Explore patrickfrank.com.
Operations
SaaS Software Stacking Email & Retention Flow Creation Reporting & KPI Goal Setting SEO & Keyword RankingGrowth
Content Creation Planning Omni-Channel Strategy Launching Paid Ad Campaigns Brand DevelopmentWork with me
1-on-1 Strategy Sessions Fractional CMOFree stuff
The Secret Menu Free tools The Founder Scorecard The 30-day planner The newsletter ↗Case files
PatientPartner Toned by Tony Float Factory Conscious Body RecoveryFour short courses for the four ways people work with Claude. Newcomers learning the ropes. Builders shipping agents. Solo founders writing code. Power users wiring up MCP, Skills, Hooks, Subagents and plugins. ~40 kit files you install in an afternoon. Bundle saves ~20% versus buying one at a time.
$99 once. Saves $24.96 vs à la carte. 30-day refund, no questions.
If any of these sound familiar, you're not alone. They're the four predictable failure modes that show up six months after every Claude rollout.
Your team types one-line prompts, gets generic drafts, sends them to you to fix. Your inbox is now a prompt-rewriting service. The force-multiplier you bought is running at 10%.
You've watched the YouTube videos. You've forked the SDK starter. Six weeks in, nothing's deployed. The gap between "agent that runs in a notebook" and "agent that ships work for two weeks under cost guardrails" is a discipline problem, not a code problem.
You're the only engineer. You should be shipping in days. Instead Claude Code is breaking what worked yesterday, your CLAUDE.md is one terrible paragraph, and you spend more time debugging the model than building features.
Your custom workflows live in tabs. Your skills live in your head. Your hooks fire on nothing. Claude opens cold every time. You know there's an extension system. You haven't connected the four pieces yet.
No bundle is for everyone. Here's exactly who this is built for, and who should pass.
Each course is short on purpose. Each ends with a kit you install. Pick one or get the bundle. The math beats à la carte once you'd buy more than one.
Stop watching agent demos. Build six. Sales, Content, Ops, Finance, Hiring, Research. Six business agents in TypeScript, plus the framework to build any agent your business needs. Cost guardrails included.
After: your first agent runs for 2 weeks under cost guardrails. The other five wait their turn. Patrick's Law.
The setup ritual + universal preferences your team installs once, benefits from forever. 30-minute install, 24 modules.
Never use em-dashes. Use commas, parentheses, or separate sentences.
Skip filler openers. Start with the answer.
Push back when something's off. Don't execute on a flawed premise.
Slash commands, hooks, subagents, and the CLAUDE.md most builders never write properly.
--- /plan-first skill --- Plan first. Don't write code yet. 1. Shape: what's the diff? 2. Data: in / out / shape? 3. Edge: name one. 4. Test: "works when X"
MCP, Skills, Hooks, Subagents. Four extension systems most users treat as separate features. Used together, they turn Claude from a tool into infrastructure. Pack them into a plugin and your team installs the whole stack at once.
After: Claude opens with your context, your tools, your skills already loaded. Every session. Without asking.
Real .ts agents you fork. Real .md templates you paste into your CLAUDE.md. Real hook scripts that fire on real events. Here's the inventory and one file in full.
Browse the inventory below by course. Then scroll down for an inline preview of one actual file (the CLAUDE.md template from the Code course). No mockups, no marketing screenshots. The actual content.
# [PROJECT NAME] · CLAUDE.md > The brief Claude Code reads first when you open a session in this > project. Replace every bracketed placeholder with your project's > specifics. Treat this as living documentation. Update when you'd > update the onboarding doc for a new hire. --- ## Project context [Two paragraphs. What is this project? What problem does it solve? Who uses it? Why does it exist?] [Example: "RevSignal is a B2B SaaS that surfaces revenue-at-risk signals from CRM data for sales leaders at 50-500 person companies. Sales leaders use it to triage which renewals need attention this quarter. Built because most CRMs surface what's open, not what's slipping."] --- ## Conventions [How code is structured here. The opinionated decisions you've made that look arbitrary without context.] - File layout: [Routes under src/routes/, business logic under src/services/, types under src/types/.] - Naming: [Functions camelCase, types PascalCase, constants SCREAMING_SNAKE.] - Tests: [Live next to the file as *.test.ts. Run with pnpm test.] - Style: [Functional preferred. No classes except for domain models.]
Forty-two files. Each one replaces about an hour of figuring it out yourself.
$99 once. Saves $24.96. 30-day refund.
The bundle isn't theory. Here's the timeline buyers report, day by day, from install to "wait, this paid for itself a month ago."
10 minutes per course. CLAUDE.md is in your repo. The slash commands are available. The first agent template is forked into your project. You haven't shipped anything new yet, but you have everything you need to start.
You write your first proper CLAUDE.md. Your first agent runs against real data with cost guardrails on. You stop re-pasting context every chat. The kit moves from theoretical to operational.
Patrick's Law kicks in: no second agent until the first ships for two weeks under guardrails. Your team adopts the universal preferences. The pattern interrupt fires when someone skips the fundamentals.
The agent that was going to be a $4k/mo SDR is running on $40 of API. The CLAUDE.md you wrote saves an hour a week per dev. The bundle's $99 is the cheapest line item in your AI budget, and you'll never re-pay it.
No, this isn't a guarantee. The discipline is on you. The kit makes the discipline cheap. That's the whole bet.
A receipt is more honest than a marketing chart. Here's the receipt.
Includes ~40 kit files, every future course, lifetime updates, and a 30-day refund.
That's ~$2.40 per kit file you don't have to write yourself. Each one replaces about an hour of figuring it out.
$99 once. Saves $24.96 vs à la carte. 30-day refund, no questions.
Thirty days to read it, install it and use it. Not worth it? Reply to your receipt. You get refunded. No questions, no "are you sure" steps. The kits are yours regardless. They're already on your machine.
You're paying $99 for a four-course bundle from someone whose face you may not have seen before. The math should make you skeptical. So here's the math reversed.
If the kit doesn't pay for itself the first week you use it, I'd rather refund you than have you angry on the internet. If I have to talk you into keeping it, the kit isn't doing its job.
Thirty full days. Read every module, install every kit, run it on real work. Then decide.
You keep the kits after the refund. The .md files, the TypeScript, the hook templates, the MCP server. All of it. That part isn't a hostage.
One reply gets it done. Reply to your receipt. No form, no ticket, no "have you tried module 4 first" gauntlet.
If $99 still feels like a lot, here's the comparison that makes it obvious. Three paths to a Claude setup that works. Only one of them costs less than your weekend.
| Option ADIY, the hard way | Option BHire a consultant | Option CThe bundle. $99 | |
|---|---|---|---|
| Time to first deployed agentFrom decision to running, with guardrails on. | DIYA weekend or three | Consultant~2 weeks, then the followups | BundleSame day. Fork the template, configure cost cap, run. |
| Cost upfrontWhat hits the credit card today. | DIY$0, plus your weekend | Consultant$3,000+ for an 8 to 12 hour engagement | Bundle$99 once, never again |
| Cost ongoingWhat it takes to keep iterating after launch. | DIYEndless evenings, Googling, redoing | Consultant$500+/hr for tweaks. They're booked. | Bundle$0. Re-read the module. Update the kit. |
| What you actually getThe deliverable you walk away with. | DIYWhatever you cobble together from blog posts and Twitter threads | ConsultantA pitch deck, one Zoom, and a Notion page | Bundle~40 kit files + 4 written courses written by an operator for operators |
| Risk on bad outputWhat happens if the work doesn't pay off. | DIYYours alone. The weekend's gone either way. | ConsultantTheirs but they're gone Friday | Bundle30-day refund. Reply to your receipt, I refund. You keep the kit. |
| Time savings on next projectWhat carries over the second time around. | DIYRebuild from zero, or copy your own brittle hack | ConsultantRehire, repay, repeat | BundleThe kit is yours forever. Fork it into every future project. |
Note: the bundle is a complement to your $20/mo Claude Pro subscription, not a replacement. It teaches you to actually use the tool you're already paying for.
Three early adopters who deployed pieces of the bundle with their teams. Their specific win, their numbers, their words. Names redacted, numbers real.
I'd been the only person on my team who could get a usable answer out of Claude. Gave them this on a Friday. By Wednesday, three different people had built things I would've assigned to myself.
Built three of the agents from the kit in a weekend. The sales agent already covers what we paid an SDR $4k a month to do. Cost guardrails kept the bill under $40 the first month.
The CLAUDE.md template alone was worth the bundle. I'd been trying to write one properly for months and getting it wrong every time. Used it Monday, shipped two features by Friday.
Used the kit with your team? Reply to your receipt and tell me what changed.
Pricing wasn't a marketing call. It was a math call. Three honest scenarios where the bundle pays itself back inside two weeks. Pick the one closest to your stack and run the numbers yourself.
Template: sales-agent.ts in the Agents kit · cost guardrails included.
Template: claude-md-template.md in the Code kit.
Replaces: the calls, the Notion docs, the Twitter-thread research nights.
Free wouldn't move you. $399 would feel like a workshop you'd schedule for next quarter. $99 is the smallest number that gets read end-to-end and installed by Tuesday. The bundle isn't priced for what it costs to make. It's priced for what it returns.
$99 once. Saves $24.96 vs à la carte. 30-day refund, no questions.
Click any question to expand. If yours isn't here, book a quick call and ask me. I'll answer it and add it.
I've been running AI for my own team and my clients' teams for a few years now. Every time I onboard a new operator, builder, or solo founder, I find myself explaining the same things. Context, planning, the magic phrase. Cost guardrails. CLAUDE.md. Why your hooks aren't firing. Over and over.
So I wrote it all down, made it look nice enough that people would actually read it, and priced the bundle low enough that there's no real reason not to put it on every operator's, builder's, and founder's desk.
If you don't find it useful, reply to your receipt within 30 days and I'll refund you. That's not a marketing line. It's the whole deal. I'd rather have zero unhappy $99 customers than a hundred I squeezed for an extra fifty bucks.
New skills and agent templates every month, with install instructions. Drop your email to know when it ships. You’ll get my newsletter too. Unsubscribe anytime.
The question is whether they're getting $99 of value out of it or thousands. The bundle makes it the second one. 30-day refund if it doesn't.
Charges show as BrandRamp, my agency.Checkout is coming soon. Every course’s first pages are free to read now.$99 once. Saves $24.96 vs à la carte. 30-day refund, no questions.
