semantic-kernel

Overview

Semantic Kernel is Microsoft's SDK for building AI agents and orchestrating AI plugins. It integrates LLMs with native code, supports multiple languages (C#, Python, Java), and provides planners for automatic function chaining.

Capabilities

• Create plugins with semantic (prompt) and native (code) functions
• Use planners to auto-select and chain functions
• Manage conversation memory and context
• Integrate with Azure OpenAI, OpenAI, and local models
• Build agents with tool use and multi-step reasoning
• Support RAG with vector stores and embeddings

When to Use

Trigger phrases:

• "semantic kernel"

• "Microsoft Semantic Kernel — AI orchestration, plugins, planners, memory, prompt "

• Building AI agents in .NET/C#/Python enterprise environments

• Needing structured plugin architecture for AI capabilities

• Wanting planners to dynamically compose function chains

• Integrating with Microsoft/Azure ecosystem

• Building RAG applications with enterprise data

When NOT to Use

• Task is outside your authorization scope
• You need to implement controls (use implementing-* skills)
• Task is about analysis, not action (use analyzing-* skills)
• You don't have access to target systems
• Task requires compliance expertise (consult professionals)
• Task is about defense, not offense (use defensive skills)

Pseudo Code

# Example workflow for this skill
def execute(input_data):
    # Step 1: Validate input
    if not input_data:
        raise ValueError("Input data is required")

    # Step 2: Process core logic
    result = process(input_data)

    # Step 3: Validate output
    validate_output(result)

    return result

Kernel Setup

import semantic_kernel as sk
from semantic_kernel.connectors.ai.open_ai import OpenAIChatCompletion

kernel = sk.Kernel()

# Add AI service
kernel.add_service(
    OpenAIChatCompletion(
        service_id="chat",
        ai_model_id="gpt-4o",
        api_key="sk-...",
    )
)

Semantic Function (Prompt Plugin)