Agent Skills 101
Agent Skills are the open standard for extending AI coding agents with specialized capabilities. Adopted by Claude Code, Cursor, Codex, GitHub Copilot, and 50+ more agents, Skills package domain knowledge, multi-step workflows, and best practices into portable folders any agent can follow. This course takes you from understanding what Skills are and how they differ from MCP and Rules, through installing and discovering Skills across the ecosystem, to building your own from scratch. You'll learn the SKILL.md spec, progressive loading architecture, cross-agent compatibility, security best practices, and how to iterate Skills into reliable playbooks. Whether you're using Claude Code, Cursor, or any compatible agent — this course gives you everything you need to 10x your agent workflows.
Course Outline
Theory: What Are Agent Skills
Understand the Agent Skills open standard and how progressive loading works.
Theory: Skills VS MCP
The critical distinction between Skills, MCP, and Rules - and when to use what.
Practice: Finding & Using Skills
Install and discover Skills across the ecosystem, set up Skills in Claude Code and Cursor.
Hands-On: Knowledge Retrieval Skill
Build a knowledge retrieval skill from scratch with progressive loading and cross-agent compatibility.
Hands-On: Deconstructing Claude Design
Deconstruct a real-world skill used in production to understand architecture patterns.
Best Practices from Anthropic
Official best practices and recommendations from Anthropic for building robust agent skills.
Building Your Own Skill
Master the SKILL.md spec, YAML frontmatter, testing, debugging, iteration, and security.
Previous Edition (Deprecated)
Earlier version of the course content, kept for reference.