story-publish gathers what a platform’s book-creation form asks for and saves it as a publishing file in the book folder. It currently supports Fanqie long novels and short stories, checks each value against the platform reference file, and marks every field as confirmed, suggested, to confirm or to verify.
It never logs in, fills the form or publishes. Fields the author has not decided stay empty and appear in a to-confirm list, and the skill does not invent characters, plot points or selling points.
What You Get
- A publishing file for the target platform with a fixed, machine-readable layout
- Candidate titles with character counts checked against the platform rules
- Category and tag choices limited to the options in the platform reference
- A synopsis for long novels, and a cover section holding the prompt and any saved image path
- To-confirm and to-verify lists that show exactly what is left to decide
Common Use Cases
- Filling in a Fanqie book-creation form after reaching about twenty thousand words
- Getting title and tag candidates checked against platform limits
- Preparing a cover prompt for a short story before publishing
What Makes It Unique
The output file is a fixed contract that a program can read later, and each field carries a status so nothing a model guessed is passed off as decided. Platform candidates come only from the skill’s own reference file with a verification date, and unsupported platforms get a platform-neutral draft that states it is not adapted.
Under The Hood
The skill reads your book’s settings, outline and character files and a platform reference file, then writes a publishing file into the book folder. It runs one local Python script that only reads the file and reports failures and warnings. For covers it hands off to the cover skill, which can call an image generation tool if your session has one. It does not browse the web, log in or submit anything to the platform.