storage SAT: split collect/self-test modes, add per-disk text reports

Check mode: read-only SMART/NVMe data collection, no self-test.
Load mode: same collection + short self-test (nvme device-self-test -s 1,
smartctl -t short). Card descriptions updated accordingly.

After each storage SAT run, a disk-N-devname-report.txt is written
per device into the runDir (auto-included in support bundles).
Web UI task page renders one card per disk directly below Task Report.

Also fixes pre-existing TestDashboardRendersRuntimeHealthTable failure:
test fixture used "inactive" status but code now treats inactive as OK
for completed oneshot services; updated to "failed" to match intent.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

audit/internal/webui/server_test.go

@@ -1227,7 +1227,7 @@ func TestDashboardRendersRuntimeHealthTable(t *testing.T) {
   ],
   "services":[
     {"name":"bee-web","status":"active"},
-    {"name":"bee-nvidia","status":"inactive"}
+    {"name":"bee-nvidia","status":"failed"}
   ]
 }`
 	if err := os.WriteFile(filepath.Join(exportDir, "runtime-health.json"), []byte(health), 0644); err != nil {
@@ -1281,7 +1281,7 @@ func TestDashboardRendersRuntimeHealthTable(t *testing.T) {
 		`Bee Services`,
 		`CUDA runtime is not ready for GPU SAT.`,
 		`Missing: nvidia-smi`,
-		`bee-nvidia=inactive`,
+		`bee-nvidia=failed`,
 		// Hardware Summary card — component health badges
 		`Hardware Summary`,
 		`>CPU<`,