01 · Coding agents¶
By 2027, working with a coding agent is a baseline skill, like using Git. This module teaches how agents work, how to give them the right context, and how to use them without shipping bugs or causing incidents.
You still need to be able to code. The agent writes faster than you; you decide whether what it wrote is right.
You'll be able to: set up an agent on a real project, write an AGENTS.md and a skill, connect an MCP server safely, and follow a spec-to-review workflow.
| # | Note | Level | Time |
|---|---|---|---|
| 1 | What is a coding agent? | Beginner | 10 min |
| 2 | Context, memory files and AGENTS.md | Beginner | 12 min |
| 3 | Agent skills and SKILL.md | Beginner | 12 min |
| 4 | MCP, the Model Context Protocol | Beginner | 12 min |
| 5 | Subagents and hooks | Intermediate | 12 min |
| 6 | The new development workflow with agents | Beginner | 12 min |
| 7 | How agent-written code goes wrong | Beginner | 10 min |
| 8 | Choosing a coding agent | Beginner | 10 min |
Project: P1 · Agent-built CLI tool. Build a small tool with an agent using a spec, tests first, an AGENTS.md and a custom skill.
Next: 02 · Building agents