refactor(rmi-backend,audit): mount Wave 3 + archive 136 dead-code files (P2.3)

PHASE 2.3 (AUDIT-2026-Q3.md):

Task 1 — Wire-in Wave 3 (1 router mounted, 2 deferred):
  - app.routers.unified_scanner_router mounted at /api/v2/scanner/* (2 routes:
    POST /api/v2/scanner/token/scan, POST /api/v2/scanner/wallet/scan).
    Refactored prefix from /api/v2 -> /api/v2/scanner to avoid future conflicts
    with the v1 /api/v1/scanner/ stub.
  - app.routers.unified_wallet_scanner DEFERRED (no router APIRouter attribute;
    library module consumed by unified_scanner_router via get_wallet_scanner()).
  - app.routers.admin_extensions DEFERRED (DORMANT per audit; 25 routes at
    /api/v1/admin/* would shadow /api/v1/admin/alerts_webhook).

Task 2 — Archive 136 dead-code files to app/_archive/legacy_2026_07/:
  - 73 routers in app/routers/ (reach graph showed zero reach into mount.py).
  - 63 flat app/*.py (domain modules never imported by live code).
  - 1 file RESTORED post-archive: app/routers/x402_bridge_health.py (caught by
    tests/unit/test_bridge_health.py which directly imports it; reach graph
    considered tests/ only as transitive reach — to be patched in next cycle).

Forced-LIVE (NOT archived per user directive):
  - app/ai_pipeline_v3.py  (3 importers in audit window, importers themselves DEAD)
  - app/splade_bm25.py       (LIVE via app.rag_service)
  - app/wallet_manager_v2.py (LIVE via x402_enforcement, x402_tools, sweep_all, sweep_now)
  - app/crypto_embeddings.py (NOT in audit ARCHIVE list; heavy import graph)

Verification (forward-import closure from mount.py + main.py + factory.py + lifespan.py):
  - imports = 348 app.* modules
  - reached = 194 files reachable from roots
  - archive set = audit_dead (186) - reached - forced_live (4) - test_live (1) = 136
  - Net delta: 136 files moved, 44,932 LOC reduction, 293->295 active routes (+2 from Wave 3)

pyproject.toml updates:
  - setuptools.packages.find: added exclude for app._archive*
  - ruff.extend-exclude: added "app/_archive/"
  - mypy.exclude: added "app/_archive/"

Smoke test: pytest tests/ — 817 passed, 3 pre-existing failures unchanged
(0 new failures; 0 routes lost; all 4 forced-LIVE files still importable).

Restoration: git mv app/_archive/legacy_2026_07/<name>.py <original-path>
and add the import to app/mount.py ROUTER_MODULES.

Refs: AUDIT-2026-Q3.md /home/dev/pry/rmi-final-deadcode-2026-07-06.md

app/_archive/legacy_2026_07/auto_healing.py

@@ -0,0 +1,84 @@
+"""Auto-Healing Infrastructure - self-diagnosing, self-repairing platform."""
+
+import os
+from datetime import UTC, datetime
+
+import httpx
+import psutil
+from fastapi import APIRouter
+
+router = APIRouter(prefix="/api/v1/health", tags=["auto-healing"])
+
+BACKEND = os.getenv("BACKEND_URL", "http://localhost:8000")
+CHECKS = [
+    ("redis", "http://localhost:6379", "Redis"),
+    ("postgres", "http://localhost:5432", "Postgres"),
+    ("ollama", "http://localhost:11434/api/tags", "Ollama"),
+    ("backend", f"{BACKEND}/health", "RMI Backend"),
+    ("clickhouse", "http://localhost:8123/ping", "ClickHouse"),
+    ("qdrant", "http://localhost:6333/health", "Qdrant"),
+]
+
+
+@router.get("/full")
+async def full_health_check():
+    """Comprehensive system health check - all services."""
+    results = {}
+    all_healthy = True
+
+    async with httpx.AsyncClient(timeout=5) as c:
+        for name, url, label in CHECKS:
+            try:
+                r = await c.get(url)
+                healthy = r.status_code < 500
+            except Exception:
+                healthy = False
+
+            results[name] = {"service": label, "healthy": healthy, "last_checked": datetime.now(UTC).isoformat()}
+            if not healthy:
+                all_healthy = False
+
+    # System resources
+    mem = psutil.virtual_memory()
+    disk = psutil.disk_usage("/")
+
+    return {
+        "overall": "HEALTHY" if all_healthy else "DEGRADED",
+        "timestamp": datetime.now(UTC).isoformat(),
+        "services": results,
+        "system": {
+            "cpu_pct": psutil.cpu_percent(interval=1),
+            "memory_pct": mem.percent,
+            "memory_available_gb": round(mem.available / (1024**3), 1),
+            "disk_pct": disk.percent,
+            "disk_free_gb": round(disk.free / (1024**3), 1),
+            "swap_pct": psutil.swap_memory().percent,
+        },
+        "alerts": _check_alerts(mem, disk),
+    }
+
+
+def _check_alerts(mem, disk) -> list:
+    """Generate alerts for concerning metrics."""
+    alerts = []
+    if mem.percent > 90:
+        alerts.append({"level": "CRITICAL", "msg": f"Memory usage at {mem.percent}% - consider restarting services"})
+    elif mem.percent > 80:
+        alerts.append({"level": "WARNING", "msg": f"Memory usage at {mem.percent}%"})
+    if disk.percent > 90:
+        alerts.append({"level": "CRITICAL", "msg": f"Disk usage at {disk.percent}% - prune logs immediately"})
+    elif disk.percent > 80:
+        alerts.append({"level": "WARNING", "msg": f"Disk usage at {disk.percent}%"})
+    return alerts
+
+
+@router.post("/heal/{service}")
+async def heal_service(service: str):
+    """Attempt to heal a failed service."""
+    if service == "backend":
+        os.system("docker restart rmi-backend")
+        return {"service": service, "action": "restarted", "note": "Backend restart initiated"}
+    if service == "redis":
+        os.system("docker restart rmi-redis")
+        return {"service": service, "action": "restarted"}
+    return {"service": service, "action": "unknown", "note": "Manual intervention required"}