Learn

The difference between a chatbot, a tool, and an agent — and why the distinction decides what you can delegate.

Why constraint beats encouragement, with the four-part structure behind every prompt in our library.

The three properties that decide which one you open daily, and why every leaderboard is the wrong input.

Files, docs, tools, and history — what to hand over, what to withhold, and why more context can make output worse.

The delegation test in full: reversible, verifiable, and describable. Fail any of the three and you want a script or a human.

The problem the protocol solves, in terms of the alternative you were probably using: custom integrations per client.

The config file, the three values that actually matter, and how to know it worked.

A five-step order that eliminates transport, credentials, arguments, and context before you blame the model.

Five checks that take ten minutes and predict most of the pain: maintenance, transport, permissions, blast radius, and install path.

The smallest useful server is about forty lines: two tools, a schema, and a decision about what not to expose.

A folder with a SKILL.md the agent loads on demand — closer to a training manual than a plugin.

The four sections every working skill has, why the fourth one is skipped, and how to test a skill before you ship it.

The upgrade path from a copied prompt to a shared practice, and the point where a skill is the wrong tool.

Five minutes of reading that predicts whether a skill will change behaviour or just consume context.

Where the tokens actually go, why tool definitions are charged on every request, and the three levers that change the bill.

A minimal eval harness: twenty cases, a deterministic check where possible, and a rule for when to re-run it.

What to hand an agent, what to keep behind a human, and how to reconstruct what it did afterwards.

Eight checks that catch the failures which only appear once real people and crawlers arrive.

4 books4 resourcesUnaffiliated Long form Books AI Engineering: Building Applications with Foundation Models ↗Chip Huyen · O'ReillyFor: Anyone shipping a product on top of…