中文网文正文扩写。当用户要把大纲/章纲写成正文、扩写某一章、续写下一章、把简介扩成场景、或要"按这个细纲写 3000 字"时使用。按平台字数和节奏产出可直接发布的网文正文,保持人设与伏笔连贯。
Scanned 6/14/2026
Install via CLI
openskills install tance-mang/chinese-webnovel-skills---
name: expand
description: 中文网文正文扩写。当用户要把大纲/章纲写成正文、扩写某一章、续写下一章、把简介扩成场景、或要"按这个细纲写 3000 字"时使用。按平台字数和节奏产出可直接发布的网文正文,保持人设与伏笔连贯。
---
# 正文扩写(expand)
把章纲变成可发布的正文,控制好节奏、爽点、字数和钩子。
## 何时用
- 把章纲/细纲写成正文
- 扩写、续写某一章
- 把一句话剧情扩成完整场景
## 写作前确认
1. **人称**:**默认第一人称“我”**(更沉浸、像写亲身经历,尤其知乎/女频/短篇)。动笔前**用一句话表明**"默认用第一人称写,要第三人称告诉我";用户指定第三人称就用第三。全书人称统一。(调 `references/pov-guide.md`)
2. **章纲/剧情点**:这一章要发生什么(没有就先用 `outline` 出章纲)
3. **平台 + 字数**:单章密度调 `references/platform-profiles.md`(**番茄长篇连载一章 2000–2200 字最佳**、起点 2000–3000);全书还差多少到签约/分发门槛调 `references/length-standards.md`
4. **人设档案 + 语言指纹**:调出本章出场角色的 speech_profile 对照注入,保证人物"声音"一致、不 OOC、不一人一腔(见 `skills/dialogue`、`project-memory.md` 七)
5. **前情衔接**:上一章结尾的钩子要在本章开头回应;长篇还要核对时间线/资源/伤势/战力是否接得上(见 `skills/continuity`)
## 写作要领
### 节奏
- **开头接钩**:第一段回应上章悬念或立刻进入场景,不要拖
- **场景推进**:对话 + 动作 + 心理交替,避免大段叙述
- **爽点落地**:本章规划的爽点要写到位(调 `references/trope-library.md`),打脸用"嘲讽→沉默→碾压→围观"四拍
- **章尾断钩**:结尾用章尾钩收口(调 `references/hook-library.md`)
### 文字
- **短句快节奏**:网文忌长难句,多用短句和对话推进
- **画面感**:用具体细节代替抽象形容("他攥紧了拳,指节发白"而非"他很愤怒")
- **情绪演出来别喊**:忌"我很心疼/我好痛苦",用神态/动作/生理反应/潜台词让读者自己感受(调 `references/show-emotion.md`,专项改用 `emotion` 技能)
- **爽感外显**:把读者要爽的点写出来——表情、围观者反应、对手的难堪
- **避免 AI 腔**:不堆砌排比和空洞抒情,写完可用 `deslop` 检查
- **标点干净**:简体网文只用常规标点(。,、?!…… 弯引号""'')。对话用**弯双引号""**,**不用直角引号「」『』**;**破折号——极少用**(一章顶多一次);不撒 ✦✨◆★ 等装饰符号,分隔场景用空行,章标题纯文字(详见 `references/anti-ai-checklist.md` 七)
- **留一点失控感(人味)**:别写得太平均太干净太顺——关键处给一个非理性行为、一句没说完的情绪、或一段不收尽的留白,让它像"情绪记录"不像"说明书"(深度调 `human` 技能 / `references/human-texture.md`)
- **句式有呼吸**:句长不均,短句长句交错,偶尔甩个 50+ 字长句;用省略号/破折号/换行做停顿,别整齐划一(调 `references/sentence-rhythm.md`)
- **可出双版本**:用户要时,给"爽文版(情绪高/反转爽)"+"高级暗黑版(极端/灰度/留刀)"两稿(调 `references/dark-style.md`);不要求则默认单版本
### 对话
- 每个角色说话方式不同(体现人设)
- 对话推进剧情或暴露信息,不闲聊
- 适当留白,不要每句都解释
- 口语化、有潜台词,避免书面 AI 腔(调 `references/dialogue-library.md`);要专门打磨台词/金句 → 用 `dialogue` 技能
## 输出
- 直接产出正文(markdown),标题 + 正文
- 字数贴近目标
- 结尾自带章尾钩
- 写完简述:本章爽点是什么、埋/收了哪些钩、字数
## 连贯性
- 长篇每写几章:用 `review` 查读者留存(钩子/爽点/伏笔),用 `continuity` 核对逻辑/时间线/资源/身份/伤势/战力是否前后矛盾
- 重要设定/人物状态/时间推进/战力变化,写完即提示用户**更新档案**(`memory`),否则下一章继续漂
## 下一步
正文有 AI 味 → `deslop`(量化打分定位 → `aidetect`);爽点不够 → `shuangdian`;查留存 → `review`;查逻辑/时间线 → `continuity`。
No comments yet. Be the first to comment!
Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.
Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...
**Complete production-ready guide for Google Gemini embeddings API** This skill provides comprehensive coverage of the `gemini-embedding-001` model for generating text embeddings, including SDK usage, REST API patterns, batch processing, RAG integration with Cloudflare Vectorize, and advanced use cases like semantic search and document clustering. ---
Interview, source-challenge, verify, save, and ADR-gate fuzzy coding requests into Codex-ready implementation specs. Use when a feature, bugfix, refactor, migration, repo-wide change, or architecture task needs user-verified requirements, source-backed decisions, durable architecture decisions, acceptance criteria, validation commands, rollout notes, saved spec/ADR files, and a Codex execution prompt. Do not use when already fully specified or when the user wants direct implementation now.
Use when a repo needs CodeGraph plus ast-grep for Codex MCP setup, exploration, impact analysis, structural search, or safe refactor planning.