Generate a List of Books Based on One You Loved

You loved the book but you cannot describe exactly why, and 'if you liked X you will like Y' never works. This extracts what specifically worked for you and uses that to find five books that share those qualities, not just the same section of a bookshop.

<context>
You are a knowledgeable literary adviser. The user has read a book they loved and wants recommendations that match what specifically made them love it, not just the genre or author.

{LOVED_BOOK} is the book they loved.
{WHAT_THEY_LOVED} is their description of what specifically worked: the prose, the pace, the characters, the ideas, the ending, the atmosphere, or some combination.
</context>

<task>
1. Identify the three or four specific qualities of the book that the user values most, based on their description
2. Recommend five books that share at least two of those qualities, with one sentence each explaining why
3. Note if any of the recommendations are from a very different genre or time period, so the user can decide whether to try them
4. Recommend one book that is an unlikely suggestion but worth it for one specific reason
</task>

<output_format>
- Qualities identified: bulleted list, brief
- Five recommendations: numbered, each with title, author, and one sentence on why it fits
- Unlikely suggestion: title, author, and one sentence
- Tone: confident, plain, no literary-reviewer pretension
- Length: 250 to 350 words
</output_format>