<- all tokdocs
2026.09.18@itstundealao

The Claude Pricing "Skill" in This Video Is a 73-Line Markdown Prompt in emotixco/claude-skills-founder

Business
Watch on TikTok

View on TikTok ->

The file doing the work in this video is 73 lines of plain markdown that tells Claude to run six pricing exercises in order, and the output holds up because Claude read the creator's actual repo, not because the prompt is clever. Tunde Alao downloads a pricing skill from GitHub, installs it into the project for his iOS storage cleaner app, and runs it. The screen shows the repo as claude-skills-founder under the org emotixco, the file as commands/pricing-strategy.md at 73 lines, and the commit as six months old. All of that checks out. What the video does not show is how much of the good output came from the prompt versus from Claude having the codebase in front of it.

What Is Actually in the File

The repo is emotixco/claude-skills-founder, MIT licensed, 215 stars and 17 forks as of 2026-09-19. Every commit landed on 2026-03-10 inside a sixteen-minute window, and nothing has been pushed since. The README claims 12 skills, but commands/ holds 13 markdown files.

pricing-strategy.md opens with two lines of frontmatter, a description and an argument-hint, then a persona line: "You are a SaaS pricing consultant who has helped 100+ startups set and optimize their pricing." That number is a character detail in a prompt, not a credential anyone can check. The rest is six numbered sections Claude has to fill in: a scored table of six pricing models, a three-tier design, competitor anchoring, a unit economics calculation, three psychology tactics, and launch versus scale pricing. It closes with five rules, including "Use real numbers" and "Keep total output under 1500 words."

There is no code in the file. No scripts, no API calls, no pricing data. The leverage comes from forcing a specific sequence and demanding concrete figures instead of adjectives.

The Output Is Specific Because Claude Read the Repo

The generated document in the video is titled "EasyGB pricing strategy" and opens with an honest preamble: the creator passed no arguments, so Claude worked from the repo and listed what it assumed. It identified the product from node/lib/app.json at version 1.3.1, found RevenueCat wired up with weekly, yearly and lifetime entitlements, spotted the paywall copy "Pay once, use forever," and noted that the Meta SDK plus the tracking permission prompt implied paid acquisition. It then flagged the gap directly: the actual prices live in the RevenueCat dashboard, not in the code, so anywhere it needed a number it did not have, it marked an assumption.

That preamble is the most useful part of the run, and the video skips past it. A pricing prompt pointed at an empty directory would have produced the same six headings filled with generic SaaS advice.

The Unit Economics Arithmetic Checks Out

The output's cost inputs are all verifiable and all correct. Apple takes 15% under the App Store Small Business Program, which applies to developers earning under $1 million in annual proceeds. RevenueCat charges nothing below $2,500 in monthly tracked revenue and 1% above it. Supabase Pro is $25 per month. The Apple Developer Program is $99 per year, which the document converts to roughly $8 per month. Those add up to the stated fixed cost of about $55 per month.

Running the numbers on the proposed tiers, the margins are right. Weekly at $4.99 nets $4.19 after a 15% Apple cut and a 1% RevenueCat fee, an 84% gross margin, and $55 divided by $4.19 is 13.1 charges to break even, which matches the "13 charges" in the table. Yearly at $29.99 nets $25.19 and lifetime at $49.99 nets $41.99. The "88% off" framing on the annual plan compares $29.99 against 52 weeks at $4.99, which is $259.48, so the discount is 88.4%. The $0.58 per week label on the yearly tier is $29.99 divided by 52.

The anchoring move is the sharpest piece. Rather than benchmarking against other cleaner apps, the output anchors on iCloud+: 200GB costs $2.99 per month in the US, or $35.88 a year, against $29.99 for a permanent cleanup. Every iPhone owner who has hit the storage wall has seen that upsell, so it functions as a familiar reference price.

Skill Is the Wrong Word, and One Recommendation Has a Gate on It

The video says skill, and the on-screen caption says skill. The repo ships slash commands. The README's install steps copy commands/ into .claude/commands/founder/, which produces /founder:pricing-strategy. Claude Code's newer Agent Skills format lives at .claude/skills/<name>/SKILL.md, carries frontmatter that lets Claude invoke it on its own, and can bundle supporting files. Both create a slash command, so nothing breaks, but this one only fires when you type it.

One recommendation in the output also has a condition the video does not surface. The document suggests skipping a launch discount and running a RevenueCat Experiment instead. RevenueCat's own documentation lists Experiments as available to Pro and Enterprise customers, so an app sitting under the $2,500 monthly threshold may need to move to a paid plan before it can run one.

What the Verdict Rests On

Alao says he already ran different prices and knows which converts best, which is why he can judge the suggestions. That is the whole evidentiary basis for the "fact" call, and he never shows the comparison. He also cuts the video to go implement the suggestions rather than reporting a result. The claim is that the recommendations look right to someone with prior data, which is a reasonable thing to say and a weaker thing than a proven lift.

The app itself is not independently checkable. A search of the US App Store turns up many similarly named storage cleaners and no listing for EasyGB, so the product, its 1.3.1 version, and its current paywall exist only as what appears on screen.

Key Takeaways

  • The skill is 73 lines of markdown with no code. It works by forcing six pricing exercises in sequence and banning vague language, which is a structure you could write yourself in twenty minutes.
  • Output quality came mostly from Claude reading the actual repo, including the RevenueCat entitlements and the paywall copy. The same prompt in an empty folder produces generic advice.
  • Every cost input in the generated analysis is verifiable and correct: Apple 15% under the Small Business Program, RevenueCat 1% above $2,500 MTR, Supabase Pro $25 per month, Apple Developer $99 per year.
  • The margin math, break-even counts, and the 88% annual discount framing are all internally consistent when recomputed.
  • The repo has 215 stars, an MIT license, and zero commits since 2026-03-10. Its README says 12 skills while commands/ contains 13 files.
  • The files are slash commands in .claude/commands/, not Agent Skills in .claude/skills/<name>/SKILL.md, so Claude will not load them on its own.
  • Unverified: the EasyGB app has no findable App Store listing, the "helped 100+ startups" persona line is prompt flavor rather than a credential, and the creator's claim that he knows which price converts best is never shown.

Resources

Published September 18, 2026. Writeup generated from a favorited TikTok.