Tracks · Learn Claude
Is this a prompt engineering course for Claude?
You get reliable output from Claude by constraining the format first and pushing on quality second — not by rewriting the request until something usable falls out. This track is prompt engineering for Claude as a set of decisions rather than tricks: where a rule belongs, when to ask for reasoning, how to get valid JSON without retrying, and how to spot a prompt injection in text you did not write. Assumes Foundations.
What a prompt engineering course cannot do
No. A prompt engineering course hands you slides and a certificate of attendance. This track puts you in the decisions — the coach asks what you would do, tells you when your reasoning is thin, and builds flashcards from the places you fumbled. You end up able to prompt, not able to describe prompting.
What you learn in this track
System vs. user vs. assistant messages
The system message sets standing rules; the user message asks for one thing. Mixing them is the most common prompt bug.
Structured output & JSON mode
Getting valid JSON out of Claude is a schema problem, not a wording problem.
Tool-use vs. response_format
Use tool calls when Claude must act, structured output when it must only answer in a shape.
Few-shot selection & anchoring
Two well-chosen examples beat ten average ones, because Claude anchors on what your examples have in common.
Chain-of-thought vs. direct answer
Ask for reasoning when the answer is a judgement, and for the answer alone when it is a lookup.
Constrain-then-coax patterns
Constrain the format first, then push on quality — the reverse order loses both.
Prompt caching & format stability
Prompt caching cuts cost, but only if the stable part of your prompt really is stable.
Prompt injection — spotting and defending
Any text Claude reads is untrusted input, including the document you asked it to summarise.
Guides in this track
How the Socratic coach teaches this
Each module is a Socratic dialogue: the coach probes your reasoning with scenario questions, surfaces weak points, and generates flashcards and study notes from your own gaps. Mastery is tracked per concept, so you can see which ideas you actually handle and which ones still need work.
See where this track sits in the curriculum on the six tracks and why they are in this order, or open the ClaudeMaster.ai hub to start it.