A professional framework for building modular, self-contained skills that extend the capabilities of AI coding agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install codex-skill-authoring
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 codex-skill-authoring 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 Authoring for AI Agents?
Skill Authoring is a systematic approach to extending AI agent functionality through modular development. By creating Openclaw Skills, developers can provide agents with the specific scripts, reference documents, and assets needed to perform complex, deterministic tasks that fall outside of general model knowledge.
This skill provides a standardized 6-step workflow—Understand, Plan, Init, Edit, Package, and Iterate—ensuring that every new capability is robust, validated, and ready for deployment. It emphasizes lean documentation and progressive disclosure to keep agent context windows efficient while maintaining high performance.
Skill Authoring for AI Agents Use Cases
- Creating custom automation workflows for specific development environments.
- Packaging reusable Python or Bash scripts into portable Openclaw Skills.
- Defining specific trigger phrases and instructions for localized or bilingual AI agent tasks.
- Standardizing organizational coding standards by bundling reference documentation for AI agents.
- Building complex skillsets that require automated validation before packaging.
How Skill Authoring for AI Agents Works
- Understand the functionality and specific trigger phrases required for the new skill.
- Plan the directory structure, identifying which components belong in scripts, references, or assets.
- Initialize the skill using the standardized CLI tool to ensure correct metadata and folder hierarchy.
- Author the SKILL.md file with precise instructions and imperative logic for the AI model.
- Execute the packaging script which automatically validates the skill structure and creates a compressed .skill file.
- Deploy and iterate based on real-world task performance and agent feedback.
Skill Authoring for AI Agents Setup
To begin creating Openclaw Skills, ensure you have the core toolkit installed. Use the following commands to initialize and package your work:
# Initialize a new skill directory
python3 scripts/init_skill.py <name> --path <output-dir> --resources scripts,references,assets
# Package a completed skill for distribution
python3 scripts/package_skill.py <path/to/skill-folder>
Skill Authoring for AI Agents Data Schema & Taxonomy
Every project within the Openclaw Skills ecosystem follows a strict directory and metadata taxonomy:
| File/Folder | Requirement | Purpose |
|---|---|---|
SKILL.md |
Required | Contains name, version, description, and core instructions. |
scripts/ |
Optional | Executable Python or Bash code for deterministic tasks. |
references/ |
Optional | Heavy documentation loaded by the agent on-demand. |
assets/ |
Optional | Templates, icons, and boilerplate files for output generation. |
Naming Convention: All skill names must be lowercase, use hyphens only, and stay under 64 characters.
Skill Authoring for AI Agents Advanced Features
- Automatic validation runs during the packaging phase to catch structural errors before deployment.
- Support for bilingual triggers (English and Chinese) to enable global AI agent interaction.
- Progressive disclosure architecture that keeps the primary SKILL.md lean by offloading details to the references directory.
- Multi-resource bundling allowing scripts, assets, and documentation to function as a single atomic unit.
SKILL.md
Loading
Related Openclaw Skills
Xiaohongshu Anti-Detection Post-Processor
A professional post-processing tool designed to reduce AI detection signatures on images by applying metadata obfuscation and subtle pixel modifications.
Summer Countdown & Beijing Solar Dashboard
A real-time seasonal dashboard providing live countdowns to the summer solstice and precise solar data for Beijing, China.
rohoon-6sigma
A professional-grade Six Sigma support toolkit providing advanced statistical process control, measurement analysis, and experiment design based on AIAG-VDA 2026 standards.
Retention Phase Intervention Selector
A diagnostic framework that identifies which retention phase is failing and prescribes specific growth interventions based on product-type benchmarks.
Product Design 0→1
A professional-grade automated pipeline for designing internal process management systems from initial concept to high-fidelity interactive prototypes.
Simp Dog Skill Creator
A sophisticated persona distillation tool that converts social media and chat exports into persistent, subservient AI agents with evolving memories.
Featured*
