diff --git a/scripts/check-backup-freshness.sh b/scripts/check-backup-freshness.sh
index 820ff8f..c0c48bf 100755
--- a/scripts/check-backup-freshness.sh
+++ b/scripts/check-backup-freshness.sh
@@ -232,10 +232,17 @@ for want in $EXPECTED_NH3; do
 done
 
 # --- Layer: PBS VM images (newest per guest, all namespaces) ---
+# ⚠ The age comes from the snapshot's NAME (its backup time, e.g. 2026-10-04T10:00:04Z),
+# never from the directory's mtime (2026-10-06). PBS rewrites index.json.blob inside the
+# snapshot dir at every verify (first one at the next 02:30, then every 30 days, from
+# verification.cfg outdated-after 30) and on notes or protect edits, which moves the
+# dir mtime forward. Measured: ana-pve/vm/101's 10-04 snapshot read "2026-10-06" here.
+# A guest whose backups had stopped would alarm up to a day late, then go quiet again
+# for ~2 days after each monthly re-verify.
 pbs=$($SSH pbs-ana 'for ns in /mnt/pbs-datastore/ns/*/; do n=$(basename "$ns")
   for d in vm ct; do for g in "$ns$d"/*/; do [ -d "$g" ] || continue
     nb=$(ls -d "$g"20*T* 2>/dev/null | sort | tail -1)
-    [ -n "$nb" ] && echo "$n/$d/$(basename "$g") $(stat -c %Y "$nb")"
+    [ -n "$nb" ] && echo "$n/$d/$(basename "$g") $(date -d "$(basename "$nb")" +%s)"
   done; done; done' 2>/dev/null)
 if [ -z "$pbs" ]; then errors+=("PBS-ANA: unreachable or no snapshots"); else
   for ns in "${!NS_NODE[@]}"; do load_jobs "$ns" || errors+=("coverage for namespace '$ns' UNREADABLE — nothing in it will be suppressed"); done