What problem does it solve? Scattered light novel ebook files (epub) often have unreliable filenames, missing metadata, and unknown series status, making a personal library impossible to maintain. This Skill identifies each volume by cross-checking the Ranobe no Mori index, colophon data, RanobeDB, and BookWalker, then records verified bibliographic data (label, author, illustrator, release date, ISBN) in an append-only ledger and places each file on a flat, status-organized shelf. ## Core Features & Use Cases - Bibliographic identification: Matches each epub against multiple sources with strict evidence rules (ISBN is definitive; running titles, colophons, and OPF metadata ranked by reliability), and never guesses when evidence conflicts. - Append-only ledger and safe placement: Records decisions in an append-only ledger, then plans and applies file moves/renames through three safety gates with full undo support; nothing is ever deleted, only moved or quarantined. - Collection maintenance: Audits the shelf for duplicate names, missing tags, vanished files, and misidentified bonus booklets; detects duplicate volumes, finds bonus-item parent volumes, and generates placeholder markers for missing volumes. - Use Case: Point the Skill at an inbox folder of 20-30 unsorted epub files; it surveys each file's contents, presents evidence for your judgment, records your decisions, and shelves the batch under names like [GA文庫][小林湖底][りいちゅ][251213] ひきこまり吸血姫の悶々13.epub. ## Quick Start Ask the AI to run python scripts/cli.py status to confirm the shelf, inbox, and working directories, then run python scripts/cli.py materials --count 20 to begin identifying the first batch of unsorted light novel epubs.
