# Foundations

This lane explains the core concepts behind agent systems before readers need
to care about any specific framework, vendor, or implementation stack.

## What belongs here

- What agent systems are
- When not to use agents
- Agents versus deterministic workflows
- Tool use and environment boundaries
- Memory and context at a conceptual level
- Single-agent and multi-agent basics

## Editorial intent

Start with universal questions, plain language, and durable mental models.
This is the first stop for global readers who want conceptual clarity before
they go deeper into implementation details.

## Current pages

- [The Agent Systems](./the-agent-system.mdx): a practical definition
  of agent systems as software loops, not just chat surfaces.
- [Agents Vs Workflows](./agents-vs-workflows.mdx): when to use deterministic
  orchestration, bounded autonomy, or a hybrid design.
- [History Of Agent Ideas](./history-of-agent-ideas.mdx): a short problem-driven
  history from symbolic systems to modern agent stacks.
- [LLM Foundations For Agent Systems](./llm-foundations-for-agent-systems.mdx):
  the minimum model concepts needed to design agent systems well.