sumeru-finalize
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
Chinese网文完稿校验
Web Novel Manuscript Finalization Verification
触发关键词
Trigger Keywords
小说写完了帮我检查下、导出适合起点/番茄的格式、查有没有错别字、检测敏感词、整理成发布版本、帮我导出小说发布格式、检查小说错别字、敏感词检测、小说完稿检查、适配起点格式导出、番茄小说格式导出、小说排版整理、多平台格式导出、小说完稿导出
Help me check my finished novel, Export formats suitable for Qidian/Fanqie, Check for typos, Detect sensitive words, Organize into publishable version, Help me export novel publishing format, Check novel typos, Sensitive word detection, Finished novel inspection, Export adapted to Qidian format, Fanqie Novel format export, Novel typesetting organization, Multi-platform format export, Finished novel export
核心功能
Core Features
- 错别字、标点符号、语法错误检查
- 敏感内容、违规内容排查
- 格式规范统一:章节标题、段落格式、标点规范
- 全文字数统计、完稿报告生成
- 适配不同平台发布格式导出
- 批量替换功能
- 自动分段功能
Skill 边界说明:技术性文字校验(错别字/标点/语法)由本 Skill 负责,Skill 专注于文笔和内容层面的优化,两者互补不重叠。用户在润色后仍应通过 finalize 进行最终技术校验。sumeru-polish
- Check for typos, punctuation marks, and grammatical errors
- Screening of sensitive and non-compliant content
- Unified format specification: chapter titles, paragraph formatting, punctuation standards
- Full text word count and manuscript finalization report generation
- Export adapted to different platform publishing formats
- Batch replacement function
- Automatic paragraph segmentation function
Skill Boundary Description: This Skill is responsible for technical text verification (typos/punctuation/grammar), while theSkill focuses on stylistic and content-level optimization. The two complement each other without overlap. Users should still perform final technical verification via finalize after polishing.sumeru-polish
子Agent并行校验机制
Sub-Agent Parallel Verification Mechanism
当需要校验/导出的章节数量大于3章时,自动启用子Agent并行处理模式:
⚠️ 遵循全局约束:每个子Agent最多负责3个章节(详见 AGENTS.md "子Agent并行处理规则")
- 所需Agent数 = ceil(总章节数 / 3)
- 相邻章节分配给同一Agent,保持格式处理一致性
调度逻辑
mermaid
flowchart LR
A[批量校验任务] --> B[读取章节列表与校验配置]
B --> C[创建任务队列,按每Agent最多3章分配]
C --> D[计算所需Agent数 = ceil/总章数/3/]
D --> E[启动N个并行子agent]
E --> F[子Agent拉取任务 → 校验章节 → 生成错误报告 → 导出格式]
F --> G{队列是否为空?}
G -->|否| F
G -->|是| H[汇总校验结果,合并错误报告,生成完稿报告]章节分配规则
- 按章节顺序连续分配(如Agent1负责第1-3章,Agent2负责第4-6章)
- 尾部不足3章的Agent按实际剩余章节数分配
每个子Agent的校验范围
- 错别字、标点、语法错误检查
- 敏感词检测(按三级标准)
- 格式规范统一
- 对应平台格式导出
When the number of chapters to be verified/exported exceeds 3, the sub-Agent parallel processing mode is automatically enabled:
⚠️ Follow global constraints: Each sub-Agent is responsible for a maximum of 3 chapters (see AGENTS.md "Sub-Agent Parallel Processing Rules")
- Required number of Agents = ceil(total chapters / 3)
- Adjacent chapters are assigned to the same Agent to maintain consistency in format processing
Scheduling Logic
mermaid
flowchart LR
A[批量校验任务] --> B[读取章节列表与校验配置]
B --> C[创建任务队列,按每Agent最多3章分配]
C --> D[计算所需Agent数 = ceil/总章数/3/]
D --> E[启动N个并行子agent]
E --> F[子Agent拉取任务 → 校验章节 → 生成错误报告 → 导出格式]
F --> G{队列是否为空?}
G -->|否| F
G -->|是| H[汇总校验结果,合并错误报告,生成完稿报告]Chapter Allocation Rules
- Allocate continuously according to chapter order (e.g., Agent 1 is responsible for Chapters 1-3, Agent 2 for Chapters 4-6)
- Agents with less than 3 chapters remaining at the end are assigned according to the actual number of remaining chapters
Verification Scope of Each Sub-Agent
- Check for typos, punctuation, and grammatical errors
- Sensitive word detection (according to three-level standards)
- Unified format specification
- Format export corresponding to the platform
输出内容
Output Content
- 错误列表:错别字、标点错误、语法问题
- 敏感内容提示:需要调整的违规内容
- 校验后的纯净版全文
- 完稿报告:总字数、章节数、核心内容摘要
- 多平台发布格式版本(起点、番茄、晋江等)
- Error list: typos, punctuation errors, grammatical issues
- Sensitive content prompts: non-compliant content that needs adjustment
- Clean full-text version after verification
- Manuscript finalization report: total word count, number of chapters, core content summary
- Multi-platform publishing format versions (Qidian, Fanqie, Jinjiang, etc.)
各平台导出格式规则
Export Format Rules for Each Platform
起点中文网(qidian)
Qidian Chinese Network (qidian)
- 章节标题格式:,居中对齐
第X章 标题内容 - 段落首行缩进2字符
- 每段空一行
- 标点符号使用中文全角
- 章节字数建议3000-5000字
- 禁止使用特殊符号作为章节标题
- 对话单独成段
- Chapter title format: , centered alignment
Chapter X Title Content - First line indent of 2 characters for paragraphs
- One blank line between paragraphs
- Use Chinese full-width punctuation marks
- Recommended chapter word count: 3000-5000 words
- Prohibited use of special symbols as chapter titles
- Dialogue is in separate paragraphs
番茄小说(fanqie)
Fanqie Novel (fanqie)
- 章节标题格式:
第X章 标题内容 - 段落首行不缩进,段落间空一行
- 每句尽量简短,适合移动端阅读
- 章节字数建议2000-3000字
- 重点内容可使用加粗标记
- 对话使用引号包裹,说话人单独成段或句尾注明
- Chapter title format:
Chapter X Title Content - No first line indent for paragraphs, one blank line between paragraphs
- Keep sentences as short as possible for mobile reading
- Recommended chapter word count: 2000-3000 words
- Key content can use bold formatting
- Dialogue is wrapped in quotation marks, with speakers in separate paragraphs or noted at the end of the sentence
晋江文学城(jjwxc)
Jinjiang Literature City (jjwxc)
- 章节标题格式:
第X章 标题内容 - 支持HTML格式标签
- 段落首行缩进2字符
- 章节字数建议2500-4000字
- 作者有话要说区域单独设置
- 支持章节提要
- Chapter title format:
Chapter X Title Content - Supports HTML format tags
- First line indent of 2 characters for paragraphs
- Recommended chapter word count: 2500-4000 words
- Separate setting for the "Author's Notes" section
- Supports chapter summaries
纵横中文网(zongheng)
Zongheng Chinese Network (zongheng)
- 章节标题格式:
第X章 标题内容 - 段落首行缩进2字符
- 章节字数建议3000-6000字
- 支持分卷设置
- 标点规范使用中文全角
- Chapter title format:
Chapter X Title Content - First line indent of 2 characters for paragraphs
- Recommended chapter word count: 3000-6000 words
- Supports volume division settings
- Use Chinese full-width punctuation marks in accordance with standards
17K小说网(17k)
17K Novel Network (17k)
- 章节标题格式:
第X章 标题内容 - 段落首行缩进2字符
- 章节字数建议2000-4000字
- 支持章节预览
- 每章结束可设置下章预告
- Chapter title format:
Chapter X Title Content - First line indent of 2 characters for paragraphs
- Recommended chapter word count: 2000-4000 words
- Supports chapter previews
- Can set a preview for the next chapter at the end of each chapter
敏感词检测标准
Sensitive Word Detection Standards
一级敏感(必须修改)
Level 1 Sensitivity (Must be modified)
- 违反国家法律法规的内容
- 涉及政治敏感人物、事件
- 色情、淫秽描写
- 暴力、恐怖内容
- 分裂国家、破坏民族团结言论
- 宗教极端内容
- Content violating national laws and regulations
- Involving politically sensitive figures and events
- Pornographic and obscene descriptions
- Violent and terrorist content
- Speech splitting the country and undermining ethnic unity
- Religious extremist content
二级敏感(建议修改)
Level 2 Sensitivity (Recommended modification)
- 过于血腥暴力的细节描写
- 低俗用语、粗口
- 可能引起不适的医疗描写
- 涉及未成年人的不当内容
- 赌博、毒品相关描写
- 侵犯他人隐私的内容
- Excessively bloody and violent detailed descriptions
- Vulgar language and swear words
- Medical descriptions that may cause discomfort
- Inappropriate content involving minors
- Descriptions related to gambling and drugs
- Content infringing on others' privacy
三级敏感(优化建议)
Level 3 Sensitivity (Optimization suggestions)
- 网络用语过多影响阅读
- 容易产生歧义的表述
- 可能引起争议的话题
- 过度使用网络热梗
- 重复冗余的表述
- Excessive internet slang affecting reading
- Ambiguous expressions
- Potentially controversial topics
- Overuse of internet memes
- Repetitive and redundant expressions
错误分级提示
Error Level Prompt
严重错误(红色标记)
Severe Errors (Marked in red)
- 错别字导致语义完全改变
- 敏感词一级违规内容
- 章节标题格式完全不符合规范
- 段落结构严重混乱
- 标点符号大面积错误
- Typos that completely change the semantic meaning
- Level 1 sensitive word violations
- Chapter title format completely non-compliant with specifications
- Severe confusion in paragraph structure
- Large-scale punctuation errors
中等错误(黄色标记)
Moderate Errors (Marked in yellow)
- 一般错别字
- 标点符号使用不规范
- 敏感词二级内容
- 段落格式不统一
- 语法错误影响理解
- General typos
- Non-standard use of punctuation marks
- Level 2 sensitive word content
- Inconsistent paragraph formatting
- Grammatical errors affecting understanding
轻微错误(蓝色标记)
Minor Errors (Marked in blue)
- 建议优化的用词
- 标点符号使用可以更规范
- 敏感词三级内容
- 段落排版可进一步美化
- 重复性表述建议
- Recommended optimized wording
- Punctuation marks can be more standardized
- Level 3 sensitive word content
- Paragraph typesetting can be further beautified
- Suggestions for repetitive expressions
批量替换功能说明
Batch Replacement Function Description
功能特性
Feature Characteristics
- 支持全局批量替换指定词汇
- 支持正则表达式替换
- 支持替换前预览确认
- 支持多组替换规则同时执行
- 支持替换历史记录查询
- Supports global batch replacement of specified vocabulary
- Supports regular expression replacement
- Supports preview confirmation before replacement
- Supports simultaneous execution of multiple sets of replacement rules
- Supports querying replacement history records
使用场景
Usage Scenarios
- 角色名统一修改
- 地名、设定名称批量调整
- 敏感词批量替换
- 标点符号统一规范
- 网络用语批量转换
- Unified modification of character names
- Batch adjustment of place names and setting names
- Batch replacement of sensitive words
- Unified specification of punctuation marks
- Batch conversion of internet slang
操作方式
Operation Methods
- 预设规则:选择常用替换规则模板
- 自定义规则:手动输入查找内容和替换内容
- 正则模式:使用正则表达式进行复杂匹配
- 确认替换:查看替换预览后确认执行
- Preset rules: Select commonly used replacement rule templates
- Custom rules: Manually enter search content and replacement content
- Regular expression mode: Use regular expressions for complex matching
- Confirm replacement: View replacement preview before confirming execution
自动分段功能说明
Automatic Paragraph Segmentation Function Description
功能特性
Feature Characteristics
- 智能识别对话与叙述内容
- 根据句子长度自动分段
- 支持自定义分段字数阈值
- 保持段落逻辑完整性
- 对话自动单独成段
- Intelligent recognition of dialogue and narrative content
- Automatic paragraph segmentation based on sentence length
- Supports custom paragraph word count threshold
- Maintains logical integrity of paragraphs
- Dialogue is automatically placed in separate paragraphs
分段规则
Segmentation Rules
- 对话优先:对话内容自动单独成段
- 字数控制:单段建议100-300字
- 逻辑完整:避免在句子中间分段
- 场景切换:场景转换时自动分段
- 心理活动:大段心理描写适当分段
- Dialogue priority: Dialogue content is automatically placed in separate paragraphs
- Word count control: Recommended 100-300 words per paragraph
- Logical integrity: Avoid segmentation in the middle of a sentence
- Scene switching: Automatic segmentation when scenes change
- Psychological activities: Proper segmentation of long psychological descriptions
可配置参数
Configurable Parameters
- 最大段落字数(默认300字)
- 最小段落字数(默认50字)
- 是否强制对话单独成段
- 是否在场景切换时加分隔线
- 是否保留原有分段结构
- Maximum paragraph word count (default 300 words)
- Minimum paragraph word count (default 50 words)
- Whether to force dialogue into separate paragraphs
- Whether to add dividing lines when scenes switch
- Whether to retain original paragraph structure
数据持久化
Data Persistence
完稿数据自动保存到 目录:
.sumeru/finalize/- :校验后的纯净版全文
clean/full-text.md - :按章节拆分的纯净版文件
clean/chapters/ - :错误列表,含错别字、标点、敏感词等所有问题
error-report.json - :完稿统计报告,总字数、章节数、平均章节长度等
stats.json - :各平台导出配置参数
export-config.json
用户可见输出(当前工作目录):
- :各平台导出版本,按平台名分类存放
publish/
Manuscript data is automatically saved to the directory:
.sumeru/finalize/- : Clean full-text version after verification
clean/full-text.md - : Clean version files split by chapter
clean/chapters/ - : Error list including all issues such as typos, punctuation, sensitive words, etc.
error-report.json - : Manuscript finalization statistical report, including total word count, number of chapters, average chapter length, etc.
stats.json - : Export configuration parameters for each platform
export-config.json
User-visible output (current working directory):
- : Export versions for each platform, stored by platform name
publish/
与其他 Skill 配合
Cooperation with Other Skills
- 前置 Skill:读取最终章节内容
- 默认从 读取最新章节内容(review 修复和 polish 润色直接修改 chapters/,无需额外操作)
chapters/
- 默认从
- Preceding Skills: Read final chapter content
- By default, reads the latest chapter content from (review fixes and polish revisions directly modify chapters/, no additional operations required)
chapters/
- By default, reads the latest chapter content from
数据复用
Data Reuse
- 可随时重新导出其他平台格式,无需重新校验
- 错误报告可作为后续写作的规避参考
- 支持增量导出,修改部分章节后仅重新生成对应章节的平台版本
- Can re-export other platform formats at any time without re-verification
- Error reports can be used as reference for avoiding issues in subsequent writing
- Supports incremental export: only regenerate platform versions for corresponding chapters after modifying some chapters