feat(game): Option A auto-critical, discretionary proposals, remove CRITICAL_FAILURE

(train) pre-test train

training loop v1

Merge branch 'main' into feature/llm-adapters

Modularize LLM integration and restructure parliamentary adapters

Merge main and fix benchmark adapter review issues

Fix baseline adapter voting validity

fix(server): repair OpenEnv contract tests and environment reset

Align LLM adapters with environment validation

Implement baseline benchmark evidence pipeline

Implement LLM policy adapters

Tighten parliament action validation

Implement parliamentary action resolution interface

Merge pull request #4 from Kr0issant/event-reward-fixtures

fix(core): align treasury settlement semantics

fix(telemetry): harden rollout metrics logging

test(core): add deterministic economy reward fixtures

feat(telemetry): add rollout event logging

feat(core): implement simulation primitives

chore: scaffold RL adapter pipeline