Components of A Coding Agent

This article by Sebastian Raschka explores the design and components of coding agents, such as Claude Code and Codex CLI, focusing on how they leverage tools, memory, and repository context to improve LLM effectiveness in practical coding tasks. It distinguishes between LLMs, reasoning models, and agents, explaining how agent harnesses enable better tool use and context management. The piece covers six main building blocks of coding agents, emphasizing that recent progress in LLM systems relies not just on better models but on surrounding systems. It serves as a reference for readers interested in agentic coding tools and their real-world applications in IT/technology.