Claude Code in action

What is Cloud Code?

• Cloud Code is not just a code-writing tool—it's like having another engineer working alongside you on a project
• Can delegate all project tasks to Cloud Code: setup, feature design, testing, deployment, and production support

Getting Started with Cloud Code

• Download and extract the project zip file
• Open the project directory in your code editor
• Launch Cloud Code by executing the cloud command in the terminal

The INIT Command

• Scans your entire codebase to understand project architecture and coding style
• Generates a special file called claude. md that stores findings
• This file is automatically included as context in all future Cloud Code requests
• You can optionally add special instructions for Cloud to focus on specific areas
• Three types of Claude files exist: Project, Local, and User memory

Managing claude.md

• Can manually edit the file or rerun INIT to update it
• Use # followed by a note to append small directions to the file
• Choose which memory level to add notes to (project, local, or user)
• Project memory is shared with all team members working on the project

Git Integration

• Cloud Code can stage and commit changes automatically
• Writes descriptive commit messages based on code changes

Two Effective Workflows for Complex Tasks

Workflow 1: Analysis → Planning → Implementation

• Identify relevant files in the codebase
• Ask Cloud Code to read and analyze those files
• Request Cloud Code to plan the solution (without writing code yet)
• Ask Cloud Code to implement the planned solution

Workflow 2: Test-Driven Development

• Ask Cloud Code to read relevant context files
• Request Cloud Code to suggest tests for the feature (no code yet)
• Select the most relevant tests and ask Cloud Code to implement them
• Ask Cloud Code to write code that makes the tests pass

Key Principle: Effort Multiplier

• Cloud Code effectiveness increases significantly with better-structured instructions
• Putting effort into how you direct Cloud Code yields dramatically better results
• Simple directions work, but detailed planning produces superior outcomes

Takeaways

• Cloud Code is an AI engineer that can handle entire project workflows from setup through production support
• The INIT command scans your codebase and creates claude. md to store project context, which is automatically included in all future requests
• Use either Analysis → Planning → Implementation or Test-Driven Development workflows for complex tasks
• Better-structured instructions to Cloud Code yield dramatically better results—effort in planning pays off
• Cloud Code can automatically stage, commit changes, and manage git with descriptive messages