Claude Code Foundations

You already use AI. Here's how to run your work on it.

Most people use AI like a smarter search box. Open a chat, ask a question, copy the answer, close the tab. That's a sliver of what these tools do, and it's why a lot of capable people feel stuck at "helpful, but not essential."

Claude Code changes the shape of the work. Instead of one-off questions, you build skills Claude reruns on command, automations that run on a schedule, and connectors that let it read your email, your Notion, your calendar, and act for you. You stop doing the busywork and start directing the thing that does it.

You don't need hundreds of tools or tens of hours. You need the few foundations that make it click, and some practice using them. In four weeks you go from install to a working system you actually use: skills and automations handling real recurring work in your week, plus the judgment to know when to trust the output. No coding required. You finish with an operating dashboard that hands you a digest of your email, meetings, and tasks each morning, built by you, running on your machine.

Foundations first, on purpose. The tools change every month. What you learn here keeps working.

What you'll learn

Set up Claude Code and think like a builder, not a chatter

Install and configure Claude Code, and get comfortable with how your computer's files actually work
Write a CLAUDE.md file that gives Claude lasting context about you and your work, plus memory files it reads every session
Pick the right model for the job, and read token costs so a bill never surprises you
Pull free, open-source tools from GitHub instead of paying for another subscription

Build skills and automations that do real work

Create your first custom skill: a workflow Claude reruns on command
Set up scheduled automations that run without you, like a morning digest or a weekly report
Build loop and goal skills that work a long task on their own until they hit a bar you set, like checking a draft against a quality checklist before stopping
Stand up a content pipeline (research, draft, review, publish) and a research skill that scans TikTok, Instagram, and X for what's trending

Connect Claude to the tools you already use (MCP)

Wire up Notion, Slack, Gmail, Google Calendar, Google Sheets, Google Analytics, Chrome, and iMessage
Turn a voice note or a rough paragraph into a full content brief, pushed straight into Notion
Pull the story out of your Google Analytics or Instagram numbers by asking, no dashboards

Run advanced workflows and let Claude operate

Split big jobs across sub-agents
Use goal-based tasks, memory, and the prompting patterns that make Claude reliable instead of flaky
Build an operating dashboard: one daily digest from your email, meeting notes, and tasks
Automate the meeting loop end to end: Zoom recap into a Notion notes database, tasks created from the notes, Claude starting on them
Set up a Slack watcher that follows threads, researches answers, replies to teammates, and pings you only when there's a real decision or its confidence is low

Who teaches this

Dylan, the engineer. Founder and AI strategist, currently leading AI adoption at a billion-dollar tech company. He's built software products for over a decade and used AI in his day-to-day for the past two years (Techstars x Cedars-Sinai, 2016). He builds and leads the cohorts, so what you learn comes straight from someone using these tools in real work every day. He brings a practitioner's take, not theory.

This is the course the rest of the catalog builds on. It's also where MANA Labs' own systems come from, including a full sales-pipeline setup built for a client's interior design studio that you'll see as a teaching example.

Who this is for

You use AI every day but suspect you're scratching the surface
You're a founder, leader, or creative who wants leverage, not a CS degree
You're tired of stitching five subscriptions together to do what one well-set-up agent could
You want a room where you build real systems, not watch demos

Not for you if you're already writing custom skills, running MCP servers daily, and scheduling self-checking automations. You'd be ahead of the room.

Prerequisites

A computer you can install software on (Mac or Windows), ideally not an admin-locked work laptop
A personal Claude account with Claude Code access and room to practice daily
Willingness to use the terminal. You won't write code, but you'll run commands, and we walk you through each one
About 4 to 6 hours a week

Setup instructions and an optional install office-hours session come before Week 1.

What's included

Live weeknight sessions, small enough for real guidance
Replays and lifetime access to the materials
Written guides with screenshots for every workflow, so you can rebuild it later
A prompt and skill library you keep and adapt
A cohort of peers building alongside you
Weekly office hours for your actual setup
Certificate of completion

Course syllabus

Week 1 — Core setup and foundations

Get installed, get oriented, give Claude a memory.

Cohort kickoff (Mon, 30 min): how the four weeks work, what you'll walk away with, your first goal
Optional: installation and setup office hours (Tue/Wed, 2 hrs): everyone's environment working, no one left behind
Live session: foundations (Fri, 90 min): files made simple, your CLAUDE.md and memory files, choosing models, token costs, and free tools from GitHub
Homework: stand up your CLAUDE.md and automate one small real task

Week 2 — Skills and automations

Build things Claude reruns for you.

Lab #1: your first skill (Mon, 60 min): build a reusable skill together, live
Optional: office hours (Mon, 60 min)
Live session: automations and pipelines (Fri, 90 min): scheduled automations, loop and goal skills with acceptance criteria, and a full content pipeline
Homework: ship one skill and one scheduled automation on your own work

Week 3 — Connectors and integrations (MCP)

Plug Claude into your actual tools.

Lab #2: connect your stack (Mon, 60 min): wire up Notion, Gmail, Calendar, and Slack live
Optional: office hours (Mon, 60 min)
Live session: working across your tools (Fri, 90 min): voice-note-to-Notion brief, analytics pull-throughs, Sheets and Chrome connectors, and keeping it secure
Homework: connect three tools and build one cross-tool workflow

Week 4 — Advanced workflows and your operating system

Let Claude operate, then show your work.

Lab #3: sub-agents and goal tasks (Mon, 60 min): split a big job across sub-agents; advanced prompting and memory
Optional: office hours (Mon, 60 min)
Live session: build your operating dashboard + demo day (Fri, 90 min): the morning digest, the automated meeting loop, the Slack watcher, then everyone demos their system
Deliverable: your own AI operating system, documented and running

Schedule

Live sessions: about 2 hrs/week, plus optional office hours
Hands-on practice: 2 to 3 hrs/week on your own work
Total: about 4 to 6 hrs/week over 4 weeks

FAQ

Do I need to know how to code? No. You'll run some commands and read some config files, all walked through. If you can follow a recipe, you can do this.

Will this work on my work computer? Usually. Locked-down corporate laptops sometimes block the tools we use, so a personal machine and personal Claude account run smoothest. We sort this out in the setup office hours before Week 1.

What if I miss a live session? Every session is recorded and yours for life. Office hours are where you catch up on anything hands-on.

How is this different from a prompting course? Prompting is one week of this. The rest is building: skills, automations, connectors, and agents that do work while you're not watching.