Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
anian0 avatar

Novel Review

  • 9 installs
  • Updated August 2, 2026
  • anian0/pick-skills

novel-review is a Claude skill that reviews novel chapters and settings across four modes using reflective questions rather than scoring or rewriting.

About

This skill reviews novel chapters and story settings across four modes: text review with eight-dimension questions, pending-setting promotion review, canon consistency check, and a per-five-chapter arc retrospective. It works by raising professional questions to prompt author thinking rather than scoring or rewriting. A novelist uses it after drafting a chapter or at review milestones, with actual edits handed off to novel-lite.

  • Reviews novel chapters and settings in four modes: text review, pending-setting review, canon health check, and every-5-
  • Uses professional questions to prompt author reflection rather than scoring or rewriting the draft
  • Coordinates handoffs to novel-lite (revise), novel-setup (macro), and novel-style-extract (style) without editing files

Novel Review by the numbers

  • 9 all-time installs (skills.sh)
  • Ranked #829 of 1,352 Code Review & Quality skills by installs in the Skillselion catalog
  • Data as of Aug 3, 2026 (Skillselion catalog sync)
At a glance

novel-review capabilities & compatibility

free; runs locally, no API keys

Capabilities
manuscript review · consistency checking · quality review · arc retrospective
Use cases
code review
Runs
Runs locally
Pricing
Free
From the docs

What novel-review says it does

**审查原则**:不是打分,而是**用专业问题激发思考**,让作者在回答问题中发现可提升点。
SKILL.md
审查不改稿 | 只产出问题与建议,实际修改走novel-lite
SKILL.md
npx skills add https://github.com/anian0/pick-skills --skill novel-review

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs9
Last updatedAugust 2, 2026
Repositoryanian0/pick-skills

What it does

Review novel chapters and settings in four modes using questions that prompt reflection, without rewriting the draft.

Who is it for?

Prompting an author to find weak points in a chapter or settings through structured review questions before finalizing.

Skip if: Rewriting the draft, scoring it numerically, or editing settings; those are handed to novel-lite/novel-setup.

When should I use this skill?

The user asks to review a chapter, check settings, evaluate pending settings, run a canon check, or do an arc retrospective after chapter 5/10/15.

What you get

A review report of multi-dimension questions and differences that drives revision through novel-lite.

  • A review report (eight-dimension questions, setting decisions, or arc retrospective)

By the numbers

  • 4 review modes
  • 10-20 questions across eight dimensions in text review
  • arc retrospective every 5 chapters

Files

SKILL.mdMarkdownGitHub ↗

小说章节审查

定位

四件套分工(详见novel-setup):

  • novel-review(本skill):中观层(审)。单章审查、阶段回顾、设定体检。
  • novel-setup:宏观层。项目启动、大纲、状态回写。
  • novel-lite:中观层(写)。单章创作,4阶段流程。
  • novel-style-extract:风格层。抽象化风格提取。

审查原则:不是打分,而是用专业问题激发思考,让作者在回答问题中发现可提升点。

---

四种审查模式

模式简称对象触发场景输出
正文审查review-text章节正文+三件套写完一稿、定稿前审查报告(八维度问题)
新设定审查review-pending待定设定草案待定项≥3条、定稿前设定决策报告
设定体检review-canon既有设定文件卷末/百章里程碑设定健康报告
阶段回顾review-arc最近5章正文+设定每5章定稿后阶段回顾报告

默认推断规则

  • "审查第X章" → review-text
  • "审查待定设定" → review-pending
  • "设定体检" → review-canon
  • "阶段回顾" → review-arc

---

参考文档

文档何时加载
references/reference-模式A-正文审查.mdreview-text
references/reference-模式B-新设定审查.mdreview-pending
references/reference-模式C-设定体检.mdreview-canon
references/reference-模式D-阶段回顾.mdreview-arc

---

通用工作流程

1. 加载上下文:按模式要求读取全部相关文件 2. 执行审查:按维度产出问题/差异/建议 3. 产出报告:写入固定路径,不改稿 4. 驱动循环:用户逐题回应 → 调用novel-lite修订 → 询问是否再次审查 5. 沉淀经验:循环结束时询问是否沉淀优秀表达/反面案例

---

与novel-lite的衔接

原则说明
审查不改稿只产出问题与建议,实际修改走novel-lite
必看文档是审查依据审查时读取02_必看文档.md,验证正文是否遵循
报告路径固定正文审查:工作区/第X章VX/审查历史/
定稿前必审移入正文前,至少跑一轮"定稿终审"
待定项触发novel-lite累积≥3条待审项时建议跑模式B
阶段回顾触发novel-lite在N%5==0时提示,由本skill跑模式D

---

与novel-setup的衔接

时机本skill动作novel-setup动作
审查涉及大纲偏离标注"由setup接手"评估并更新大纲
待定项属宏观范畴标注"由setup接手"执行晋升
模式C整改项标注分工各自落地
模式D设定更新标注分工各自落地
发现风格分片需调整建议调整落地调整

---

与novel-style-extract的衔接

场景本skill动作novel-style-extract动作
发现风格分片缺失/薄弱建议补充增量提取
发现量化指标不合理建议调整重新校准
发现自检分数虚高建议调整阈值

---

边界

  • 不打分(80/100这种),质量是多维的
  • 不替代用户判断,永远给方向不给结论
  • 不主动改写原文,除非用户明确要求
  • 不擅自修改任何设定文件;发现问题只在报告中提出
  • 遇到设定冲突,先指出、再询问以哪边为准
  • 模式D不擅自演化当前冲突.md,只产出建议
  • 模式B不替用户做晋升;模式C不替用户改设定
  • 审查风格一致性时,以必看文档中列出的规则为基准
  • 若发现自检报告与正文实际不一致,标注为P0级问题
  • 不直接生成风格知识库内容,由novel-style-extract执行

---

创作标准更新评估

每次讨论后执行:

评估项处理
发现优秀表达?建议写入 创作标准/好的示例.md
发现反面案例?建议写入 创作标准/差的示例.md
发现风格类问题?建议写入对应风格分片
发现新约束?建议写入 写作注意事项.md

本skill负责"发现与建议",不负责直接写入。写入规范见novel-lite的reference-创作标准更新规范.md。

Related skills

FAQ

Does it rewrite my chapter?

No. It only produces questions and suggestions; actual revisions are done through the novel-lite skill.

What are the four review modes?

Text review, new/pending-setting review, canon health check, and the per-five-chapter arc retrospective.

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.