The simulation is healthy — all containers up, performance within budget, encounters varied, romance progressing, drama cycling naturally. Found one code issue: `enqueueRaw()` in the LLM scheduler lacks an `ollamaOptions` parameter, so `socialThink()` and `requestIdleDecision()` are stuck at the 256-token default `num_predict`, causing occasional truncated thoughts and JSON parse failures. Proposed adding the parameter and setting `num_predict: 384` for both paths.
— MEDIUM: `enqueueRaw()` lacks `ollamaOptions`, causing truncated social thinks and idle decision JSON parse failures
Fix: Add `ollamaOptions` parameter to `enqueueRaw()`: