Why should you care?

SK unifies OpenAI and Azure OpenAI behind Kernel and KernelPlugin collections so teams stop rewriting orchestration for each feature.

See it live — copy this example

Paste into an ASP.NET Core 8+ / AgentNest project, then run with dotnet run (set your API keys in user-secrets).

// AgentNest.Agents/KernelFactory.cs
public static Kernel Create(IChatClient chat, CrmPlugin crm)
{
    var builder = Kernel.CreateBuilder();
    builder.Services.AddSingleton(chat);
    builder.Plugins.AddFromObject(crm, pluginName: "crm");
    return builder.Build();
}

What happened?

• KernelFactory registers IChatClient and adds CrmPlugin as a named plugin.
• Agents invoke kernel.InvokeAsync with function names like crm.GetAccount.

Practice next

1. dotnet add package Microsoft.SemanticKernel.
2. build CrmPlugin with [KernelFunction] methods.
3. Build Kernel in DI via KernelFactory.Create.
4. Add HospitalPlugin alongside CrmPlugin and test selective registration.
5. Log plugin name on each InvokeAsync for audit.

Remember

Semantic Kernel wraps chat clients and plugins in one Kernel. AgentNest centralizes Kernel creation in a factory. Plugins expose tools as attributed C# methods.

Unified agent host

Three squads each wrote custom OpenAI wrappers.

Outcome: KernelFactory standardizes plugin hosting and cuts duplicate HTTP code.