openclawd/config
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Add library/genres commands and recommendation skills

Browse source 
- bin/audiobooks: add library [limit] [--full] and genres commands
- bin/jellyfin: add library [limit] [--full] and genres commands
- skills/book-recommender: new skill for audiobook recommendations
- skills/media-recommender: new skill for movie/show recommendations
This commit is contained in:
Hoid 2026-02-09 22:00:57 +00:00
parent 503f88820d
commit df0693f01d
4 changed files with 112 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

16
skills/book-recommender/SKILL.md Normal file
View file

@ -0,0 +1,16 @@
---
name: book-recommender
description: Recommend books and audiobooks based on the user's listening history and preferences. Use when the user asks for book recommendations, what to read/listen to next, or wants suggestions similar to books they enjoyed. Analyzes the user's Audiobookshelf library to understand taste.
---
## How to Recommend Books
1. Run `bin/audiobooks library` to get the user's full library with genres and metadata
2. Run `bin/audiobooks current` and `bin/audiobooks finished` to see what's in progress / recently finished
3. Analyze reading patterns: favorite authors, genres, series preferences, German vs English
4. Use `web_search` to find books similar to their favorites if needed
5. Recommend 3-5 books with explanations of WHY they'd like each one based on their library
6. Include: title, author, brief pitch, which books in their library it's similar to
7. Check if a German audiobook version exists (user prefers German audiobooks)
8. Note: user has ADHD — keep recommendations punchy, not walls of text
9. Format for WhatsApp (no markdown tables, use bold + bullet lists)

15
skills/media-recommender/SKILL.md Normal file
View file

@ -0,0 +1,15 @@
---
name: media-recommender
description: Recommend movies and TV shows based on the user's watch history and preferences. Use when the user asks for movie/show recommendations, what to watch next, or wants suggestions. Analyzes the user's Jellyfin library and watch history to understand taste.
---
## How to Recommend Movies & Shows
1. Run `bin/jellyfin library` to get the full library with genres and ratings
2. Run `bin/jellyfin watched` and `bin/jellyfin resume` to see recent activity
3. Analyze watch patterns: favorite genres, preferred types (movies vs shows), completion rates
4. Use `web_search` to find similar content if needed
5. Recommend 3-5 items with explanations of WHY based on their watch history
6. Check availability (is it on streaming services they might have? on Jellyfin already?)
7. Note: user has ADHD — keep recommendations punchy
8. Format for WhatsApp (no markdown tables)