fix(sqlite): respect --lock-timeout flag in direct mode (bd-2zd.4)

The ensureStoreActive() function was ignoring the user-configured
--lock-timeout flag and always using the 30s default via sqlite.New().

This fix changes ensureStoreActive() to use sqlite.NewWithTimeout()
with the configured lockTimeout, allowing users to specify shorter
timeouts for multi-agent scenarios.

With this change, users can now run:
  bd --lock-timeout=500ms sync

to fail fast if the database is locked, rather than waiting 30s.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

cmd/bd/direct_mode.go

@@ -89,7 +89,9 @@ func ensureStoreActive() error {
 		}
 	}
 
-	sqlStore, err := sqlite.New(getRootContext(), path)
+	// Use configured lock timeout (default 30s, can be overridden via --lock-timeout)
+	// This fixes bd-2zd.4: SQLite lock contention during multi-agent operations
+	sqlStore, err := sqlite.NewWithTimeout(getRootContext(), path, lockTimeout)
 	if err != nil {
 		// Check for fresh clone scenario
 		if isFreshCloneError(err) {