numman-ali/openskills
Overview
This skill is an introductory example that demonstrates how to structure a procedural guidance package for AI coding agents. It serves as a minimal, practical template to learn skill authoring and to validate a universal skills loader. The focus is on progressive disclosure and clear resource resolution.
How this skill works
The skill provides a base directory layout and shows how relative paths resolve to bundled resources. It includes concise procedural instructions for creating a new skill directory, adding metadata, and organizing supporting files. Load the skill to inspect examples of instructions, references, scripts, and assets.
When to use it
- Learning how to author a skill for AI coding agents
- Testing or validating a universal skills loader implementation
- Understanding the expected skill file format and structure
- Teaching progressive disclosure and concise procedural writing
Best practices
- Write instructions in imperative/infinitive form: 'To do X, execute Y'
- Avoid second-person phrasing such as 'you should...'
- Keep primary guidance concise (recommendation: under 5,000 words)
- Move detailed or large content into a references directory
- Place executable helpers in a scripts directory and templates in an assets directory
Example use cases
- Onboarding new contributors to author consistent skills
- Verifying that a loader correctly resolves relative paths to bundled resources
- Using as a checklist for structuring metadata and support files
- Demonstrating how to separate concise guidance from in-depth references
FAQ
The loader provides a base directory for the skill; relative paths in the skill resolve from that base to the referenced files.
Where should long-form or detailed content live?
Place detailed documentation in a references directory and keep the main guidance concise to improve readability and loading performance.
