Guide · Agent skill
How to use the Import an existing novel agent skill (story-import)
Updated · written with AI from the project's own documentation and checked against our catalogue data
story-import is an agent skill from the oh-story-claudecode set. It takes a novel you have already written, whether an unfinished draft or a completed book, and reverse-parses it into a standard writing project with settings, outlines, manuscript folders and tracking files. This guide is for authors who want to keep writing that book with AI help, and it explains how to install the skill, trigger it with plain requests, and check the results before relying on them.
Quick start
- 1Install the skill. Download the ZIP from the page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
- 2Run story-setup. Run /story-setup first so the writing environment and analysis helpers are ready. This is recommended, not required.
- 3Start a new conversation. Open a fresh conversation in the project so the import has room to work.
- 4Trigger the import. Type /story-import or a request such as 把我的书导进来 and provide the file path or paste the text.
- 5Confirm scope. Answer the agent's questions about title, genre, platform, length and whether the last chapter is complete.
- 6Review the report. When the agent reports back, check the structure and the open decisions before confirming activation.
What the story-import skill makes the agent do
The skill reverse-engineers an existing book into a project that the other oh-story writing skills can continue. It first analyzes the text by reusing the story-long-analyze or story-short-analyze pipeline, choosing the path by length. Then it migrates the results into a project layout.
The analysis output is kept as a source library in 拆文库/{书名}/. The project itself is built from that analysis, and the agent reports back to you for confirmation before activating the book.
- Long novels: settings, chapter outlines, manuscript folders and tracking files are created, then a chapter-by-chapter outline is reverse-engineered in batches.
- Short stories: the result is three single files rather than the long-form folder structure.
- Continuation: after import, you can continue with /story-long-write or /story-short-write.
Install the story-import skill
You can install the skill from the download on our page. It arrives as a ZIP file. Unzip it into one of these locations: ~/.claude/skills/ for personal use across projects, or .claude/skills/ inside a specific project folder.
If you use the Claude apps instead, upload the skill under Settings → Capabilities → Skills.
- Personal install: unzip into ~/.claude/skills/
- Project install: unzip into the project's .claude/skills/ folder
- Claude apps: Settings → Capabilities → Skills, then upload
Set up the writing environment first (recommended)
The skill works best with the rest of the oh-story-claudecode skill set. The source recommends running /story-setup before importing so that a dedicated analysis helper can process the book section by section. Without it, the import still runs, but everything happens in the main conversation, which is slower and can fill the context.
Some phases use bundled scripts: a Python 3 script (build_outline_brief.py) and a Node.js script (check-outline-contract.js). Agent sub-agent support is optional for the faster analysis path.
- Recommended: run /story-setup, then start a new conversation and use /story-import
- Optional: sub-agent support for faster analysis; without it the main session is used
- Required for some phases: Python 3 and Node.js
/story-setupTrigger the skill with natural requests
You can invoke the skill with the slash command or with a plain request. The skill's description lists these triggers: /story-import, 「导入小说」, 「反向解析」, 「导入」 and 「把我的书导进来」. These mean roughly 'import a novel', 'reverse-parse', 'import' and 'bring my book in'.
When you start, the agent first asks what you want. If you only want an analysis and no project files, it says so and points you to the analysis skill instead.
- /story-import
- 导入小说
- 把我的书导进来
- If you only want analysis: /story-long-analyze or /story-short-analyze
/story-importWalk through an import step by step
The import runs in phases. At the start the agent confirms the book title, genre, target platform, whether the book is complete, the length class (long or short), and whether the final chapter is complete. For an unfinished draft, you decide whether to continue from the partial chapter or finish it before importing. The agent records your decisions in 导入记录.md inside the book folder.
Later phases run analysis, migrate the structure, initialize tracking for long works, and then report back. The report asks you to check the results and decide the open points before the book is activated through .active-book.
- Phase 1: confirm scope and write the import record
- Phase 2: deep analysis into 拆文库/{book name}/
- Phase 3: structure migration into the project folders
- Phase 4 (long works only): tracking initialization
- Phase 5: report, self-check and set .active-book
Know what the skill needs and what it touches
The skill can read and write files and runs scripts. It also uses the network, and sub-agents may be network-connected. Read this before uploading a manuscript you have not published. The source does not describe where content is sent beyond these capabilities, so check your setup and the skill set's documentation yourself before importing sensitive material.
The skill does not register your book as a benchmark. Your own book is never added to the benchmark folder, and no benchmark is set unless you name an external reference work yourself.
- Runs scripts: yes (Python 3 and Node.js helpers)
- Uses network: yes
- Reads and writes project files
- Keep unpublished or confidential manuscripts out until you have checked where content goes
What people use it for
- Continuing a partly written Chinese web novel with an AI agent that knows the existing settings and outline.
- Turning a completed book into an organized settings, outline and tracking structure for a sequel.
- Bringing a short story into single-file settings, outline and manuscript form for further short-form writing.
- Using story-long-write or story-short-write after import to write the next chapter.
Frequently asked questions
Is the story-import skill free to use?
The source states that the skill is released under the MIT licence. It does not list prices, so check the repository for the licence text and any terms that apply to your own use.
What do I need before installing story-import?
You need the oh-story-claudecode skill set. Running story-setup is recommended for the writing environment. Some phases need Python 3 and Node.js, and sub-agent support is optional.
Can I use story-import with a book that is not finished?
Yes. The skill is designed for both partial drafts and completed books. For a partial draft, it asks whether to continue from an unfinished last chapter or complete it before importing.
Does story-import only analyze a book?
No. It creates project files. If you only want a one-off analysis, the source says to use /story-long-analyze or /story-short-analyze instead.
Why does the agent stop and ask me to confirm?
Before activating the book, the agent reports the structure and asks you to check the results. Where the original text has no clear volume breaks, the volume division must be confirmed by you.
Is it safe to upload an unpublished manuscript?
The skill can read and write files, runs scripts and uses the network, and sub-agents may be network-connected. Check where your content goes before uploading confidential or unpublished work.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use story-import if you have a partial or completed Chinese web novel and want an AI agent to rebuild it as an organized writing project you can keep building on. Skip it if you only need a one-off analysis of a book, or if you cannot yet confirm where your unpublished manuscript will be processed.
Download the skill