fix(ci): tear down legacy capakraken compose project before Docker Deploy

After the Phase 3 rename the project name flipped from 'capakraken' to 'nexus'.
The QNAP runner may still have capakraken-redis-1 running (holding port 6380).
The down step only cleaned up the 'nexus' project, leaving the old container
alive and causing "Bind for 0.0.0.0:6380 failed: port is already allocated".

Add an explicit `docker compose -p capakraken ... down` before the normal
cleanup so stale pre-rename containers are always removed first.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

.github/workflows/ci.yml

@@ -477,7 +477,11 @@ jobs:
         # runs. A previous run's failed migration entry in _prisma_migrations
         # causes P3009 on the next migrate deploy; wipe volumes for a truly
         # fresh deploy test every time.
-        run: docker compose -f docker-compose.yml -f docker-compose.ci.yml down -v --remove-orphans || true
+        # Also tear down the legacy "capakraken" project (pre-Phase-3 rename)
+        # in case old containers are still holding host ports 5433/6380.
+        run: |
+          docker compose -p capakraken -f docker-compose.yml -f docker-compose.ci.yml down -v --remove-orphans || true
+          docker compose -f docker-compose.yml -f docker-compose.ci.yml down -v --remove-orphans || true
 
       - name: Start infrastructure (postgres + redis)
         run: docker compose -f docker-compose.yml -f docker-compose.ci.yml up -d postgres redis