1. Collect book file path, target language, concurrency, and custom instructions.
2. Convert the book into Markdown chunks using a conversion script.
3. Discover which chunks need translation by comparing source and output files.
4. Launch parallel sub-agents to translate each chunk independently.
5. Verify the completeness of the translation and retry any failed chunks.
6. Translate the book title to the target language.
7. Merge the translated chunks and build the final output in various formats.
8. Report the results to the user.