- Fix swipe data retrieval to check both message.extra and swipe_info sources
- Fix user_message_end position to inject into preceding (not next) user message
- Add ordered content-matching for text completion prompt injection
- Add ordered content-matching for chat completion prompt injection
- Remove unnecessary HTML entity normalization
- Clean up unused imports and variables
- Use persistent event listeners instead of once() to inject into ALL generations
- Don't clear context map on GENERATION_ENDED so prewarm gets the same context
- Remove unused onGenerationEndedCleanup function
This prevents any risk of injected context being accidentally saved to the chat.
Instead of modifying chat[].mes directly, we now:
1. Build a context map during GENERATION_STARTED
2. Inject into the prompt string (GENERATE_AFTER_COMBINE_PROMPTS) for text completion
3. Inject into the message array (CHAT_COMPLETION_PROMPT_READY) for chat completion
The original chat messages are never modified.
Features:
- Complete tracker configuration UI with add/remove functionality
- User Stats: Custom stats, status fields, skills section
- Info Box: Configurable widgets (date, weather, temp, time, location, events)
- Present Characters: Custom fields, relationships, character stats, thoughts
- Character-specific stats with color interpolation
- New multi-line format for cleaner AI generation and parsing
- Auto-cleanup of placeholder brackets in AI responses
- Relationship badges with emoji mapping
- Advanced inventory v2 system with multi-location storage
- Responsive mobile support with horizontal scrolling
- Removed legacy format support for cleaner codebase
- Fixed context injection for together mode (no duplication)
- Updated README with new features and configuration guide
Folds in changes from commits 6d97992, cc53c69, 6987c0c:
- Update stat label/value font sizes (clamp 0.9-1vw → 0.5-0.7vw)
- Update inventory items font size (clamp 0.4-0.6vw → 0.7-0.6vw)
- Center dice save button with auto margins
- Add debug console.log for committedTrackerData
- Comment out verbose console logs for plot progression and persona avatar
- Clarify prompt instruction to remove brackets
Extract AI generation and parsing logic from monolithic index.js into
modular architecture under src/systems/generation/.
**Modules Created:**
- promptBuilder.js (319 lines) - AI prompt generation functions
- parser.js (152 lines) - Response parsing and stats extraction
- apiClient.js (154 lines) - Separate mode API call handler
- injector.js (216 lines) - Prompt injection for both modes
**Changes:**
- All functions preserve exact behavior from original
- Import paths calculated for browser module resolution
- Zero functionality changes, pure code organization
Reduces index.js by ~700 lines when combined with function removal
(to be committed separately).
Spicy_Marinara
tomt610
Subarashimo
Andy Mauragis
Lucas 'Paperboy' Rose-Winters