feat: implement full-platform STT API (v2.3 consensus plan)

Batch+realtime transcription API: faster-whisper engine w/ EngineOwner
(single GPU owner, OOM downgrade chain, persisted attempted_profiles),
hardware-adaptive device manager (T0-T3 VRAM tiers), Redis/in-proc job
queue w/ leases + crash recovery, postprocess (glossary/rules/LLM w/
egress guard), privacy-first result store (UUID keys, source deleted
after transcribe), retention sweeper, API-key auth (HMAC digests,
scopes, job ownership), WebSocket realtime lane (LocalAgreement),
CLI (detect/transcribe/bench/serve/key), Docker, benchmark runner.

127 mock-based tests pass; ruff clean. Includes verification checklist
and autoplan review notes.

benchmarks/manifest.yaml.example

@@ -0,0 +1,20 @@
+dataset_version: "1.0"
+hotword_sets:
+  none: []
+  domain: [API, vLLM, FastAPI, Kubernetes, LLM, GPU]
+clips:
+  - id: ko-en-001
+    audio_path: /secure/local/path/ko-en-001.wav
+    reference_path: references/ko-en-001.txt
+    language: ko
+    duration_sec: 30.0
+    entities:
+      - canonical: API
+        surface: API
+        start_char: 3
+        end_char: 6
+      - canonical: vLLM
+        surface: vLLM
+        start_char: 12
+        end_char: 16
+    tags: [clean, single-speaker]