Stage 3: Skills & Knowledge
Duration: ~5 minutes
What happens: AgentSherpa loads its domain skills and explains each one so you know what expertise it brings.
What are agent skills?
Skills are structured knowledge files that give the AI agent domain expertise. Without skills, the agent only has general knowledge. With skills, it knows:
-
How to build MCP servers with the Python SDK v2
-
How to deploy using MCPServer CRs on OpenShift
-
How to register with the MCP Gateway
-
How to troubleshoot common issues
-
How to score your progress and generate certificates
The 17 skills
AgentSherpa has 17 pre-loaded skills organized by domain:
| Category | Skills |
|---|---|
Workshop flow |
workshop-guide, voice-and-pacing, presentation-mode |
Build |
build-mcp-server, mcp-concepts |
Deploy & operate |
build-deploy-openshift, use-mcp-tools, troubleshooting-coach, workshop-cleanup |
Assessment |
knowledge-check, workshop-scoring, interactive-challenge, achievement-system, workshop-certificate |
Reference |
redhat-mcp-resources |
What AgentSherpa does
-
Lists all loaded skills with one-line summaries
-
Explains key skills in more detail based on your experience level
-
Asks if you have questions before moving on
Why skills matter
Skills are the difference between a generic chatbot and a domain expert. They encode:
-
Correct patterns — e.g., MCPServerRegistration uses
prefix, nottoolPrefix -
Best practices — e.g., always wait for MCPServer Ready=True before registering
-
Error recovery — e.g., if the build fails, check
oc logs bc/<name>