Behind the book

The book was built through the framework it describes.

Every chapter draft, research sweep, fact-check, and revision pass ran through a project-local Claude Code harness. Seven skills did the work. Read the SKILL.md files directly. Copy them. Adapt them. The patterns transfer cleanly to any rigorous nonfiction project; the specific banned-word lists and chapter-slug conventions are the author's, and you swap those for your own.

The same pattern that produced Builder-Leader is the pattern the book argues every senior leader should start running. The proof is that the book exists.

book-platform-sweep
Three-tier research orchestrator

Per-chapter sweep across Perplexity Sonar Pro, Hacker News (with comment bodies), Reddit (with post bodies), Grok x_search, Gemini grounded search, and a curated Substack pool. Tier 2 layers an open-ended discovery sweep with URL body extraction. Tier 3 is a hybrid two-pass synthesis: a Sonnet mechanical rollup, then an inline Opus judgment pass into a consolidated findings file. No chapter draft starts without a completed sweep.

book-draft-chapter
Voice-sensitive drafter

Reads canon (thesis, outline, audience, voice), locates the target chapter, pulls source material, drafts to disk, and runs a banned-phrase + sentence-variance scan against the global voice profile. Inline on main-context Opus only — voice work does not delegate cleanly.

book-red-team
Structural and adversarial critique

Runs a chapter draft through four adversarial lenses (contrarian, hostile reviewer, sympathetic-skeptical operator, craft editor), plus cross-chapter contradiction checks and promise-versus-delivery alignment. Sequential runs across the manuscript surface systemic patterns invisible at the per-chapter level. Operates as both an early-stage critic and a final rigor gate.

book-fact-check
Assertion registry + claims queue

Extracts every factual claim in a draft, matches each to the chapter findings file or the master source index, and classifies as SOURCED, PARTIAL, UNSOURCED, or DRIFT. Writes a per-chapter assertion registry and updates a rolling claims-to-verify queue.

book-adversarial-reader
Multi-persona reader simulation

Reads each chapter through five reader personas drawn from the audience profile: in-position VP, aspiring Director, Pascal-type senior executive, hostile augmented-leadership reviewer, and one near-Group-3 reader who is one promotion away from the seat. Each persona produces independent feedback. Convergence across personas is the signal to advance.

book-voice-polish
Mechanical and judgment voice gate

Final-stage voice pass. Mechanical scan for em-dashes, banned words, banned constructions, sentence-length monotony, paragraph-rhythm uniformity, and content-word over-frequency. Per-hit Opus judgment for context-justified versus violation. Strictly clause-and-sentence-level — never restructures.

book-export
Manuscript assembly and typesetting

Pandoc plus xelatex pipeline that assembles the chapters into a typeset PDF and a clean EPUB. Front matter, table of contents, page numbers, citation handling. Ship mode versus preview mode. Reproducible from the same source.

Get the starter

Slopless

The scaffolding side of the author's own Claude Code setup. Rules, hooks, statusline, a base CLAUDE.md. Not personal, not opinionated about your projects. Designed to bootstrap a working rig in under five minutes from an empty directory.