Build AI agents with agent skills

In this free, on-demand technical lab, learn how to build a declarative AI agent with an agent skill that deploys applications to Kubernetes.

• Build a skill: Use a predefined Kubernetes deployment skill that knows how to generate manifests and deploy apps.
• Package the skill: Build a Docker image bundling the skill and push it to the local registry.
• Create a Kubernetes agent: Define a declarative agent, reference the skill, and deploy it to your local cluster.
• Interact with the agent: Start a chat session in the kagent Dashboard and instruct the agent to deploy a sample application, such as httpbin.
• Analyze skill execution: Observe how the agent reads the skill instructions, invokes the script, and applies the manifest using its MCP tools.
• Verify deployment: Use a curl pod to make requests against the deployed app and confirm it is running correctly.

Gain hands-on experience extending AI agents with reusable skills, automating Kubernetes operations, and combining declarative agent behavior with MCP-powered tools.

‍