Fix mouse look swallowing by the HUD

Move guard look handling to the _input stage, recapture the cursor on
window-focus return and when clicking back into gameplay, and make HUD controls
mouse-transparent. Add regression tests for yaw/pitch and HUD mouse filters.
Verified under Xvfb with real relative mouse motion. Headless 237/237, Python
94/94, movement smoke 60 fps.

MEMORY.md

@@ -42,6 +42,7 @@ Session memory for durable decisions, state, and gotchas. Keep concise, remove o
 - Shift length is the scenario's `shift_duration_s` (900 s in all three venues). Tests and the screenshot tool override elapsed time; they do not change the default.
 - RNG: NPC reactions and event scheduling are seedable for deterministic tests.
 
+- **Mouse look (v0.7.3):** guard look is handled in `GuardController._input()` (before GUI), not `_unhandled_input`; focus/click-back recaptures the cursor and the HUD is made mouse-transparent recursively. Regression tests cover yaw/pitch and HUD filters.
 - **Tutorial foundation (0.8.0 work):** `docs/tutorial.md` is the eight-beat design; `src/game/tutorial.gd` is a pure `RefCounted` state machine with skip/restart and tests. It is not yet wired into `GameManager`; when wiring it, add an explicit QA bypass because captures/smokes would otherwise see prompts.
 - **Documentation ownership (2026-11-01):** `docs/architecture.md` §12 defines the source-of-truth map and consolidation rules. Obsolete/historical plan documents are removed rather than kept; current work belongs in `TODO.md`. README/development-status/Blender AO/Forward+ stale claims were corrected. The v1.0.0 gap register and 0.8/0.9/0.9.5/1.0 roadmap live in `TODO.md`.
 - **Vision judge (v0.7.2):** `tools/openrouter_judge.py` reads `/api/v1/models`, filters image-input models with `response_format`, and ranks by advertised prompt/image/completion prices. `--model auto` (default) uses the cheapest (`z-ai/glm-5.3-flash` on 2026-11-01) with fallbacks; explicit `--model`/`GIG_LINE_JUDGE_MODEL` wins. Auth errors abort, provider/model errors fall through. Full screen judge 13/13 match, 0 diff; A/B judge ~$0.006.