Claude Ecosystem Guides
Short, evergreen explainers for the building blocks and everyday techniques of engineering with Claude — what each one is, why it matters, and how they fit together.
The first four guides are the working vocabulary of building with Claude, and they stack on each other. Tool use is the foundation: it lets the model call your functions and read back the results, which is what turns a chat model into something that can act. The Model Context Protocol standardises how those tools and data sources are exposed, so one integration is reusable across many applications instead of rewritten for each. The Agent SDK packages the entire loop — with built-in tools for files, shell and search — so you can ship a working agent on your own infrastructure. And extended thinking is the reasoning layer beneath all of it: the model works a hard problem out before it answers.
The other four are the everyday craft. Prompt engineering is the cheapest lever of all — clear instructions, examples and structure usually beat anything heavier. Retrieval-augmented generation grounds answers in your own documents instead of the model's trained-in memory. Computer use lets Claude drive a desktop by screenshot when there is no API to call. And prompt caching cuts the cost and latency of any request that reuses a large, stable prefix.
Read them in any order — each page is self-contained, free to read in full, and links to the others where the concepts meet. When you want to practise instead of just read, the interactive Master Claude track turns these into questions, spaced-repetition cards and a progress record.
The guides
Work through it interactively
Every area has questions, spaced-repetition cards and a progress record. Those need an account, which is free and takes a moment.
Open the interactive track Create a free account