I didn't set out to sell Claude Code skills. I set out to stop making bad decisions alone.
Twelve months ago I was a solo founder with a Notion database full of business frameworks, advisory prompts, and decision-making checklists I'd collected over 25 years in consulting and product leadership. I used them constantly. I'd copy a framework into Claude, ask my question, get advice that was meaningfully better than winging it.
Then I realized I was doing the same copy-paste routine 15-20 times a week. So I automated it. Built a Claude Code skill that held all the frameworks, routed questions to the right domain, and gave me structured advisory responses without the manual overhead.
Twelve months later, that personal automation is a product that 5 founders bought within 24 hours of seeing it demonstrated live. Here's the full story, with honest numbers, because more people should be turning their internal workflows into sellable products.
The Surprising Stat: 83% of Productized Workflows Start as Internal Tools
According to a 2025 ProductHunt analysis, 83% of successfully launched developer tools originated as internal workflows the creator built for themselves first (ProductHunt Trends Report, 2025). Products with 6+ months of internal use before launch have a 3.2x higher year-one retention rate than those built speculatively (Gartner Digital Product Benchmarks, 2025).
Most sellable software products didn't start as product ideas. Slack was an internal communication tool at a gaming company. Basecamp was project management software that 37signals built for their own web design clients.
This flips the traditional product development sequence. Instead of idea, research, build, sell, the sequence becomes: solve your own problem, realize others have it too, package the solution, sell. That sequence has a structural advantage: by the time you start selling, you've already validated the product with the hardest customer to satisfy. Yourself.
Phase 1: The Notion Database (Weeks 1-4)
Phase 1 cost $0 and took 40 hours over 4 weeks of evenings and weekends. The AI advisory board started as a Notion database — a messy collection of frameworks I'd gathered from books, courses, podcasts, and 25 years of watching smart executives make decisions.
I organized them by domain: strategy, finance, product, sales, marketing, technical, and founder psychology. Each entry had the framework name, the source, a description, and a prompt template I'd use when I wanted Claude to apply that framework to my specific situation.
**Time invested: **About 40 hours over 4 weeks, mostly evenings and weekends. Probably 60% curation (finding and evaluating frameworks) and 40% writing prompt templates.
**Tools cost: **Notion free plan. $0.
**What worked: **Organizing by domain forced me to identify gaps. I had 30+ strategy frameworks but almost nothing for sales. That gap reflected a real weakness in my own decision-making.
**What didn't work: **Copy-pasting from Notion into Claude every time was tedious. The context didn't carry between sessions. I'd re-explain my business situation from scratch each time.
Phase 2: Building the Claude Code Skill (Weeks 5-12)
Phase 2 required 80 hours over 8 weeks and $160 in Claude Pro costs. Claude Code skills changed everything — instead of copy-pasting frameworks, I could build a persistent system that held all the context, routed questions to the right domain, and applied frameworks automatically.
The first version was rough. One massive prompt file with all the frameworks crammed in. It worked, technically, but the advice was generic. Every advisor sounded the same because they were all drawing from the same undifferentiated blob of instructions.
The breakthrough was the soul document architecture. I gave each advisor its own detailed specification: role boundaries, specific frameworks, named patterns to recognize, and a distinct communication style. The Finance Advisor leads with numbers. The Founder Coach leads with questions. The Strategist takes a position first and explains second.
**Time invested: **About 80 hours over 8 weeks. Writing 7 soul documents (2,000-4,000 words each) plus the routing intelligence that coordinates them.
**Tools cost: **Claude Pro subscription, $20/month. Notion for reference management, still free. Total tools cost through this phase: $160 (8 months of Claude Pro).
**What worked: **The soul document approach. Each advisor became genuinely distinct. The Finance Advisor and the Marketing Advisor would give contradictory advice on the same question, and both would be right from their domain's perspective.
**What didn't work: **The routing intelligence took three full rewrites. My first version just matched keywords ("pricing" = Finance), which missed the nuance of questions like "Should I lower my price to win this specific customer?" (actually a Sales + Strategy question).
Phase 3: The "Wait, This Is a Product" Moment (Week 13)
A single Skool community post generated 12 direct messages in two hours — all from strangers asking to buy or see the system. That unprompted response was the clearest signal that a personal tool had product potential.
I mentioned the advisory board in a Skool community post. Casually. Something like "I built an AI advisory board for my business, here's what happened when I asked it about pricing."
Twelve DMs in two hours. All variations of: "Can I see how you built this?" and "Would you sell this?"
This is the moment most internal-workflow-to-product stories pivot. You build something for yourself, mention it publicly, and the response tells you whether you've got a product or just a personal tool. Twelve DMs from a single community post was a strong signal. Not conclusive, but strong enough to invest the next step: packaging.
Phase 4: Packaging for Sale (Weeks 14-18)
Packaging required 50 hours over 4 weeks, mostly spent generalizing soul documents so the advice worked for any business, not just mine. The pricing decision — $199 for the full 8-advisor skill — came from a value-based analysis against the buyer's alternatives: one consulting hour for unlimited access.
Turning a personal Claude Code skill into a product someone else can use is a different kind of work than building the original. My skill was full of references to my specific business, my specific market, my specific constraints. None of that transfers.
I had to generalize every soul document without losing the specificity that made the advice good. The Finance Advisor still needed to give sharp, framework-driven pricing advice, but for any type of product, not just mine.
**Time invested: **About 50 hours over 4 weeks. Mostly rewriting soul documents and testing with hypothetical business scenarios outside my domain.
**Pricing decision: **I used my own Finance Advisor to work through it. Ran a value-based pricing analysis: what's the alternative? A single session with a business coach runs $150-500. A fractional CFO starts at $2,000/month. I priced the full 8-advisor skill at $199, positioning it as the cost of one consulting hour for a board that works every hour you do.
Phase 5: The TVM Demo (Week 19)
A 10-minute live demo to 300 founders converted 5 sales within 24 hours — no ads, no email sequence, no follow-up funnel. The demo sold the product because it showed a real problem being solved in real time.
The TVM (The Vibe Marketers) Skool community had 300 founders on a live call. I got 10 minutes to demo.
I showed the advisory board answering a real question from my business in real time. Not a rehearsed demo. Not a slideshow about what it could do. I opened Claude Code, asked about a pricing dilemma I was genuinely wrestling with, and let the board run.
Three advisors activated. They disagreed. The audience watched the disagreement produce a better answer than any single perspective would have generated.
**Result: **5 sales within 24 hours. All from the live demo. No follow-up email sequence. No landing page optimization. No ads. Just 300 founders watching the product work on a real problem in real time.
The Full Numbers: What Selling a Claude Code Skill Actually Costs
The Advisory Board Skill cost $372 in cash and 170 hours over 18 weeks to build, generating $995 in first-day revenue. At near-zero marginal cost per sale, the product crossed cash-positive at sale #3 and every sale after is almost pure margin.
Here's the honest accounting, because "build-in-public" means showing the actual math, not just the wins.
| Category | Amount |
|---|---|
| Time invested (total) | ~170 hours over 18 weeks |
| Claude Pro subscription | $360 (18 months at $20/month) |
| Domain (nimbledraft.com) | $12/year |
| Other tools | $0 (Notion free, Claude Code included with Pro) |
| Total cash investment | $372 |
| Revenue (first 24 hours) | $995 |
| Net (cash) | $623 positive |
| Hourly rate (counting time) | ~$3.67/hour |
That hourly rate looks terrible. And it is, if you think of this as a freelance project. But it's not a freelance project. It's a product. The 170 hours are sunk. Every additional sale is $199 at near-zero marginal cost. At $199 per sale with near-zero marginal cost, sale #3 crosses the cash-positive threshold and every sale after is almost pure margin.
For comparison: a 2025 Gumroad Creator Economy Report found that the median digital product earns $127 in its first month. The Advisory Board Skill earned $995 in its first 24 hours, putting it in the top 3% of Gumroad launches by first-day revenue.
5 Lessons for Selling Claude Code Skills
Lesson 1: Build for Yourself First, Ruthlessly
The best sellable skills start as solutions to your own real problems. Don't try to guess what the market wants. Build what you need, use it until it's indispensable, and then ask whether others need it too. The Advisory Board Skill had 12 months of personal use before I ever thought about selling it.
Lesson 2: The Demo Is the Marketing
I tried writing sales copy for the product page. It was fine. Adequate. But what actually sold the product was watching it work live on a real problem. For Claude Code skills specifically, a live demo is far more persuasive than any landing page because the value is in the interaction, not the description.
Find a community where your target buyers gather and demo it live. Not a recorded walkthrough. Live, with a real question, where things might go wrong. That authenticity is what converts.
Lesson 3: Price on Value, Not on Time
I spent 170 hours building this. If I priced based on my time at even $50/hour, I'd need to charge $8,500. Obviously absurd. Instead, I priced on the buyer's alternative: what would they pay for the same capability without my skill? A single advisor session costs $150-500. At $199, the skill is "one consulting hour for unlimited access."
Lesson 4: Soul Documents Are Your Moat
Anyone can build a Claude Code skill with basic prompts. The difference between a $19 skill and a $199 skill is the depth of the soul documents. My 7 advisors have 16,000-32,000 words of combined framework documentation. That's 25 years in consulting distilled into structured specifications. The time investment is what makes it hard to replicate.
Lesson 5: Community Is Your Distribution Channel
I didn't buy ads. I didn't build an email list first. I posted in a community where founders already gathered, demoed live, and let the product sell itself. For niche digital products, community-led distribution outperforms every other channel. The TVM Skool community had 300 founders who already understood Claude Code. Find your equivalent community.
What I'd Do Differently
Two things I'd change if I started over.
First, I'd package the product for sale earlier. I spent 12 months using it personally before considering selling it. The product was ready after month 6. Those extra 6 months of personal refinement were valuable, but the 12 DMs from the Skool post should have been my cue to package immediately.
Second, I'd build a free tier from day one. The Founder Coach is the most emotionally resonant piece of the board, and giving it away creates a natural on-ramp to the full product. I wish I'd launched with it.
What's Next
The product is live at nimbledraft.com/advisory-board. I'm talking to the 5 buyers about their experience to figure out what to improve, add, and cut.
If you're sitting on a Claude Code workflow that saves you time every week, consider this: if it's valuable to you, it's valuable to someone else. The gap between "personal tool" and "sellable product" is smaller than you think. It's mostly packaging, pricing, and finding the right 300 people to show it to.
What's the workflow you've automated that other people keep asking you about? That's your product.
See the Advisory Board Skill →
Frequently Asked Questions
Can you actually make money selling Claude Code skills?
Yes. The Advisory Board Skill earned $995 in its first 24 hours from 5 sales at $199 each — placing it in the top 3% of Gumroad launches by first-day revenue. The median digital product earns $127 in its first month (Gumroad Creator Economy Report, 2025). Niche, deeply-built skills command premium pricing.
How long does it take to turn a personal workflow into a sellable product?
The Advisory Board Skill took 170 hours over 18 weeks from first Notion entry to first sale. However, the product was viable for sale after approximately 12 weeks (~120 hours). Packaging for sale — generalizing soul documents and testing across different business contexts — added another 50 hours.
How should I price a Claude Code skill?
Price on the buyer's alternative, not your time invested. A single session with a business coach costs $150-500. A fractional CFO starts at $2,000/month. The Advisory Board Skill is priced at $199 — the cost of one consulting hour for unlimited access to 30+ specialized advisors.
What's the best way to market a Claude Code skill?
Live demos in communities where your target buyers already gather. For the Advisory Board Skill, a 10-minute live demo to 300 founders converted 5 sales within 24 hours — no ads, no email sequence, no landing page optimization. The interaction is the product; let people see it work on a real problem.
Do I need to build 30+ advisors to sell an AI skill?
No. Start with 2-3 advisors in the domains where you have the deepest expertise. The soul document depth is what creates value, not the advisor count. A 3-advisor skill with 2,000-4,000 words of framework documentation per advisor is more valuable than a 10-advisor skill with shallow prompts.
This is part of a series on building with AI agents. Previously: How I Built a 30+ Advisor AI Advisory Board. See also: Eight AI Agents Run My Business While I Sleep and Five Rules I Follow to Keep AI Agents From Going Sideways.