A free, MIT-licensed pack of fourteen agent skills walks a solo technical founder through positioning, first users, launch, and pricing, which is the same scope a senior fractional CMO bills $200 to $350 an hour to cover.
A technical founder ships a dev tool that works, and then hits the part nobody warned them about. Who is this actually for. What does the homepage say. Where do the first fifty users come from. What do you charge. The usual fix is to rent judgment: a senior fractional CMO bills $200 to $350 an hour, and early-stage retainers start between $2,000 and $5,000 a month. The GTM Co-Founder, which topped Product Hunt's daily leaderboard on August 8, is that scope of work compressed into fourteen files, MIT licensed, free, running inside the coding agent the founder already has open.
The mechanic is worth understanding because it is not a prompt pack. The repository ships a set of agent skills organized by stage: discovery skills like who-is-this-for and talk-to-users, positioning skills like positioning-and-story and value-prop-that-converts, then first-50-users, launch-it, pricing, founder-led-sales, and a know-if-its-working skill for after the launch. A review-the-work gate runs throughout. You answer five questions once, the agent writes a founder brief to a file, and every skill after that reads the brief back before it does anything.
That file is the whole trick. Anyone who has run a positioning exercise inside ChatGPT knows the failure mode: session one produces a sharp answer about who the product is for, session four produces a homepage that contradicts it, and by session nine the pricing page is aimed at a completely different buyer. The brief-on-disk pattern is the cheap fix for that drift, and it is the same reason the good consultants write a positioning doc before they touch the copy. Install is npx skills add AIDevGTM/gtm-cofounder, or you can skip the terminal entirely and paste a skill's text into whatever assistant you already pay for.
What it costs against what the engagement costs
Free. MIT. Fork it, gut it, ship it into your own repo.
Set that against the thing it is standing in for. Take a narrow scope, positioning plus a launch plan, and call it twenty hours of senior advisory at Growtal's published range. That is $4,000 to $7,000. Take the retainer route at the low end and three months of an early-stage engagement runs $6,000 to $15,000. The skill pack's competing bill is your existing agent subscription plus whatever tokens fourteen interview-style workflows burn, which for a solo founder is a rounding error against either number.
Now the honest part, because the arithmetic above is doing something slightly dishonest. What $10,000 buys is not the framework. The frameworks have been public for years, and this repo says so out loud, crediting playbooks from Adam Frankl and Jakub Czakon. What the money buys is a person who has watched forty of these launches, who will tell you the thing you do not want to hear, and who is on the hook when the plan is wrong. A skill file does none of that. It will run the interview, structure your answers, and produce a roadmap that is internally consistent with whatever you told it, including the parts where you were flattering yourself.
Who it is wrong for
The repo scopes itself tightly and deserves credit for it: this is built for a solo technical founder spending their own money on a dev tool or AI product. Almost everything about that scope is load bearing.
If you already have a marketing team, this is a worse version of a conversation you can have in a room. If you are not selling to developers, roughly half the content is aimed at the wrong buyer, and market-to-devs-sell-to-buyers is not a skill a regional services business needs. If you are funded, you are optimizing the wrong constraint, because your problem is speed and coordination, not access to a playbook. And if what you actually need is someone to disagree with you, an agent reading a brief you wrote about yourself is the last tool you should reach for.
There is also the ordinary open-source math. This is a young project with a couple hundred stars, one primary maintainer, and an MIT license, which is another way of saying there is nobody to call. That is a fine trade at a price of zero. It is a different trade if you were planning to build a quarter of your company's go-to-market on top of it.
The interesting thing here is not that a founder can skip a consultant. It is that a consulting deliverable turned out to be shaped like software all along. Fourteen skills, a brief on disk, a review gate. Once the framework is a file the agent can read, the retainer has to justify the part that was never the framework, and most retainers have never had to say out loud what that part is.