learn.vdmitriev.com
Тема

Claude Code с нуля

От установки Claude Code до собственного агента. Четыре модуля, квизы, практика с проверкой в терминале, без регистрации.

Модуль 1. Как пользоваться Claude Code

Установка, первая задача, CLAUDE.md, режимы разрешений, контекст и стоимость. Без программирования.

  1. 01Getting Started with Claude Code Не начатClaude Code is an agentic coding assistant that operates directly in your terminal. It can read and edit files, run commands, search your codebase, and interact with Git and GitHub -- all through natural language. This guide walks you through installation, authentication, and your first productive session.5 минут
  2. 02Prompt Tips for Claude Code Не начатThe quality of Claude Code's output depends heavily on how you communicate your intent. This guide covers practical techniques for writing clear, effective prompts that get the results you want on the first try.5 минут
  3. 03Writing Effective CLAUDE.md Files Не начатCLAUDE.md is a special Markdown file that gives Claude Code persistent context about your project. Think of it as a briefing document -- it tells Claude your conventions, tech stack, testing commands, and preferences so you don't have to repeat them every session.5 минут
  4. 04Permission Modes Не начатClaude Code operates within a permission system that controls which tools and actions it can take without asking for approval. Understanding these modes lets you balance safety with speed depending on your task and comfort level.15 минут
  5. 05Build a landing page with Claude Code Не начатHands-on practice. Describe your product, let Claude Code build a one-page site, review it and ship it.45 минут
  6. 06Managing the Context Window Не начатClaude Code operates within a context window whose size depends on your model and plan. Opus 4.8 supports a 1M token context window by default on Max, Team, and Enterprise plans. Sonnet and Haiku have smaller windows. How you manage that window directly affects response quality, speed, and cost. This guide covers strategies for keeping context focused and knowing when to start fresh.20 минут

Модуль 2. Один агент: цикл, инструменты, контроль

Как устроен агент для кода изнутри, шаг за шагом на Python.

  1. 01The Agent Loop Не начатOne Loop Is All You Need5 минут
  2. 02Tool Use Не начатAdd a Tool, Add Just One Line5 минут
  3. 03Permission Не начатCheck Permissions Before Execution5 минут
  4. 04Hooks Не начатHang on the Loop, Don't Write into It10 минут
  5. 05TodoWrite Не начатAn Agent Without a Plan Drifts Off Course5 минут
  6. 06Subagent Не начатGive a Subtask Its Own Context5 минут

Модуль 3. Контекст и долгие задачи

Скилы, сжатие контекста, память, система задач и фоновая работа.

  1. 01Skill Loading Не начатLoad Skills When Needed5 минут
  2. 02Context Compact Не начатMake Room Before the Context Fills Up15 минут
  3. 03Memory Не начатKeep Useful Knowledge Across Sessions10 минут
  4. 04Task System Не начатFrom an Execution Checklist to Coordinated Task State10 минут
  5. 05Background Tasks Не начатSlow Operations Go to the Background10 минут

Модуль 4. Автономия и команды агентов

Планировщик, команды агентов, плагины MCP, собранная обвязка, рабочие процессы и целевой цикл.

  1. 01Cron Scheduler Не начатStart Work on a Schedule10 минут
  2. 02Agent Teams Не начатRuntime and Coordination Protocols20 минут
  3. 03MCP Tools Не начатDiscover and Invoke External Tools10 минут
  4. 04Integrated Harness Не начатMany Mechanisms, One Loop15 минут
  5. 05Workflow Runtime Не начатThe Model Decides Each Step; a Script Decides the Orchestration15 минут
  6. 06Goal Loop Не начатThe Model Proposes a Stop; an Independent Evaluator Decides Whether to Continue10 минут