via open API access.
WeChat Official Account Operations
An all-in-one AI agent skill for managing the complete lifecycle of WeChat Official Account operations, from trend research to automated publishing.
Travel Grid Generator
Travel Grid Generator creates high-quality, character-consistent 3x3 social media photo collages in a realistic travel blogger style.
Featured*
