mirror: patches 0138-0144 (backfill c5 gap + c6) + full docs re-sync

c5: gpu-share busy accounting via llama_synchronize (bug-2244) — pa4 + MTP gates pass; patch 0141 recaptured

docs+patches: gpu-share (0141) — USAGE.md, patch series, design doc

bench: defaults — parallel 8, tps-floor 15 scheduler-driven, stall timeout auto-scales with steps (bug-2222)

docs: package_courier harness = reference LangGraph implementation

c4: package_courier doc

c4: polykv_api doc

Upload folder using huggingface_hub

Upload docs/features/fused_nextn_mtp.md with huggingface_hub

Upload docs/features/rys_layer_duplication.md with huggingface_hub

Upload folder using huggingface_hub