feat(recipes): add manual checkpoint reconnect for broken recipe entries

Checkpoint entries that cannot be restored by download (deleted,
unresolvable hash, or name-only remnants with no CivitAI identifiers)
now get the same remediation chain LoRAs already had:

- scanner: parameterized reconnect-suggestion ranking, update/restore/
  set-hash-invalid for the checkpoint entry, and clear hashInvalid on
  rematch write-back (was only done for LoRAs)
- persistence/handlers/routes: reconnect/restore/reconnect-suggestions/
  mark-hash-invalid endpoints under /api/lm/recipe/checkpoint/*
- modal: checkpoint reconnect UI (deleted/hash-invalid badges, inline
  form with suggestions, undo for reconnected entries); download
  failures mark the hash invalid only on explicit unresolvable signals
  (not found/deleted/404/410), matching the LoRA rule
- css: checkpoint undo button shares the LoRA undo styles
- i18n: the 14 new keys translated in all 9 locales

static/css/components/recipe-modal.css

@@ -957,8 +957,10 @@
 }
 
 /* Restore icon for manually reconnected entries: its presence on the info
-   row doubles as the "was reconnected" marker. */
-.lora-undo-reconnect {
+   row doubles as the "was reconnected" marker. Shared by LoRA and
+   checkpoint entries, which use the same info-row flex layout. */
+.lora-undo-reconnect,
+.checkpoint-undo-reconnect {
     margin-left: auto;
     background: none;
     border: none;
@@ -973,7 +975,9 @@
 }
 
 .lora-undo-reconnect:hover,
-.lora-undo-reconnect:focus-visible {
+.lora-undo-reconnect:focus-visible,
+.checkpoint-undo-reconnect:hover,
+.checkpoint-undo-reconnect:focus-visible {
     opacity: 1;
     color: var(--lora-accent);
     background: var(--lora-surface);