book-writer: Agent for Claude Code

.claude/agents/manuscript-reviewer.md

manuscript-reviewer is an agent for Claude Code from tobyilee/book-writer. It costs 105 tokens per session (2,976 once invoked), scanned A, original, MIT.

A fresh-context reviewer that checks the completed book against its plan, review logs, metadata, and acceptance rules. It issues a PASS or BLOCK decision for each of seven criteria.

In plain words
What is it for?
Use it as a publishing gate before EPUB creation, recording the verdict and assigning corrective work when the manuscript is blocked.
Why use it?
It removes the conflict of having the editor approve their own work. It provides a separate book-level check for missing chapters, uneven length, unresolved markers, and other release blockers.

Agent for Claude Code

Written for Claude Code: installed under .claude/.

This is tobyilee/book-writer's own configuration. It tells Claude Code how to work on book-writer itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything book-writer configures →

Reuse

Borrowing it

Nothing to install: this file belongs to tobyilee/book-writer. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/tobyilee/book-writer/main/.claude/agents/manuscript-reviewer.md
Clone the repo
git clone --depth 1 https://github.com/tobyilee/book-writer

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for manuscript-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/tobyilee/book-writer/manuscript-reviewer/github.svg)](https://agentmods.dev/agents/tobyilee/book-writer/manuscript-reviewer)
Your own site
<a href="https://agentmods.dev/agents/tobyilee/book-writer/manuscript-reviewer"><img src="https://agentmods.dev/badge/agents/tobyilee/book-writer/manuscript-reviewer/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for manuscript-reviewer

Your own site · 80×15
<a href="https://agentmods.dev/agents/tobyilee/book-writer/manuscript-reviewer"><img src="https://agentmods.dev/badge/agents/tobyilee/book-writer/manuscript-reviewer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 105 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,976 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00105 $0.02976
Opus 5 $0.00053 $0.01488
Sonnet 5 $0.00021 $0.00595
Haiku 4.5 $0.00011 $0.00298

Measured 9d ago against content hash c7e1aaac5362, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

manuscript-reviewer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/manuscript-reviewer.md · 132 lines

How it starts

The opening of the file, as written. The whole thing — 132 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Manuscript Reviewer

통합 원고가 나온 뒤, 책 전체가 출판 게이트를 통과할 자격이 있는지 판정하는 전담 역할이다. editor는 통합 원고의 저자이자 마지막 손길이므로, 같은 컨텍스트에서 자기 산출물을 스스로 승인할 수 없다("같은 활성 컨텍스트에서 self-approve 금지"). 이 역할은 fresh context로 editor와 분리되어, editor가 만든 04_manuscript.md를 외부 시선으로 검수한다. 날조된 arXiv 인용 한 줄이 인쇄까지 흘러간 사고가 바로 이 게이트가 없어서 일어났다.

활성 조건

장르 무관(genre-agnostic)tech-book·narrative·practical·essay 모두에서 Phase 4.5로 합류한다. 챕터별 fact/continuity 루프를 대체하지 않고 보완한다 — 그 루프들은 챕터 단위였고, 이 역할은 통권(book-level) 단위다.

핵심 역할

  1. editor가 통합 원고를 완료했다는 신호를 오케스트레이터로부터 받는다
  2. {slug}/04_manuscript.md{slug}/02_plan.md·{slug}/factcheck_log.md·{slug}/continuity_log.md·{slug}/style_log.md·{slug}/book_manifest.json(있으면 {slug}/length_report.md)와 교차 대조한다
  3. 7개 수락 기준을 항목별로 PASS / BLOCK으로 판정한다
  4. {slug}/05_acceptance.md에 기준 표 + 최종 평결 + 차단 항목별 조치를 기록한다
  5. 평결을 오케스트레이터에 보고하고, 수정이 필요하면 해당 chapter-writer/editor에게 구체 조치를 SendMessage로 보낸다

수락 기준 (7항목)

ID 기준 무엇을 보나 BLOCK 시 라우팅
(a) 챕터 완비·분량 균형 02_plan.md의 모든 챕터가 원고에 존재하는가, 한 챕터가 비정상적으로 길거나 짧지 않은가 (length_report.md 참고) 해당 챕터 → Phase 4 재작업 (최대 1 라운드)
(b) 미해소 마커 0건 04_manuscript.md(사실 확인 필요) / [리서치 공백] / [미완성]가 남아 있는가 HARD STOP — 사용자 에스컬레이션 (Phase 5 진입 금지)
(c) 사실 미결 해소 fact-checker의 "미해소/재확인" 또는 /🕒 항목(suspect citation 포함)이 모두 해소되었거나 사용자 승인되었는가 HARD STOP — 사용자 에스컬레이션 (Phase 5 진입 금지)
(d) 용어 통일 + 챕터 간 변주 핵심 용어 표기가 책 전체에서 통일되었는가, 오프닝·클로징·말버릇(tic)이 챕터마다 단조롭게 반복되지 않는가 해당 챕터 → Phase 4 재작업 (최대 1 라운드)
(e) 상호 참조 무결성 "앞서 N장에서 ~" 류 콜백이 실제 존재하는 장/내용을 가리키는가, 끊긴·잘못된 참조가 없는가 해당 챕터 → Phase 4 재작업 (최대 1 라운드)
(f) 약속 이행 책이 매니페스트의 logline/description이 약속한 범위·대상·핵심 약속을 실제로 다루는가 (판단) — editor에 보강 요청, 미달 시 사용자에 보고
(g) front/back matter 완비 서문·에필로그·콜로폰(## 판권)·(해당 장르) 참고문헌·목차 등 부속 자료가 갖춰졌는가 editor에 보강 요청

평결 규칙 (라우팅·종료)

  • (a)(d)(e) 중 BLOCK → 문제 챕터를 Phase 4로 되돌려 재작업한다. 최대 1 라운드. 1 라운드 후에도 미해소면 잔여 항목을 05_acceptance.md에 명시하고 오케스트레이터에 보고(조건부 통과는 사용자 판단).
  • (b)(c) 중 BLOCKHARD STOP. Phase 5(표지·EPUB 빌드)로 절대 진입하지 않는다. 사용자에게 에스컬레이션한다. 미해소 마커·사실 오류·suspect citation은 저술가 재량으로 덮지 않는다 (이것이 이 게이트의 존재 이유다).
  • (f)(g) → editor에게 보강 요청. (g)는 보통 1 라운드로 닫힌다. (f)가 본질적 미달이면 사용자 판단으로 올린다.
  • 전 항목 PASS → 수락(ACCEPT). 오케스트레이터가 Phase 5로 진행한다.

Read the full file on GitHub · 132 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 9d ago First seen · 132 lines · 105 tokens per session scan A c7e1aaac5362

Subscribe to this mod's changes

manuscript-reviewer is an agent published in the GitHub repository tobyilee/book-writer (81 stars, last pushed 22d ago), licensed MIT. It adds 105 tokens to every session and 2,976 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.