Stage 1: Welcome & Concepts

Duration: ~5 minutes
What happens: AgentSherpa introduces itself, learns your name and experience level, and teaches the fundamentals of MCP.

How it starts

When you type your first message, AgentSherpa:

  1. Introduces itself with your name and the workshop overview

  2. Asks about your experience level (beginner / intermediate / advanced) using clickable options

  3. Adapts all future explanations to your level

What AgentSherpa teaches

MCP fundamentals

AgentSherpa explains MCP using a concept card:

  • MCP is an open protocol for connecting AI agents to tools — like a USB standard for AI

  • Three roles: Host (the AI app), Client (built into the host), Server (your code)

  • Tool definitions include names, descriptions, and typed parameters so the LLM can decide when to use them

  • The protocol uses JSON-RPC 2.0 over HTTP with SSE (Streamable HTTP transport)

The three lifecycle layers

If you are at beginner or intermediate level, the agent explains the OpenShift MCP lifecycle:

Layer Analogy

Deploy (MCPServer CR)

Like a "shipping manifest" — you declare what you want, the operator builds and runs it

Discover (AI Hub)

Like an "app store" — deployed servers appear for browsing

Route & Govern (MCP Gateway)

Like a "switchboard" — one entry point, many servers, with auth and policies

Points

Action Points

Complete welcome conversation

5

What to expect

  • AgentSherpa will explain before executing any command — you always know why before it does something

  • You will be asked to confirm before moving to the next stage

  • If anything is unclear, just ask — the agent adapts