Add htn to v1.0/v1.1 baselines, fix stale fp401-t3a note, add box_threshold and tier4 investigation addenda

Add v1.1 (9 scenes, tier4_multi_step, qwen2.5:3b comparison) alongside frozen v1.0

Cross-reference the live milo_benchmark_companion Space

Add perception-grounded tier1_locate finding: perceived_by_agent diverges from exists_in_scene (real sim-to-real detection gap, not fixed)

Flatten tasks.json to 25 top-level rows; fix broken origin-repo links in dataset card

Publish milo_benchmark v1.0: 25-task/5-scene/3-tier dataset with real 3-planner baseline