Book Companion: AI Reading Partner

Problem

Many local AI chats still behave like support tools: a user wants to discuss a book, but gets summaries, recommendations, or service-desk phrasing. Book Companion reframes the conversation as reading companionship. It reads user_profile.md, the current book, and prior discussion, then responds with one of three personalities and updates reading preferences, mood, and life events after each session.

How It Works

The skill follows a practical loop:
- Initial setup: choose personality, name, genre preferences, current reading, and secret code, then run scripts/start.sh to initialize the local knowledge base.
- Intent routing: chat, co-reading, daydreaming, listening, help, personalization, and voice requests map to different modes described in conversation-modes.md, personalities.md, and voice-config.md.
- Memory update: writes to user_profile.md and discussions.md, while scripts/anniversary.py, scripts/daily_snippet.py, and scripts/mood_tracker.py handle milestones, nightly snippets, and mood logs.
- Intake: accepts text, .txt/.md files, photo OCR, Douban links, or book-title mentions, with a bias toward natural dialogue instead of listy output.

Boundaries

It is not a general customer-support agent or an automatic book recommender. Quality depends on local files, scripts, and optional voice configuration. It fits scenarios where a stable persona, reading memory, and light ritual matter more than task automation.