Reference

Install, login, working directory conventions.

Project memory: what to put in it, what to leave out.

Built-in commands and when each one earns a keystroke.

How it works, when to invoke it, how to read plans critically.

Parallel work, specialization, isolation as a guardrail.

Connecting external tools to Claude Code; when it helps vs hurts.

Shell commands that fire before/after tool calls.

Scoping risky operations per-project.

Building your own slash commands for repeated workflows.

Non-interactive runs in pipelines and automation scripts.

Keeping sessions focused and token spend predictable.

System prompts, XML structure, few-shot, prefilling.

Permission scoping, deny rules, test-first loops, eval harnesses.

Common mistakes that waste tokens or cause agents to go off the rails.

Projects, Artifacts, Skills, Connectors, Research mode.

Competitor analysis, customer interview synthesis, ICP definition.

Landing page copy, SEO drafts, launch posts.

Models, caching, batch, tool use, extended thinking — what exists.