Course Wiki

Coding Agents

LLM assistants that read, write, and modify a codebase — the subject AI Dev Tools Zoomcamp disciplines with specs, review, and tests.

A coding agent is an LLM wired into a repository: it reads project files, plans changes, edits code, runs commands, and opens pull requests. AI Dev Tools Zoomcamp opens by comparing the landscape — Claude Code, Codex, OpenCode, Cursor, GitHub Copilot, Aider, Windsurf — and then teaches the workflow that makes any of them safe to use: right context, right job, review the output, test the result.

The course’s stance is capability-first: tools change monthly, but the durable skills are giving agents the right context, breaking work into reviewable tasks, and verifying everything a machine writes. Every later module — MCP, skills, subagents, worktrees — extends or constrains this core loop.

Taught in

• AI Dev Tools Zoomcamp

  • Module 4: DevOps and Observability for AI-Built Apps
    • DevOps and Observability for AI-Built Apps
  • Module 5: Coding Agent Building Blocks: Reusable Skills and Specialized Subagents
    • Module 5 — Coding Agent Capabilities: MCP, Skills, Plugins, and Custom Agents ## Related concepts

• Spec-Driven Development

• Context Engineering

• MCP

• Agent Skills and Subagents

• Loop and Graph Engineering