İçeriğe geç
TakoAPI
Becerilere dön

calibre-catalog-read

nextaltair tarafından

Web & Frontend Development
918İndirme sayıları clawskills.sh tarafından bildirilir
0Görüntülenmeler
0 beğeniBu beceriyi beğenmek için giriş yapın
—Henüz değerlendirilmedi
Puanınız

Özet

Read Calibre catalog data via calibredb over a Content server, and run one-book analysis workflow that writes HTML.

Açıklama

Bu Beceri Ne Yapar

Reads book catalog data from a Calibre Content server using calibredb, supporting list, search, and ID lookups. Runs a one-book AI analysis pipeline that extracts text, generates a structured analysis via a subagent, and writes the result as an HTML block back to the book's comments field. Analysis state is cached in SQLite to avoid reprocessing unchanged books.

The two-turn async design and SQLite cache let long analysis jobs run without blocking the chat session or reprocessing books that have not changed.

Ne Zaman Kullanılır

  • Searching a remote Calibre library for books in a specific series
  • Looking up full metadata for a single book by its Calibre ID
  • Running AI analysis on a novel and saving the result to its Calibre comments
  • Checking whether a book has already been analyzed before triggering a new pipeline run
  • Listing all books from a Calibre Content server in structured JSON

Örnek İş Akışı

GİRİŞ

User asks: analyze book ID 3 from my Calibre library in Japanese

AJAN
  1. 1

    Validate book ID and confirm subagent settings (model, thinking level, timeout) for the session

  2. 2

    Run prepare_subagent_input.mjs to extract text and split it into chunked source files

  3. 3

    Use subagent-spawn-command-builder to generate the sessions_spawn payload, then call sessions_spawn

  4. 4

    Register the run in runs.json with runId, book_id, title, and status running, then reply to user that analysis is running in background

  5. 5

    On completion event, call handle_completion.mjs to load the analysis JSON, apply HTML to Calibre comments via calibredb, and remove the run record

ÇIKIŞ

Structured AI analysis written to the book's Calibre comments field as HTML, with cache record stored in calibre_analysis.sqlite

Gereksinimler

  • CALIBRE_PASSWORD environment variable
  • CALIBRE_USERNAME environment variable (optional, for authenticated servers)
  • Calibre application installed with calibredb and ebook-convert on PATH
  • Running Calibre Content server accessible at a known host and port with library ID

Kurulum

clawhub install nextaltair/calibre-catalog-read

CLI komutunu çalıştırın ya da istemi Claude Code, Cursor, Codex veya kullandığınız kodlama ajanına yapıştırın.

Bağlantılar