A hands-on training

Building with Agents

Skills, subagents, slash commands, hooks, MCPs, CLIs: there are many ways to setup and extend an AI agent. Learn what each one is, how to structure them, and when an agent should use them.

Start the course →
~44 min· Interactive examples

Context is key.

System prompt, your input, the agent's thinking, every skill: they all share one finite window. Press the ▶ Simulate Interaction button to watch a simulated agent interaction. You'll go deeper in Module 2.

The context window
active memory 0 / 256k tokens
  1. System prompt + AGENTS.mdalways loaded
    You are an agent. Tools: read, edit, bash… Follow the house rules. Be concise.
  2. Skill Libraryalways loaded

    Available Skills

     design-guide: Apply the house visual system… coding-standards: Applies the coding standards… commit-message-format: Shows the agent how to format commit message
  3. User inputyour turn
    > Add a dark-mode toggle to the settings page.
  4. Agent thinkinggrows as it works
    The toggle needs a saved preference… read settings.js → found themeContext editing settings.js… read theme.js, storage.js, 38 more files… + dozens of tool results, reasoning over many turns
  5. Compactionreplaces the conversation
    ↺ user input & agent thinking summarized: “built dark-mode toggle, wired to themeContext, tests green” system prompt & skills stay; the bulky turns are folded away

↑ Press play, or tap a section, to watch each one enter the same finite 256k window, and what compaction does when it fills.

Jump to module →

Why it works

Load a little.
Then a little more.

Skills stay cheap through Dynamic Context: three tiers that load only as far as the task needs. You'll build one in Module 3.

Jump to the module →
  1. 1

    Metadata

    At startup, only name + description load. Enough to decide relevance.

  2. 2

    The body

    When the skill becomes relevant, its SKILL.md body is read.

  3. 3

    Bundled files

    References and scripts open only when actually needed.

The course

Six modules. Build as you go.

Start with The Agents Active Memory, end able to compose a real multi-part system, and connect it to the outside world. Every module ends with a hands-on build and a quick check.

Ready when you are.

The first module takes about six minutes and needs nothing but curiosity.

Begin Module 1 →