founder-playbook turns fifteen business books into routed agent skills

The premise is that the frameworks are the part worth keeping

The stated motivation is that business books have great frameworks buried in 300 pages of stories, so you read the book, highlight 20 pages, and forget the frameworks six months later. The repository's answer is a set of skills that extract what matters, and the project describes itself as 15 proven business books distilled into structured AI skills that any LLM can use. Each skill captures the frameworks, decision trees, case studies and templates from a single book, without the 300 pages of anecdotes. The five things every skill contains are worth listing, because they are the difference between a summary and a usable reference. There are decision trees that ask which framework applies to your situation, the actual models with scored checklists and step by step processes, case studies of real companies that applied or failed to apply the ideas, fill in the blank templates, and integration maps showing where frameworks from different books conflict or complement each other. The last piece is the one that is hard to produce and the one that makes the collection more than a bookshelf.

A meta-skill routes the vague questions

One entry in the table is not a book at all. The diagnose skill is described as a meta-skill that routes across all 15, and its use when is the case where you do not know where to start, you have multiple problems, or nothing is working. The examples in the usage section show the routing in action, with a question about what is wrong with your startup loading Diagnose, which in turn routes you to the right skill, an interview question loading Mom Test, a pivot question loading Lean Startup, a marketing channel question loading Traction, a pricing question loading Monetizing Innovation, a website conversion problem loading StoryBrand, and a pitch that is not landing loading Made to Stick. You can also invoke any skill directly, with slash commands such as /lean-startup, /traction and /mom-test. The meta-skill is what makes the collection usable by someone who does not yet know which book answers their problem, which is arguably the person who would benefit most from it.

Fifteen books, chosen for founders rather than for completeness

The list is a curriculum, and the selection is the argument. The Mom Test by Rob Fitzpatrick is for customer interviews and validating ideas without leading questions. The Four Steps to the Epiphany by Steve Blank is for finding first customers, Customer Development and Market Type. The Lean Startup by Eric Ries covers Build-Measure-Learn, MVPs, pivots and innovation accounting. Obviously Awesome by April Dunford is about positioning, category choice and competitive context. Crossing the Chasm by Geoffrey Moore is about tech adoption, beachhead strategy and mainstream scaling, while Blue Ocean Strategy by Kim and Mauborgne is about category creation and escaping competition. Monetizing Innovation by Ramanujam and Tacke is about pricing strategy, willingness to pay and designing product and price together, and SPIN Selling by Neil Rackham is for B2B sales and multi-stakeholder deals. The three Hormozi books cover offer design and packaging, lead generation and advertising, and offer sequencing with upsells, downsells and continuity. Influence by Cialdini is about persuasion, negotiation and defence against manipulation. Traction by Weinberg and Mares covers growth channels and the Bullseye Framework, Building a StoryBrand by Donald Miller covers brand messaging and website copy with the SB7 Framework, and Made to Stick by the Heaths is about memorable messages with the SUCCESs Framework. The selection is skewed towards early stage founders, which is where the frameworks are most concentrated and most useful.

Progressive disclosure is what keeps the context cost down

Each skill follows a progressive disclosure format, which is the mechanism that makes fifteen books loadable at all. The entry point, SKILL.md, is around 250 to 500 lines and holds the decision trees, key frameworks and a quick reference. The supporting files are loaded only when needed, with frameworks.md for detailed framework breakdowns, cases.md for the real company case studies from the book, examples.md for templates, worksheets and worked examples, and integration.md for how the skill relates to and conflicts with the others. The readme quantifies the cost, roughly 100 tokens per skill at startup for just the description, around 500 lines when a skill is triggered, and the supporting files only on demand. That is the difference between a model that has read the book and a model that has read a table of contents, and it is why the routing descriptions are written to be about 100 tokens each.

Installing into Claude Code, or into anything else

The quick install is a single command.

npx skills add getagentseal/founder-playbook

The manual route clones the repository and symlinks each skill directory into the skills folder, one link per skill.

git clone https://github.com/getagentseal/founder-playbook.git

for skill in founder-playbook/skills/*/SKILL.md; do
  dir=$(dirname "$skill")
  name=$(basename "$dir")
  ln -sfn "$(pwd)/$dir" ~/.claude/skills/$name
done

The project is built for Claude Code, where the skills auto-trigger based on your question, and it also works with ChatGPT, Gemini, Cursor or any LLM as reference documents. For a plain chat model you open the relevant SKILL.md file and paste it into the conversation as context, and the readme notes that the YAML frontmatter is just metadata that every LLM ignores in favour of the content. For deeper detail you also paste the supporting file you need, such as frameworks.md, cases.md or examples.md. For Cursor, Windsurf or Cline you drop the repository into your project's rules or docs folder, and the AI will reference the skills when relevant.

What a distillation costs you

The honest limitation is that a summary of a framework is a compression somebody else made, and the value depends entirely on whether you agree with the compression. A book like The Mom Test has a specific and somewhat counterintuitive argument about what makes customer interviews useless, and a skill file that captures the framework without the argument may leave you with the technique and none of the reasoning. The repository is aware of this, and every skill includes an honest scope section covering what the book got right, what is dated, and where it does not apply, which is a good sign and not a complete answer. The second limit is that the frameworks are generic. A skill about positioning or pricing cannot know your market, your customers or your constraints, so the routing will hand you the right framework and you still have to supply the facts. The third is that the collection reflects a point in time, and books like Crossing the Chasm and The 4 Steps to the Epiphany are not new, which is why the scope sections mention what is dated. The last practical point is that this is a set of instructions to a model, so how well it works depends on the model following them, and on how much of the framework survives that translation.

Editorial conclusion

founder-playbook suits a founder who has read the books and lost the frameworks, and it suits a team that wants a shared vocabulary for decisions rather than a private pile of highlights. It is the wrong tool if you need the books themselves, since the premise is that the 300 pages of anecdotes are what you skip, and it is the wrong tool if you want the advice applied to a specific company, because the skills are generic frameworks with no knowledge of your market. Before installing, read one skill file end to end and check that the distilled version matches your own reading of the book, since a summary of a framework is a compression you have to trust. The last push was on 2026-10-06 and the repository publishes no releases.

Frequently asked questions

Official sources

1. getagentseal/founder-playbook on GitHub
2. Issues
3. License: MIT
4. README