Builder Methods library
Become AI-Native
Work faster, think sharper, and create more, with AI woven through everything you do.
Example 1: The Consultancy
Turn a document you keep rewriting from scratch into a skill that asks the right questions, then rebuild it from interchangeable parts.
Consultancies with AI
A quick tour of the first example business, a consultancy that sells operations expertise, and the proposal process we're about to turn into a skill.
Build an agent skill
From an empty folder to a self-describing repo and a working proposal skill built from two markdown files.
Test the agent skill
We run the new proposal skill on a real prospect and read the draft closely: what it got right, and the one choice it made without asking.
Human in the loop
The improvement half of the loop: give the skill a short interview so it asks for the judgment only you have, then re-run and compare the two drafts side by side.
Modular agent skills
The skill writes one kind of proposal well; now we restructure it into a library of interchangeable sections, step files, and a shared training folder.
Friction-driven iterations
We run the modular skill on two new prospects and let it hit the limits of the library on camera, patching each run by hand and noticing what keeps coming up.
Grow your skills library
Kill the modules nobody used, promote the sections you kept asking for, and run the skill clean: real usage rewrote the design in one pass.
Agent Skills Patterns
The first pattern lesson names what you just watched: the loop, where skills live, the anatomy of a skill, the interview convention, and the three improvement moves.
Example 2: The Content Business
Build a drafting skill, train it on your voice and taste from finished work, and connect it to the platform that sends the thing.
Content creation with AI
A quick tour of the second example, a one-person newsletter business where the writer's voice is the product, and what this module builds for it.
Content development repo
An empty folder becomes the working home for the newsletter: nine finished issues, raw idea captures, brand docs, and an AGENTS.md that explains it all.
Content drafting skill
We build a newsletter drafting skill that interviews you before it writes, run it on a raw voice memo, and read a draft that has the right shape and none of the voice.
Voice training for AI
Point the agent at nine finished issues to build a voice doc, seed a slop list by hand, wire both into the skill with a self-check, and run it again.
MCP & Agent Skills
The last mile: add a push step so the finished issue lands in your email platform as a draft, then swap platforms and watch nothing in the skill change.
Training AI with taste
The second pattern lesson makes the course-wide voice thread explicit: two mechanisms, a self-check, where the training lives, and the kit for doing it on your own work.
Example 3: The Sales Pipeline
Build a file-based CRM and the skills that run it, then put the pipeline on a schedule and steer it from the report.
AI Sales Pipeline
A quick tour of the third example, a managed IT provider with founder-led sales, and the biggest build in the course: a sales assistant with a real job.
Sales Pipeline Repo
A CRM with no app, no database, and no server, from one prompt: six numbered stage folders, eight prospects filed into their stages from a seed doc, and an AGENTS.md that tells any agent how the pipeline works.
Leads intake skill
Build the intake skill that files every inbound item, then run it on two days of forms, emails, voicemails, and a call transcript.
Daily report skill
The fix for sifting: two report skills that surface what needs you, plus the first skill that calls another skill, then the rest of the week's inbound.
HTML sales dashboard
A visual board that reads the repo: one HTML page, no app, no database, no server, kept fresh by a skill that other skills call.
Automate agent tasks
The same three skills, run by a scheduler instead of by you: hourly intake, a daily report, a weekly report, and a week that runs while you do nothing.
Async interactive agents
Turn the daily report into a two-way command surface: the agent recommends and leaves you checkboxes, you tick or scribble, and the next run carries it out.
Agent employee patterns
The third pattern lesson names the course's biggest idea, the agent employee, assembles the whole AI-native operating model, and hands the loop over to you.
The Reference Shelf
Short standalone explainers on the file types, apps, GitHub basics, and costs behind everything in the course.
File types for AI
Markdown, JSON, HTML, YAML, and scripts: what each file type you saw in the course is for, and why none of them belongs to a company.
Agentic apps & harnesses
How to make sense of the AI apps out there: chat apps vs file-based apps, UIs vs terminals, agent bots, and the three ways an agent reaches your other tools.
GitHub repos, explained
What a repo is, why syncing to GitHub is worth it, the three words you need (commit, push, pull), and the one thing that must never go in a repo.
AI costs
How AI pricing is structured, tokens versus subscriptions, how I mix models and plans, and the progression I recommend as your usage grows.