feat(training): add airborne_penalty and a stage-6 "unmask" curriculum run

Stage 5 (aggression) passed (41-47 vs grounded curric-s2-defend, within
the lenient gate but not yet a clear win). Rather than keep the locomotion
mask on indefinitely, stage 6 reopens full 3D controls on top of the
aggression retune and pairs it with a new dense airborne_penalty (scaled
by height above the floor) so the policy learns to prefer staying grounded
through incentives instead of a hard mask — same regime shift that
regressed stage 3, but this time with a mitigation and ~12x the training
time (~240M timesteps / ~24h vs ~20M / ~2h) to actually re-converge
instead of stalling mid-shift.

airborne_penalty follows the existing SHIP_AI_OVERRIDES pattern: default
0 (off) on ship_ai_controller.gd, exposed via train.py's new
--airborne-penalty flag, added to training_mode.gd's allow-list. Also adds
a per-stage timesteps override in curriculum.py (STAGES[n]["timesteps"])
since this is the first stage to need a different budget than the rest.

training/train.py

@@ -96,6 +96,10 @@ def parse_args():
         "--ball-touch-reward", type=float, default=None,
         help="Overrides ShipAIController.ball_touch_reward (event reward on ball contact, cooldown-gated)",
     )
+    curriculum.add_argument(
+        "--airborne-penalty", type=float, default=None,
+        help="Overrides ShipAIController.airborne_penalty (dense per-tick cost scaled by height above the floor)",
+    )
 
     return parser.parse_args()
 
@@ -116,6 +120,7 @@ def _curriculum_kwargs(args) -> dict:
         "ai_velocity_to_ball_weight": args.velocity_to_ball_weight,
         "ai_ball_distance_penalty": args.ball_distance_penalty,
         "ai_ball_touch_reward": args.ball_touch_reward,
+        "ai_airborne_penalty": args.airborne_penalty,
     }
     return {key: value for key, value in mapping.items() if value is not None}