# Book Bestseller Studio > Use when the user wants a complete book-production system aimed at commercial viability, bestseller positioning, agent/editor readiness, or launch readiness. Orchestrates book-genesis, book-researcher, literary-agent-panel, book-editor, copy-editing, humanizer, editorial-package, launch-strategy, and content-strategy. - Skill: `felipelobomotta-blip/book-bestseller-studio` (Agent Skill, multi-file: 4 files) - Install (CLI): `npx skillmds@latest add felipelobomotta-blip/book-bestseller-studio` - Raw SKILL.md: https://api.skillmd.com/api/skills/felipelobomotta-blip/book-bestseller-studio/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Marketing & Growth - Author: felipelobomotta-blip (https://skillmd.com/u/felipelobomotta-blip) - Updated: 2026-09-22 - Page: https://skillmd.com/skills/felipelobomotta-blip/book-bestseller-studio --- # Book Bestseller Studio This is the umbrella skill for creating a complete book at market level. It does not replace canonical `book-genesis`; it coordinates specialist skills and adds commercial gates. Load the installed `book-genesis` skill and its `references/pipeline/host-contract.md` first. Resolve this skill's `references/` locally and `skills//` through sibling installed skills. The repository Python helper is optional; use native file tools when it is absent. Optional external skills and services are enhancements, never required to begin writing. Use this when the user says things like: - "create a complete book" - "bestseller-level" - "commercially viable book" - "prepare for agents/editors" - "complete book pipeline" - "market-ready" - "publication or launch prep" ## Specialist Stack - `book-genesis`: canonical project pipeline, state files, chapters, audits, Genesis Score. - `book-researcher`: genre map, comp titles, reader gaps, word count targets, nonfiction evidence. - `book-swarm-panel`: MiroFish-style simulated reader swarms, niche-risk scouting, public-opinion tests, interviews, heatmaps, and revision tickets. - `literary-agent-panel`: simulated agent/editor/bookseller review for market acceptance. - `book-editor`: surgical revision by mode after evaluation. - `copy-editing`: final language cleanup. - `humanizer`: remove synthetic phrasing and restore lived voice. - `editorial-package`: logline, back-cover copy, synopsis, query letter, cover brief. - `launch-strategy`: go-to-market, offer, launch calendar, proof-driven positioning. - `content-strategy`: audience-building and content plan around the book. `launch-strategy` and `content-strategy` are optional external companions. Do not block manuscript completion when they are not installed. ## Operating Rule Create durable files. Do not keep the book only in chat. Default project layout: ```text / PROJECT_STATE.yaml ASSUMPTIONS.md research/ foundation/ manuscript/ chapters/ evaluations/ delivery/ launch/ ``` ## Market-Level Pipeline 1. Run `book-researcher` before architecture unless user already supplied market research. 2. Run `book-genesis` phases 0-3 to build intake, foundation, architecture, and manuscript. 3. After first full draft or major act, run `literary-agent-panel`. 4. Run `book-swarm-panel` when the project needs many fictional readers, niche/sensitivity-risk scouting, public-opinion reaction, or package-framing tests. 5. Run `book-genesis` adversarial audit and Genesis Score. 6. Use `book-editor` only against specific findings. Pick mode: structural, connective, prose-texture, rag-rewrite, or factual. 7. Use `humanizer` and `copy-editing` after structure is stable. 8. Run `literary-agent-panel` and/or `book-swarm-panel` again after revision. 9. Run `editorial-package` only after manuscript passes product gates. 10. Run `launch-strategy` and `content-strategy` for public-launch readiness. ## Portable Agent Dispatch Read `references/dispatch.md` and use `references/agent-registry.yaml` for specialist ownership. Prepare packets with `python runner/cli.py prepare-agent-packet ` when the runner is available. - Claude Code: dispatch a custom or general-purpose subagent with the packet. - Codex: dispatch an isolated subagent with the packet when available. - Kimi Code: dispatch a fresh built-in subagent with the packet. - No subagents: run roles sequentially, mark evaluation independence as degraded, and never call self-scoring independent. Editorial rules remain identical across adapters. Platform-specific agent files may select tools or models, but must not redefine gates or scoring. ## Commercial Gates Do not call a project "market-ready", "best-seller-ready", or "publication-ready" unless all gates pass or blockers are documented. Gate 1: Product Fit - clear genre/category - current comp titles - target reader defined - reader promise stated in one sentence - format and word count match category expectations Gate 2: Manuscript Fit - complete manuscript or explicit serial/novella positioning - opening creates immediate reader contract - chapter endings sustain forward motion - weakest Genesis Score dimension is not hiding behind average score - continuity and factual checks complete Gate 3: Market Package - logline - 100-word cover copy - 300-word editorial synopsis - query/pitch letter - cover brief with genre shelf fit - author positioning Gate 4: Launch Fit - proof points or platform plan - audience acquisition path - content angles - launch calendar - review/ARC strategy - retailer/category metadata assumptions ## Bestseller Discipline "Bestseller" is a positioning target, not a promise. Treat it as a standard for product, package, and launch quality: - write for a specific reader, not "everyone" - use comp titles from the current market, not classics - make the first 10 pages commercially ruthless - preserve voice while improving readability - document risks instead of pretending certainty - separate craft excellence from market readiness ## Output Contract For any full run, maintain: - `research/market-research.md` - `foundation/positioning.md` - `manuscript/chapters/chapter-XX.md` - `evaluations/agent-panel.md` - `evaluations/book-swarm//SUMMARY.md` - `evaluations/genesis-score.md` - `delivery/editorial-package.md` - `launch/launch-plan.md` If user asks only for one part, run only the relevant specialist and do not force full pipeline.