Hey --
Quick honest update: this newsletter went quiet because I was heads-down building.
Not "building" as in thinking about building. Building as in: I set up a team of AI helpers inside Claude Code that now runs my product business, a game studio, and the blog you might have found me through. Some of it worked right away. Some of it broke in ways that taught me more than any tutorial.
The stuff I learned is too useful to keep in my notes. So I started writing it down.
Claude Code Decoded -- a new series on the blog
Here's the problem I kept hitting: Claude Code can do a lot, but the official docs read like reference manuals. They tell you what exists. They don't tell you when to reach for it, what it feels like to get wrong, or which thing you actually need when you're staring at three options that sound the same.
The subreddits are full of people asking those exact questions. Good questions, real answers available -- just not written in plain English anywhere.
So I'm writing them. One stuck question at a time, with a straight answer in the first two sentences and notes from my own setup to back it up.
Here's what's live now:
What Are Claude Code Subagents? Subagents are like calling in a specialist. You give them a narrow job, they do it in their own workspace without cluttering yours, and they hand back a result. The article covers when that's worth the trouble and when it's overkill. https://echov4ult.com/blog/what-are-claude-code-subagents/
Skills vs. Subagents vs. Commands -- Which Do You Need? Think of it this way: a skill is a recipe card, a subagent is a second cook, and a command is the bell you ring to start the order. If you're not sure which one you need, you probably want a skill. The article explains why, including the mistake I made building a giant roster because it looked impressive. https://echov4ult.com/blog/skills-vs-subagents-vs-commands/
And dropping this week:
How to Build Your First Claude Code Subagent -- the actual steps, not the concept. Where to put the file, what to put in it, and the one question to ask yourself before you start.
What is CLAUDE.md? -- the instruction file that tells Claude Code how to behave in your project. What goes in, what doesn't, and what I've learned from months of editing mine.
Claude Code Hooks Explained -- automatic actions that trigger before or after Claude Code does something. When that's useful and when it creates more problems than it solves.
Three articles this week. The series keeps going from there. Each one answers a question I've seen someone actually get stuck on.
One more thing
Everything I write about in these articles -- the specialist agents, the skills, the rules that keep them in line -- I ship a pre-built version of all of it as a product called Full Party. It turns Claude Code into a ten-person AI game studio. There's a free demo if you'd rather see the pattern working than read about it. I'll mention it where it's relevant, but this newsletter is about the questions, not the product.
What are you stuck on?
I'm building this series around real questions, not keyword lists. If you're using Claude Code and something isn't clicking -- or if you're curious about it but haven't started yet -- hit reply and tell me what's on your mind. Your question might become the next article, and either way I'll give you a straight answer.
-- Chris