Trim oversized structural-impossibility test loops

[stphung]

Summary

• Ran the test-audit skill (3 parallel Explore agents across core/combat, systems, and items/achievements/misc test scopes)
• Found several tests looping 200–1000x to check invariants that hold by construction (guard conditions, structural range/membership bounds) rather than probabilistically — capped these at the project's documented 100-iteration norm for structural-impossibility checks
• Removed a stale comment in tick_integration_test.rs claiming spawn_enemy_if_needed/apply_tick_xp use thread_rng() internally; verified against source that both are already fully deterministic
• Other findings (duplicate test names, helper-function duplication across files, redundant coverage between files, a 20k-iteration full-combat-sim loop) were flagged as needing human judgment rather than auto-fixed, per the audit skill's policy — not included in this PR
• ~40-occurrence known brute-force-RNG-search-loop debt (tracked since 2026-07-02) intentionally left untouched, per skill guidance

Test plan

• cargo test --test game_tick_tests --test stormglass_tests --test fishing_tests --test deep_tests — all green
• make check (fmt, clippy, full test suite, progression check, security audit) — all green
• cargo test run 10x consecutively — 0 failures across all runs

---

Generated by Claude Code