ultimate-best-seller-studio
AGENTS.md2 starsChanged 5 months agoMITView on GitHub
This repository contains Ultimate Best Seller Studio (engine codename: Book Genesis), an agent-agnostic book-production workflow made of markdown skills, manifests, prompts, scoring rules, and reference files — with a measured human-band gate powered by the Better Humanizer. When asked to create, plan, draft, audit, score, revise, or package a book, use the Universal Book Genesis Core unless the user explicitly asks for legacy V4/V5: The folder name is historical. Treat this as the current universal pipeline for Claude Code, Codex,…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,542 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository27 other files it gives its agents.
- beta-readerSkill
- book-bestseller-studioSkill
- book-editorSkill
- book-genesis-coreSkill
- book-genesis-fullSkill
- book-genesisSkill
- book-researcherSkill
- book-swarm-panelSkill
- copy-editingSkill
- chaos-engineSkill
- dialogue-polishSkill
- hook-craftSkill
- mechanical-preprocessSkill
- quality-gateSkill
- editorial-packageSkill
- humanizer-proSkill
- literary-agent-panelSkill
- manuscript-managerSkill
- narrative-foundationSkill
- book-autoSkill
- continuity-guardianSkill
- entity-trackerSkill
- reader-personaSkill
- voice-fingerprintSkill
- production-prepSkill
- prose-craftSkill
- series-architectSkill
The file
We split this AGENTS.md into the rules it gives an agent
10 rules. About 0 are common, 8 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 7
When asked to create, plan, draft, audit, score, revise, or package a book, use the Universal Book Genesis Core unless the user explicitly …
- Distinctive: line 15
When the user asks for bestseller-level, market-ready, agent/editor-ready, or launch-ready work, layer the market umbrella skill on top of …
- Distinctive: line 21
When the user asks for MiroFish-style reader swarms, public-opinion simulation, niche-risk scouting, or many simulated readers, use:
- Distinctive: line 27
When the prose reads synthetic, fails the anti-AI scan, or needs to sit inside the real human distribution (de-AI / humanization), use the …
- Very specific: line 34
Use the literary register for fiction.
- Very specific: line 45
Keep PROJECT_STATE.yaml synchronized with reality.
- Distinctive: line 48
Do not skip Phase 4: Adversarial Audit.
- Distinctive: line 49
Do not run final scoring before the adversarial audit.
- Distinctive: line 50
Write Portuguese artifacts and prose in Portuguese when the book is in Portuguese.
- Distinctive: line 51
Treat legacy V4 material as historical reference unless the user asks for it.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Agent Instructions
This repository contains Ultimate Best Seller Studio (engine codename: Book Genesis), an agent-agnostic book-production workflow made of markdown skills, manifests, prompts, scoring rules, and reference files — with a measured human-band gate powered by the Better Humanizer.
Default Behavior
When asked to create, plan, draft, audit, score, revise, or package a book, use the Universal Book Genesis Core unless the user explicitly asks for legacy V4/V5:
skills/book-genesis-core/SKILL.mdThe folder name is historical. Treat this as the current universal pipeline for Claude Code, Codex, Antigravity, Kimi, and other file-aware agents.
When the user asks for bestseller-level, market-ready, agent/editor-ready, or launch-ready work, layer the market umbrella skill on top of the universal core:
skills/book-bestseller-studio/SKILL.mdWhen the user asks for MiroFish-style reader swarms, public-opinion simulation, niche-risk scouting, or many simulated readers, use:
skills/book-swarm-panel/SKILL.mdWhen the prose reads synthetic, fails the anti-AI scan, or needs to sit inside the real human distribution (de-AI / humanization), use the Better Humanizer instead of the old rule-based humanizer:
skills/humanizer-pro/SKILL.md # bridge: points at external/better-humanizer/
external/better-humanizer/ # the actual tool (git submodule, single source of truth)Use the literary register for fiction. To gate a whole manuscript on measured human-likeness, run ubss humanize-score <project> (writes evaluations/human-band/REPORT.md). See docs/humanizer-pro-integration.md.
Load the manifest before advancing phases:
skills/book-genesis-core/references/pipeline/manifest.yamlRules
- Persist important decisions to files.
- Keep
PROJECT_STATE.yamlsynchronized with reality. - Keep
ASSUMPTIONS.mdexplicit. - Load only the prompt for the active phase.
- Do not skip Phase 4: Adversarial Audit.
- Do not run final scoring before the adversarial audit.
- Write Portuguese artifacts and prose in Portuguese when the book is in Portuguese.
- Treat legacy V4 material as historical reference unless the user asks for it.
Agent-Specific Notes
- Claude Code can run
/book-genesis-coreafter installing the full skill folder. - Codex can use this repo directly through
AGENTS.mdand the skill folder. - Antigravity can use this file as the repo-level playbook.
- Kimi can use the full skill folder or the active phase prompt plus project state files.
- The optional local runner in
runner/cli.pycan scaffold projects and prepare phase packets, but it does not call a model or write literary output.
Legacy Commands
/book-genesis: V5 Craft Mode legacy orchestrator./book-genesis-full: full legacy production pipeline./book-genesis-core: current portable command name kept for compatibility.
Public Documentation
README.md: public overview and commercial positioning.docs/book-genesis-core.md: universal core architecture.docs/runner.md: local runner and mechanical demo.docs/portability.md: Claude Code, Codex, Antigravity, Kimi, and generic agent usage notes.docs/humanizer-pro-integration.md: the Better Humanizer integration and human-band gate.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.