Prompting
Prompt engineering and LLM interaction patterns
CLAUDE.md gives Claude the overview, but rule files handle the details. They're focused documents in .claude/rules/ that Claude reads automatically based on what it's working on.
Some prompts you'll type over and over. Custom commands let you save them as reusable recipes. Type a short name instead of pasting a paragraph every time.
Every time you start Claude Code, it forgets who you are. CLAUDE.md is a text file that gives it permanent memory. One file, read automatically at the start of every session.
Set up your screen so you can see Claude Code and your project at the same time. This one change makes the whole experience feel different. Terminal on one side, results on the other.
Claude Code can split itself into multiple copies and work on several things at the same time. You don't need to manage this. Just describe what you want and Claude figures out the rest.
The quality of what Claude Code builds depends entirely on the quality of your instructions. Vague instructions get vague results. Specific instructions get exactly what you want.
When you start Claude Code from a folder, that folder becomes its world. Everything inside it is visible, everything outside doesn't exist. This is the most important thing to understand.