From features
to infrastructure.
first
Four features
Or one stack.
MCP showed up. Skills showed up. Hooks showed up. Subagents showed up. Each got a blog post. Each got a Twitter moment. Each got treated as a feature. Almost no one talks about how they compose. That's the gap this course fills.
Used as features, they're nice. Used as a stack, they turn Claude into infrastructure. A wrapped tool layer (MCP) that the Skills know about, with Hooks that catch what Skills miss, plus Subagents for the small set of tasks that need a different shape. That's a different product than "Claude plus a few add-ons." Package it as a plugin and a teammate gets the same stack in one install.
MCP
Where Claude touches your tools. The integration layer.
Skills
Reusable workflows Claude picks up automatically. The behavior layer.
Hooks
Guardrails Claude can't skip. The safety layer.
Subagents
When one Claude is the wrong shape. The shape layer.
Four pillars. One stack. Compose them and Claude works as infrastructure, not features.
Who this course is for.
People who already use Claude well and want to push deeper. Operators with the basics down. Builders past the "just use claude.ai" stage. Anyone who's noticed they keep typing the same prompts, wrapping the same APIs, forgetting the same guardrail. The four pillars eliminate that work.
Some technical comfort assumed (you can edit a JSON config, run a script). You don't need to be a senior engineer to compose any of this. The kit ships with thirteen starter files you fork into your setup.
The single most common mistake at this level: trying to install all four pillars at once on day one. Don't. Install ONE, get a week of use out of it, then reach for the next. The compounding only works if each piece earns its keep before the next one lands.