Four playbooks, one pipeline. Each governs a different question, asked in the order building actually requires an answer to it: whether a thing should exist, how it's arranged, how it looks, and what it says. Every rule inside renders live — this is the door in front of all four.
Whether it should be built at all, and what would prove that wrong.
How the pages, terms, and sections get sorted before anything ships.
The tokens, type, motion, and governance the rest gets built on.
What actually gets written, once there's something worth writing.
Each one is a complete, working front door on its own — open any card below and you land on a full index with its own hero, its own sequence, its own rules rendered live rather than described. This page is only the room they all sit in.
Whether it should exist. Four stages — Map, Talk, Doubt, Price — that refuse to price, build, or believe anything before the evidence says so.
5 pages · 4 refusalsOpen → FoundationHow it's arranged. The layer underneath the other three: the Three Fogs, the Say and Never lists, the Nesting Test, and the rule that sorting is always an argument.
1 page · 9 named artifactsOpen → Design systemHow it looks. Eighteen living styleguide pages — tokens through accessibility — plus the demos and working files behind them, every rule rendered rather than written about.
29 pages · 18 styleguidesOpen → Copy systemWhat it says. Thirteen pages in the order you actually write in — Decide, Build, Check, Run — with two enforcement passes for AI tells and manufactured pressure.
14 pages · 4 stagesOpen →Sorting is never neutral — see the Architecture Playbook's own Sorting Is Arguing — so the sequence above is stated on purpose rather than left implicit.
The four questions have a dependency order, and skipping forward is the same failure mode Living Brainstorm names inside its own four stages: pricing before the problem is proven, building before the doubt list. Deciding whether something should exist before arranging it, arranging it before skinning it, and skinning it before writing on it isn't a preference — building a page's copy before its layout exists means rewriting the copy once the layout arrives anyway.