Skip to content

Part 1

Overview

Before diving into patterns and architectures, you need a solid understanding of what agents are, how they differ from traditional software systems, and the fundamental components that make them work.

This section builds the mental models you'll use throughout the rest of this knowledge base.

What You'll Learn

  • Agent Anatomy - The components that comprise an agentic system
  • Agent Loop - The perception-reasoning-action-reflection cycle
  • Design Philosophy - Core principles for thinking about agents
  • Historical Context - How we got here from classical AI

Quick Concepts

Agent Formula (Lilian Weng)

Agent = LLM + Memory + Planning + Tool Use

Core Loop

Perceive → Reason → Act → Reflect → Repeat

Key Distinction

System Type How It Works When It Acts
Chat Responds to user input Only when prompted
RAG Retrieves and augments response Only when prompted
Agent Autonomously pursues goals Continuously until goal met

Chapter Map

01-Agent-Anatomy
 - What is an agent?
 - Core components
 - Agent vs other systems
 - Anatomy breakdown

02-Agent-Loop
 - Perception phase
 - Reasoning phase
 - Action phase
 - Reflection phase
 - Loop closure

03-Design-Philosophy
 - Principle 1: Autonomy with Oversight
 - Principle 2: Composability
 - Principle 3: Observability
 - Principle 4: Safety by Design
 - Principle 5: Explicit Intent

04-Evolution-From-Classical-AI
 - Classical AI approaches
 - Learning from history
 - Why LLM agents differ
 - The 2025 turning point

Key Takeaways

Agents are autonomous decision-makers with 4 core components The perception-reasoning-action-reflection loop is the runtime Agency requires separating planning from execution LLM-based agents succeed where classical systems struggled

Prerequisites

  • Basic understanding of LLMs (how they work, what they can do)
  • Familiarity with concepts like API calls, databases
  • Optional: Experience with Python or other programming languages

How to Read This Section

  1. Start with 01-Agent-Anatomy.md - Understand the structure
  2. Move to 02-Agent-Loop.md - Understand the behavior
  3. Read 03-Design-Philosophy.md - Understand the mental models
  4. Optional: 04-Evolution-From-Classical-AI.md - Understand the context

Each file builds on the previous one, but can also be read independently.

-

Next Steps

After mastering Foundations:

-

Last Updated: August 9, 2026