fix(simulation): address PR #37 re-review — stale strings, test coverage, confrontation symmetry

Hoshe re-review (3 items):
- content_scaling.rs:185: doc "StableId 0-51" → references constant
- content_scaling.rs:256: assertion message "id<=51" → "id <= max_baseline_id"
- input.rs: teleport test now asserts WalkAwayRequest + ConfrontationDelivered
  are cleared (was only checking TalkRequest + ActiveDialogue)

Tyre re-review (2 items):
- input.rs: same teleport test coverage (overlaps Hoshe #3)
- dialogue.rs: process_confrontation_response now inserts RoutineDeviation
  with DeviationTrigger::Confrontation — symmetric with walk-away path.
  Test updated to verify deviation is recorded.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

server/tests/content_scaling.rs

@@ -182,8 +182,9 @@ fn scaling_tick_timing_within_budget() {
 }
 
 /// Determinism test: baseline entities produce identical snapshots regardless
-/// of extra NPCs being present. The original Gauntlet entities (StableId 0-51)
-/// should have the same positions and visibility after the same number of ticks.
+/// of extra NPCs being present. The original Gauntlet entities (StableId 0
+/// through RESET_PLATE_STABLE_IDS.1) should have the same positions and
+/// visibility after the same number of ticks.
 #[test]
 #[cfg(feature = "gauntlet")]
 fn extra_npcs_dont_affect_baseline_behavior() {
@@ -253,6 +254,6 @@ fn extra_npcs_dont_affect_baseline_behavior() {
 
     assert_eq!(
         baseline_original_ids, scaled_original_ids,
-        "Original Gauntlet entities (id<=51) should be identical in both runs"
+        "Original Gauntlet entities (id <= max_baseline_id) should be identical in both runs"
     );
 }