Getting Started with Semantic Kernel .NET

1. Using InvokePrompt

• This step demonstrates how to define a prompt and get the result from an AI Model using the InvokePrompt method in Semantic Kernel.

1.2 Using InvokePrompt with arguments

• Introduces the concept of templating in prompts, allowing dynamic arguments to be passed at the invocation time using KernelArguments dictionary.

1.3 Using InvokePromptStreaming with arguments

• Explains how to use streaming to update the UI in real time as information arrives from the Model, reducing the perception of delay between the request and the result.

1.4 Using Kernel Arguments

• Demonstrates how to specify execution settings while invoking the AI Model by creating a KernelArguments class with specific configuration settings.

1.5 Getting JSON with Custom Settings

• Shows how to specify custom settings, such as ResponseFormat, to instruct the model to return results in JSON format.

2. Using Plugins with Kernel

2.1 C# Method Functions

• Details how C# functions can be imported as plugins into the Kernel by marking them with KernelFunction attribute.

2.2 Using Template to Invoke Plugin Functions

• Illustrates the usage of templates to append the result of a template function to the prompt before sending it to the AI Model.

2.3 Auto Function Invoking

• Explains how to configure the settings to allow the AI Model to call your plugin functions automatically based on the ToolCallingBehavior specified.

2.4 Auto Function Invoking with Complex Object Plugin Functions

• Highlights how the AI Model understands the structure of complex object parameters in plugin functions, providing them following the description and deserialization representation.

By following these steps, users can get started with Semantic Kernel .NET and leverage its powerful features for interacting with AI models efficiently.