refactor: harden diagnostics and consolidate runtime code

audit/internal/app/component_status_db_test.go

@@ -3,6 +3,7 @@ package app
 import (
 	"os"
 	"path/filepath"
+	"strings"
 	"testing"
 
 	"bee/audit/internal/schema"
@@ -145,6 +146,21 @@ func TestSATFailureDetailFallsBackToFailedSubJobs(t *testing.T) {
 	}
 }
 
+func TestSATFailureDetailIncludesValidatorDetail(t *testing.T) {
+	runDir := t.TempDir()
+	summary := "2-all-reduce-perf_rc=0\n" +
+		"2-all-reduce-perf_status=FAILED\n" +
+		"2-all-reduce-perf_detail=NVLink state does not match selected topology: GPU0<->GPU1\n" +
+		"overall_status=FAILED\n"
+	if err := os.WriteFile(filepath.Join(runDir, "summary.txt"), []byte(summary), 0644); err != nil {
+		t.Fatal(err)
+	}
+	got := SATFailureDetail(runDir)
+	if !strings.Contains(got, "NVLink state does not match selected topology") {
+		t.Fatalf("SATFailureDetail() = %q, want validator detail", got)
+	}
+}
+
 // TestSATFailureDetailEmptyWhenNoReasonFound guards the ultimate fallback:
 // when summary.txt carries no identifiable per-job or warnings detail (e.g.
 // unreadable or from an older binary version), callers must get "" so they
@@ -233,6 +249,21 @@ func TestApplySATResultToDBNvidiaConfigSharesGPUKeyWithOtherNvidiaTargets(t *tes
 	}
 }
 
+func TestApplySATResultToDBNvidiaInterconnectFailureReachesGPUComponent(t *testing.T) {
+	db, err := OpenComponentStatusDB(filepath.Join(t.TempDir(), "component-status.json"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	ApplySATResultToDB(db, "nvidia-interconnect", writeSATSummary(t, "FAILED"))
+	rec, ok := db.Get("pcie:gpu:nvidia")
+	if !ok {
+		t.Fatal("nvidia-interconnect failure wrote no pcie:gpu:nvidia record")
+	}
+	if rec.Status != "Warning" {
+		t.Fatalf("status=%q want Warning", rec.Status)
+	}
+}
+
 // TestRecordDeduplicatesRepeatedIdenticalStatusFromSameSource guards the
 // hardware-ingest-contract.md rule that status_history is a transition log
 // ("История переходов статусов"), not a per-poll journal. A component