What is the /loop
command in Claude Code and how can you use it?
I am starting to poke at
MCP (Model Context Protocol) servers.
And how they work and how can you connect to them in Claude.
As a first step I want to look at one aspect of an MCP servers Prompts.
Claude AI is powerful
and it needs to be contained in many cases.
The default mode for Claude code is to ask permissions before it asks.
But if that is all it does that presents two problems. One you have to click enter/yet again and
again and again. And two if you do that
eventually you are going to click a yes when you don’t mean to.
In Sonnet 4.6 of Claude
the default is 200K for the context size.
You can up it to 1
million by tweaking the ~/.claude/settings.json file
In this post I am going to be going over a few Claude commands /context, /clear, /resume, and /compact