Skip to content

Pro Workflows

The features and routines that separate occasional Claude Code users from those who truly get the most out of it: plan before you act, undo without fear, and work in parallel.

Plan Mode

In Plan Mode, Claude explores and investigates your code and proposes a detailed plan, but without touching any files. You review it and, if you're satisfied, give the green light to execute it. It's the best safety net for large tasks.

How to enable it

  • Press Shift + Tab to toggle the mode (press again to exit).
  • Or at startup: claude --permission-mode plan

Checkpoints and Rewind (undo)

Every time you send a message, Claude Code creates a checkpoint (a save point). If a change goes wrong, you can roll back both the conversation and the code to an earlier state.

How to undo

  • Use the /rewind command to choose which point to return to.
  • Press Esc Esc (twice) as a "panic button."

Background tasks

You can launch tasks or subagents that run in the background while you keep working on something else in the main session. Perfect for long-running processes (heavy tests, builds, extensive research).

How to use it

  • Launch in background: claude --bg "your task"
  • Press Ctrl + B or ask it to "run this in the background."
  • Manage processes with claude agents (view, attach, logs, stop).

Output styles

Control how Claude Code responds—especially useful for scripts and automations:

Terminal
# Plain text output (default)
claude -p "summarize the changes" --output-format text

# JSON output (for processing with scripts)
claude -p "list the TODOs" --output-format json

# Streaming JSON output (for live integrations)
claude -p "analyze the repo" --output-format stream-json

You can also define custom styles (e.g., an "explanatory" mode that teaches more—ideal for learning) through configuration or the system prompt.

Most recommended workflows

Here's what the Claude Code community is sharing and recommending as best practices:

  • Plan Mode first, always. Before any serious task: Shift + Tab → review the plan → execute. Avoid surprises.
  • Parallel subagents + background. Delegate research, coding, tests, and review at the same time. Monitor with claude agents.
  • Standardized skills. Create or install skills for review, deploy, and test. Use disable-model-invocation: true on those with side effects (like deploy) so they don't launch on their own.
  • Rewind / Esc Esc as a panic button when something goes wrong.
  • Hooks + MCP to integrate your environment: automatic formatting after edits, connection with Slack or your ticket manager.
  • Intensive review before production. A rigorous review pass (with a powerful model) before pushing important changes.
  • A lightweight CLAUDE.md + specific skills, and version the .claude/ folder in your repository so the team shares the same configuration.

Check your version

Claude Code updates very frequently (frequent releases). If a feature here doesn't show up for you, update:

Terminal
claude --version
npm update -g @anthropic-ai/claude-code

Cross-session messages

Since Claude Code v2.1.232, type @ in the prompt to mention another live session by name. Claude uses SendMessage to pass the finding directly. Use claude agents to inspect available sessions.

Use /config to accept, hold, or refuse incoming messages. Send a concise finding rather than secrets or hidden decisions, and verify the receiving session's result before integration.

Verified source

Claude Code v2.1.232 and v2.1.224 release notes, reviewed August 16, 2026.

Complete Aulafy mapSee how this lesson fits without leaving your path.

Complete Aulafy map

How all courses connect

This is not a checklist. Start with the foundation, choose an outcome, and go deeper only when your project needs more control.

  1. 1Understand
  2. 2Apply or build
  3. 3Operate with confidence
01

Choose an application

Turn the foundation into a visible outcome: a website, a business improvement, media, or an interactive experience.

Continue into the technical branch when you need to maintain code, data, or infrastructure.

02

Build with code

Prepare your environment, work with coding agents, and run models while keeping control of your projects.

This branch prepares you to design and operate reliable AI systems.

03

Take systems to production

Combine retrieval, agents, evaluation, security, deployment, and model adaptation when the problem requires it.

You do not need every course: choose the component your system needs and return as it grows.

View full catalogue