Learn MCP and Skills from Zero
A structured learning path for understanding how agents connect to tools, reuse expertise, and combine MCP with Skills in real systems.
Recommended learning path
If you are new to both concepts, do not start from implementation details. Start with the mental model, then follow one practical example, then compare when to use MCP, Skills, or both.
Start with simple explanations of MCP, Agent Skills, and how they work together before touching code.
Read MCP explained simplyLearn why agents need a standard way to discover and call tools, data, and external services.
Start with MCPLearn how repeatable task knowledge is packaged into SKILL.md, references, scripts, and templates.
MCP is the capability access layer. Skills are the reusable task procedure layer. They solve different problems.
Read MCP vs SkillExpose one capability through MCP, then write one Skill that describes how an agent should use it.
Follow MCP QuickstartUse one page to review the full learning sequence, glossary, and capstone checklist without hunting around.
Open the complete courseWhat each section teaches
The site is organized like a course, not a random glossary. Each track has concepts, practical pages, examples, and FAQ pages.
Definition, architecture, client, server, quickstart, best practices, FAQ, and an example server page.
Definition, structure, creation, quickstart, best practices, evaluation, FAQ, and reusable template.
When to use MCP, when to use Skills, and how they work together inside agent workflows.
A guided route, glossary, and troubleshooting guide that help a beginner finish the whole site in one sitting.
The mental model
An AI agent needs two things to do real work reliably: access to capabilities and knowledge about how to use those capabilities. MCP and Skills address those two layers.
Start now
Pick one path based on what you want to learn first.