Skills for AI
Skills.MD
AI Agent Skills Index
This page serves as a central registry for some of the skills I use (either created by me or that I have found very useful) and other excellent third-party tools and specifications I use in my own Agentic Workflow tech stack.
What is an AI Skill?
An AI Skill is an isolated, portable module that gives an LLM agent the precise system context, operational rules, and execution scripts it needs to perform a distinct role. Instead of relying on sprawling, fragile prompts, structuring capabilities as skills allows me to delegate specific tasks to an agent with repeatable, reliable results.
Why I Use This Approach
When tinkering with local frameworks or complex orchestration pipelines, giant monolithic codebases break down quickly. By building modular skills, I can seamlessly hot-swap tools into local environments—running open-source models via Ollama or LM Studio—without rebuilding my core applications from scratch.
Where These Skills Live
Modular skill standards are transforming how modern autonomous engines and enterprise frameworks execute long-running, multi-step tasks. These packaged capabilities are explicitly built to integrate with leading agentic platforms, terminal runners, and desktop environments, including:
My Directory Layout Standard
To keep my agentic workflows consistent, portable, and easy to deploy across various open-source orchestration runners, I structure every skill folder using this exact format:
├── assets/
├── scripts/
├── skill.md
└── readme.md
# Python, Node.js, or utility scripts triggered by the skill
# Strict machine-facing system instructions and YAML metadata
# Plain-English documentation, setup steps, and human-facing examples
*Note: Separating the LLM instructions (skill.md) from the human guide (readme.md) saves critical tokens during runtime while ensuring I don’t forget how my own code works six months down the line.