awesome-novel skill
A novel-writing skill with nine agents: setting, volume and chapter plans, prose, a machine-style check and an archive with character memory
- 836 ★ on the repository
- Last commit
- AGPL-3.0
Extended access
Why this level
- Runs Python scripts: init.py creates project files, sync-project.py updates them and install.sh copies the skill into the agent directory
- Nine agents write and overwrite book and memory files in the project folder
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/awesome-novelDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Runs Python scripts: init.py creates project files, sync-project.py updates them and install.sh copies the skill into the agent directory
- Nine agents write and overwrite book and memory files in the project folder
This is third-party code. Review the repository files before installing.
What it does
awesome-novel is a skill for writing a novel together with an AI agent. The `/awesome-novel` command in an empty folder runs `init.py`: it creates a workspace with a story.md file, settings, volumes, chapters, prompts and archives folders, definitions of nine agents, a knowledge base and memory files. After that the novel-agent takes control; it only distributes tasks and verifies results. First the author discusses genre (24 ready profiles exist), world, characters and writing style with the agent; style can be derived from text samples. Then volume-planner builds a volume plan: emotional arc, conflict steps and the gap in what characters know. chapter-planner splits a chapter into scenes. prompt-crafter assembles the request, writer drafts the chapter, anti-ai checks the text in six steps (Gate A-F) and scores it, reader gives an in-depth review on request, and updater archives the chapter, updates character state and tracks plot threads. Style rules and the author's preferences accumulate in memory files and become permanent over time.
Who it is for. For web novel and book authors who write with the help of Claude Code, OpenCode or Codex and want to keep the world, characters, chapters and edits in one project.
Good fit when
- A novel is written from scratch and you need a path from genre and world through volume and chapter plans to finished chapters
- A long book loses threads: you must track characters, plot hooks and chapter rhythm
- You want to adopt a writing manner from text samples and keep it across all chapters
- A chapter draft sounds machine-made and has to be checked and cleaned step by step with a score
- You need a short story: the skill has a separate branch with planning, writing and checking
Not a fit when
- You need articles, posts or marketing copy: the skill is built around a fiction novel
- You write the book in Russian or any language other than Chinese: instructions, genre profiles and examples in the repository are Chinese
- You need one paragraph or a fragment edit: nine agents and a project file structure are excessive
- You need the novel for commercial publication without the skill author's consent: personal use is free, commercial use needs permission
- You work in an agent missing from the skill's platform list: Claude Code, OpenCode, Codex, ZCode, DeepSeek Harness, Grok Build, Reasonix
Example request
Help me write a xianxia novel: let's discuss the world and the hero, then outline the first volume and write the first chapterLimitations
- The skill is built for Chinese web novels: the main documentation, genre profiles, examples and machine-style check rules are in Chinese (the main file has an English version).
- Python 3.9 or newer is required. For OpenCode, Codex, ZCode, DeepSeek Harness and Grok Build `pyyaml` is also installed.
- Terms are double: the repository is under AGPL-3.0, while the main file adds that personal use is free and commercial use requires contacting the author.
- The author states that the project provides no models or tokens, and payment and stability come from third-party services.
- For Reasonix the skill is installed at project level with `init.py`, not through `install.sh`.
- Version 4.28.0. Project initialization and migration run through Python scripts that the agent launches after your confirmation.
How to disable. Delete the skill folder: `~/.claude/skills/awesome-novel/` for Claude Code, `~/.config/opencode/skills/awesome-novel/` for OpenCode, `~/.codex/skills/awesome-novel/` for Codex. Agent and memory folders inside your book projects (`.claude`, `.opencode`, `.codex`) stay and can be deleted by hand. Chapter texts live in archives and do not depend on the skill.
FAQ
How do I install a novel-writing skill in Claude Code?
Clone the repository and run ./install.sh claude-code; the skill lands in ~/.claude/skills/awesome-novel/. You can also tell the agent to install awesome-novel-agent and it does the same. Then type /awesome-novel in the folder of your future book. On Windows use install.ps1 with the same platform argument.
How many agents does awesome-novel have and what does each do?
Nine: novel-agent distributes work, volume-planner plans a volume, chapter-planner a chapter, prompt-crafter assembles the request, writer drafts the chapter, anti-ai cleans machine style, reader reviews, updater archives, style-distiller derives style parameters from samples. The repository also has agents for short stories.
How do I keep a novel text from sounding AI-written?
In awesome-novel the anti-ai agent handles this: it checks a chapter in six steps Gate A-F, computes a score and assigns a level. Rules against machine style are also injected into the writer's request. The author warns that the rules are written for Chinese text.
Can I use awesome-novel for free for a commercial book?
The main file says personal use is free and commercial use requires contacting the author. The repository license is AGPL-3.0 at the same time. Before publishing a book for money, clarify the terms with the author.
How does awesome-novel remember characters and plot hooks between chapters?
After each approved chapter the updater agent appends character state, the emotional arc and relationships to project files, scans unresolved and new plot hooks and warns about stretches of tense or calm chapters that run too long. This data lives in plain Markdown files.
Discussion
No questions or reviews yet.
Related
Skills24 skills for fiction in Markdown, from premise to print, and a story command that catches plot contradictions before readers do
A set of 13 skills for writing web fiction inside a coding agent: rank scanning, hit deconstruction, outline and prose, AI-flavor cleanup, cover art
A Claude Code plugin for long web novels: volume planning, chapter writing and review, memory of characters and plot, a state dashboard
A plugin for novels and serial fiction: a muse agent leads from idea to draft, while writer, critic, editor and canon-check agents each handle one job
A skill for writing a complete Chinese novel chapter by chapter: three-layer intake, preference memory, cliffhangers and automatic validation
A system for books and scripts: a CLI, Studio and an agent skill that write a novel by chapters, short fiction and scripts, and translate long texts
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| awesome-novel (this entry) | no | 836 | |
| Webnovel Writer | no | 7,626 | a Claude Code plugin for long web novels with a state viewing dashboard |
| chinese-novelist | no | 4,124 | a skill for writing a whole novel in Chinese by chapters with a three-level interview |
README in short
The repository holds the main skill file, definitions of nine agents, a knowledge base on writing craft, scenes, characters, plot and machine-style checks, project templates, the `init.py` and `sync-project.py` scripts and installers for Linux, macOS and Windows, plus release notes. The main file leads from installation to the first chapter and describes the project structure, the chapter work cycle and integrations with seven platforms.
README badge
Original SKILL.md
--- name: awesome-novel description: 和 AI 协作写小说的工作流系统。9 个 agent 协作完成从设定到归档的完整写作流程。入口检测 → 初始化/迁移 → 交 novel-agent 调度。适用场景：从零写新小说、导入已有小说。 --- # Novel — 小说创作工作流 和 AI 一起写小说。本 skill 负责项目状态检测、新项目初始化、旧版项目自动迁移，完成后将控制权交给 novel-agent。 **唤起方式：** 用户在项目目录输入 `/awesome-novel`（Claude Code / OpenCode / Grok Build 直接输入；Codex 输入 `/use awesome-novel`；或说"帮我写本小说"）即进入下方检测流程。新目录会先询问作者，确认后运行 `init.py` 在本地初始化小说工作空间。 ## OpenCode 集成说明 本 skill 也支持 OpenCode。安装在 `~/.config/opencode/skills/awesome-novel/` 后，项目初始化脚本会自动部署 agent 定义到 `.opencode/agents/`，OpenCode 即可自动发现： - `@novel-agent` — 加载总指挥 agent - `@volume-planner`、`@chapter-planner` 等 — 加载子 agent - **Task 工具** — novel-agent 通过 Task 工具调用子 agent **调度机制：** novel-agent 写 order 文件到 `.agent/task/`（`status: pending`）→ Task 工具调用子 agent → 子 agent 读取 order 执行 → 完成后将 order 覆盖为 `status: DONE` 后退回。 ## Codex 集成说明 本 skill 也支持 Codex。skill 本体**用户级安装**到 `~/.codex/skills/awesome-novel/`；`init.py --platform codex` 初始化小说项目时，把 9 个自定义 agent 部署为项目级 `.codex/agents/*.toml`（Codex 官方 TOML 格式，含 `name`/`description`/`developer_instructions`）： - `@novel-agent` — 总指挥 agent（TOML 名 `novel-agent`） - `@volume-planner`、`@chapter-planner` 等 — 子 agent（TOML 名与源 agent 一致） - **独立工具** — `memory-recording`、`roleplay-sandbox` 部署为 `.codex/skills/<name>/SKILL.md` **调度机制：** novel-agent 写 order 文件到 `.agent/task/`（`status: pending`）→ 用 `spawn_agent` 调度子 agent（agent 名 = `.codex/agents/*.toml` 的 name）→ 子 agent 读取 order 执行 → 完成后将 order 覆盖为 `status: DONE` 后退回。order 文件协议与其余平台完全一致。 ## ZCode 集成说明 本 skill 也支持 ZCode。ZCode 的 skill 约定（目录 + `SKILL.md`）与 Claude Code 同源，天然兼容；但 ZCode 无项目级 agents 目录，agents 即 skills。skill 本体**用户级安装**到 `~/.zcode/skills/awesome-novel/`；`init.py --platform zcode` 初始化小说项目时，把 9 个 agent 部署为项目级 `.zcode/skills/<name>/SKILL.md`（与 Reasonix 同构，另含 memory-recording、roleplay-sandbox 独立工具，共 11 个 skill）： - novel-agent — 总指挥（入口调度者，由 ZCode 按描述自动发现，无 `@` 语法） - writer、volume-planner、chapter-planner 等 — 子 agent（subagent skill，由 novel-agent 用 Agent 工具按 skill 名 spawn） **调度机制：** novel-agent 写 order 文件到 `.agent/task/`（`status: pending`）→ 用 `Agent` 工具调度子 agent（子 agent 名 = `.zcode/skills/` 下的 skill 名）→ 子 agent 读取 order 执行 → 完成后将 order 覆盖为 `status: DONE` 后退回。order 文件协议与其余平台完全一致。 ## dsh 集成说明 本 skill 也支持 DeepSeek Harness（dsh，DeepSeek 官方的开源 agent harness）。dsh 的 skill 约定（目录 + `SKILL.md`，frontmatter 只认 `name`/`description`）与 Claude Code 同源，天然兼容；但 dsh 无项目级 agents 目录（subagent 是运行时能力），agents 即 skills。skill 本体**用户级安装**到 `~/.dsh/skills/awesome-novel/`；`init.py --platform dsh` 初始化小说项目时，把 9 个 agent 部署为项目级 `.dsh/skills/<name>/SKILL.md`（与 ZCode 同构，另含 memory-recording、roleplay-sandbox 独立工具，共 11 个 skill；`<项目根>/.dsh/skills/` 是 dsh 的项目级 skill 根，自动发现且优先级最高）： - novel-agent — 总指挥（入口调度者，由 dsh 按 name/description 自动路由，无 `@` 语法） - writer、volume-planner、chapter-planner 等 — 子 agent（subagent skill，由 novel-agent 用 `subagent` 工具调度） **调度机制：** novel-agent 写 order 文件到 `.agent/task/`（`status: pending`）→ 用 `subagent` 工具调度子 agent（prompt 中要求子 agent 先调用 `skill(name="<子agent名>")` 加载自身指令；子 agent 名 = `.dsh/skills/` 下的 skill 名）→ 子 agent 读取 order 执行 → 完成后将 order 覆盖为 `status: DONE` 后退回。order 文件协议与其余平台完全一致。