https://anthropic.skilljar.com/claude-code-in-action

Assuming you already installed and is authenticated...

Go to your project and run claude.

create a CLAUDE.md file

First command to be used: /init. It creates the CLAUDE.md file. Review it and make adjustments you deem necessary.

You can use # <instruction> to make claude update the CLAUDE.md file with more instructions

You can use @path/to/file to reference a file from inside CLAUDE.md.

You can paste images directly in the text input field.

Once you're happy with the changes, ask claude to commit the changes.

context management

go back to a previous point of conversation

Press Esc Esc (yes, press it twice) to jump to a previous point of the conversation and fork from it. This is useful when parts of your conversation are just noise and you want to remove them from the context history.

compact the context

Another way to decrease the amount of tokens used to keep the context, is by using the /compact command. This is ideal when:

• Claude has gained valuable knowledge about your project
• You want to continue with related tasks
• The conversation has become long but contains important context

Use compact when Claude has learned a lot about the current task and you want to maintain that knowledge as it moves to the next related task

clear the context and start a new one

If you want to start a brand new task, use /clear to remove all the context and start a new one.

custom commands

Create a file inside .claude/commands/<command_name>.md and add instructions inside the file.

Restart claude and you'll see the /<command_name> is now available.

Your custom command can also receive arguments. Just create a custom command referencing $ARGUMENTS, like in this example:

Write comprehensive tests for: $ARGUMENTS

> more instructions here...

MCP integration

TBD...