chinese-novelist skillSAFE
🎭 AI 写小说:从零生成 10-50 章完整中文小说,三层问答 · 创作记忆 · 悬念钩子 · 自动校验,长篇网文连载皆宜|开源免费,适配主流 coding agent|🤝 Infistar.cc 赞助 · 注册送 $5|AI novel writing skill
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
让 AI 为你写一部完整的中文小说
[](https://github.com/PenglongHuang/chinese-novelist-skill/releases) [](https://claude.com/claude-code) [](LICENSE)
v2.0 重构发布:引入三层递进式问答、偏好记忆系统、中断续写、三种写作模式和自动校验修复! 查看 v1.0 → v2.0 升级说明 了解详情。
为什么用这个?
写小说最难的是坚持写完。这个 Skill 专为解决这个痛点而生:
- 智能问答 - 三层递进式问答,支持快速跳过和随机生成
- 偏好记忆 - 跨会话学习你的喜好,下次创作更贴心
- 中断续写 - 意外中断?自动检测并从断点继续
- 多种写作模式 - 串行 / 子Agent并行 / Agent Teams,按需选择
- 自动校验 - 完成后自动检查字数和连贯性,不合格自动重写
- 每章必爽 - 开头即高潮,结尾留悬念
快速开始
安装skill:npx skills add PenglongHuang/chinese-novelist-skill
输入指令:使用 chinese-novelist 帮我写一部小说
赞助商
chinese-novelist-skill × [Infistar.cc](https://www.infistar.cc/register?aff=29SS8B9E&ref_source=link)
用这个技能包写长篇的朋友都知道:动辄几十万字的大纲回溯、多代理并行出章,额度往往比灵感先耗尽。
如果想放开手脚写长篇,推荐换上 Infistar:
- 零配置接入:无需修改代码,Claude Code / 客户端只需换上 Base URL 与 Key 即可开工。
- 超低成本试错:多模型价格极具优势。整卷重构、换文风重写、多代理并行爆肝不再肉疼。
- 全模型任意切:Claude、GPT、Gemini、DeepSeek 一个 Key 全搞定,改个模型名就能对比打斗与感情戏表现。
- 国内顺畅使用:RMB 直接结算,全模型支持验真防掺水。
- 专属福利:通过专属通道注册即送 $5 体验额度。建议领完拿同一份大纲让两个模型各试跑 3 章,感受文笔差异!
使用过程
Phase 1 — 交互问答 Phase 2 — 规划确认
Phase 3 — 疯狂创作 Phase 4 — 完
0867a0f23c32OBSERVED · 2026-10-20Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| claude-code | mentioned |
What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
---
name: chinese-novelist
description: |
当用户要求写小说、创作故事、分章节写作、连续剧情、长篇小说,或需要续写/校验未完成的小说项目时使用。支持悬疑/言情/奇幻/科幻/历史等全题材中文创作。
metadata:
trigger: 创作中文小说、分章节故事、长篇小说创作
source: 基于小说创作最佳实践设计
---
# Chinese Novelist: 中文小说创作助手
## 三大黄金法则
1. **展示而非讲述** - 用动作和对话表现,不要直接陈述
2. **冲突驱动剧情** - 每章必须有冲突或转折
3. **悬念承上启下** - 每章结尾必须留下钩子
## 特性说明
- **中断续写**:自动检测未完成项目,从断点继续创作
- **自动校验**:创作完成后自动检查字数和质量,不合格自动修复
- **并行写作**(可选):支持子Agent并行写作,通过 `02-写作计划.json` 协调状态
## 核心流程
进入每个阶段时,先阅读对应的流程文档以获取详细执行指令。
### 第0步:初始化与偏好加载
读取用户偏好,检测未完成项目(中断续写),展示个性化欢迎。 → 详见 [phase0-initialization.md](references/flows/phase0-initialization.md)
### 第一阶段:三层递进式问答
通过递进式问答收集创作需求,确定小说定位与标题:
- **核心定位**(必答,Q1-Q3):题材创意、主角设定、核心冲突 → 详见 [phase1-layer1-core.md](references/flows/phase1-layer1-core.md)
- **深度定制与规格**(2 次合并提问):世界观、视角基调、核心主题、读者定位、章节数量、配置确认 → 详见 [phase1-layer2-customize.md](references/flows/phase1-layer2-customize.md)
- **标题生成**:AI 基于创意元素生成候选标题,用户选择或自定义 → 详见 [phase1-layer3-title.md](references/flows/phase1-layer3-title.md)
### 第二阶段:规划 + 二次确认
创建项目文件夹(`./chinese-novelist/{timestamp}-{小说名称}/`),生成大纲、人物档案和写作计划JSON,等待用户确认。 → 详见 [phase2-planning.md](references/flows/phase2-planning.md)
### 第2.5步:写作模式选择
规划确认后,选择写作模式:
- **逐章串行**(`serial`):主 Agent 自己逐章写,全程无中断
- **子Agent并行**(`subagent-parallel`):将章节分成批次,派生子 Agent 并行写作
- **Agent Teams**(`agent-teams`):Claude Code 多 Agent 协作模式,Agent 间可通讯(需手动开启)
→ 详见 [phase3-writing.md](references/flows/phase3-writing.md)
### 第三阶段:疯狂创作(无需用户确认)
> 切记,一旦进入这个阶段,所有过程都禁止向用户确认。用户就是你的读者,你必须把完整的小说创作完成才能与用户报告
根据用户选择的写作模式(串行/并行/Teams)逐章执行创作流程。每章创作前必须读取 `01-大纲.md` 中对应章节的规划信息,严格按大纲创作。支持中断续写。 → 详见 [phase3-writing.md](references/flows/phase3-writing.md)
### 第四阶段:自动校验与修复(无需用户确认)
全程无需用户介入,自动检查所有章节完成度和字数,不合格章节自动重写(最多3轮)。 → 详见 [phase4-validation.md](references/flows/phase4-validation.md)
## 共享机制
偏好系统、写作计划系统、黄金法则详解、字数检查脚本等跨阶段共享机制。 → 详见 [shared-infrastructure.md](references/flows/shared-infrastructure.md)Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (0)
No findings outside the package's declared scope.
Gates applied: no_behavioural_pass.
0867a0f23c32full audit observations/trust-audit/skill/penglonghuang__chinese-novelist.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-20 | 0867a0f23c32 | SAFE | B | 89 | first audit |
Questions
What does the chinese-novelist — penglonghuang/chinese-novelist-skill skill do?
🎭 AI 写小说:从零生成 10-50 章完整中文小说,三层问答 · 创作记忆 · 悬念钩子 · 自动校验,长篇网文连载皆宜|开源免费,适配主流 coding agent|🤝 Infistar.cc 赞助 · 注册送 $5|AI novel writing skill
Is chinese-novelist — penglonghuang/chinese-novelist-skill safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can chinese-novelist — penglonghuang/chinese-novelist-skill access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
Which assistants does chinese-novelist — penglonghuang/chinese-novelist-skill work with?
Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (0867a0f23c32), read on 2026-10-20. The repository is watched, and a new audit runs when it changes — this is the first audit.
Add this badge to your README
Shows this page's audit result and links back to it. It updates when the audit does.
[](https://modelsatlas.com/skills/penglonghuang/chinese-novelist/)
Not your listing, or something out of date? Claim or update your listing.