The LlmService interface no longer exposes isDailyLimitReached — model
switching is handled internally. Remove all checks in thoughtSystem,
inventionSystem, narrationService, and thoughtGenerator. Update mocks
in test files to match the new interface and remove obsolete daily-limit
test cases.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Events recorded with tick: 0 now get a meaningful tick value from the
service's internal counter, ensuring selectForPrompt recency scoring
works correctly. Also fixed preferTypes to accept readonly arrays.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace the simple last-2-narration-events approach with smart selection
of up to 5 memory events, preferring types relevant to the thought trigger
(social events for post-interaction, need events for need-critical).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- determineTrigger now detects post_interaction (recent social outcome),
need_critical, idle (wandering), and periodic triggers
- Track active emoji per NPC, destroy previous before showing new one
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
root