fix: address gadfly P1 review (3 low-risk findings)

Triaged gadfly's P1 review (advisory). Fixed the three clearly-correct,
low-risk items; the rest were pre-existing mort behavior or theoretical:

- model/call.go: recordUsage dropped fully-cached responses (input==0 &&
  output==0 early-return missed CacheRead/CacheWrite-only usage, which
  Anthropic/OpenAI prompt-caching bills). Guard now also checks cache tokens.
- llmmeta/helper.go: recordLedger swallowed Storage.RecordMetaCall errors;
  now logs them (slog.Warn) so a non-logging Storage impl can't silently drop
  audit rows.
- model/cloud_sync.go: the ollama.com limit-cache used unbounded io.ReadAll;
  wrapped both reads in io.LimitReader(1 MiB) so a misbehaving endpoint can't
  exhaust memory before the 15s timeout.

Noted-not-fixed (follow-ups / pre-existing mort semantics): tier_not_allowed
ledger label on resolution failure, unknown-model usage attribution, the
cloud_sync https scheme allowlist, and several theoretical/cosmetic items.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

llmmeta/helper.go

@@ -34,6 +34,7 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
+	"log/slog"
 	"strings"
 	"time"
 
@@ -574,7 +575,9 @@ func (h *Helper) recordLedger(ctx context.Context, call MetaCall) {
 	if h.storage == nil {
 		return
 	}
-	_ = h.storage.RecordMetaCall(ctx, call)
+	if err := h.storage.RecordMetaCall(ctx, call); err != nil {
+		slog.Warn("llmmeta: failed to record ledger row", "err", err)
+	}
 }
 
 // tryParseJSON attempts to decode text as JSON. Returns the parsed