story-import takes a novel an author has already written, as a text file, a folder of chapters or pasted text, and rebuilds it as a writing project. It confirms title, genre, platform, length and whether the last chapter is complete, runs a deep analysis, then migrates the result into settings, outline, chapter and tracking folders.
The project it leaves behind works with the long-form and short-form writing skills of the same package, so the author can say “write chapter N” and carry on. Open questions, such as where volumes split, are listed for the author to decide.
What You Get
- A project folder named after the book with settings, outline, chapter text and tracking files
- Reverse-written chapter outlines for a long novel, built batch by batch from the existing text
- A tracking state initialized from the manuscript, with foreshadowing and character status checked by a script
- An import record that keeps decisions, preferences and progress so you can resume in a new conversation
- A report listing what needs author confirmation, with the next step to continue writing
Common Use Cases
- Continuing a serial novel you already published for twenty or more chapters
- Turning a finished short story into a project that the short-form writing skill can build on
- Rescuing a half-written draft whose last chapter is unfinished, after you decide how to treat it
What Makes It Unique
Most writing skills start from a blank page. This one starts from the text you already have and reuses the package’s own analysis pipelines, so the imported book lands in exactly the folder structure the writing skills expect. It keeps the imported book separate from any outside benchmark book, never registers your own work as its reference, and leaves story decisions such as volume splits to you.
Under The Hood
The skill is mostly instructions that route the agent through phases, with reference files read as each phase starts. It reads your manuscript files, checks the project for a deployed marker and for sub-agent definitions, and hands the heavy reading to helper agents when the host supports them. It writes many files into the project: the import record, the analysis library, settings, outline, chapters and tracking. Bundled Python and Node scripts run locally to count words, check outline contracts, build outline briefs and initialize and verify the tracking state. No web requests are part of the flow.