fix(curate): guard _salvage_objects against None LLM content

The LLM (Opus score stage / Haiku cards stage) can return None or empty
content, causing _salvage_objects to crash with
'NoneType' object has no attribute 'find' — which killed the curation
pipeline and returned zero matches to the user.

- Guard _salvage_objects and _salvage_kept against None/empty input
- Add explicit empty-content checks in both stages with logging
- Log content[:200] on unparseable score stage for QA diagnostics

app/engine/curate.py

@@ -121,10 +121,12 @@ def _parse_json(text: str) -> dict:
     return json.loads(t)
 
 
-def _salvage_objects(text: str, *, after: str | None = None) -> list[dict]:
+def _salvage_objects(text: str | None, *, after: str | None = None) -> list[dict]:
     """Recover valid objects from malformed/truncated model JSON: brace-scan each top-level {...} and
     json.loads it on its own, skipping the one broken/cut-off object. `after` anchors the scan past a key
     (e.g. '"kept"'); omit it to scan the whole text (the cards stage uses '"cards"', not '"kept"')."""
+    if not text:
+        return []
     s = text
     if after:
         i = text.find(after)
@@ -145,9 +147,7 @@ def _salvage_objects(text: str, *, after: str | None = None) -> list[dict]:
                     pass
                 start = None
     return out
-
-
-def _salvage_kept(text: str) -> list[dict]:
+def _salvage_kept(text: str | None) -> list[dict]:
     """Score-stage salvage — recover kept-job objects (anchored after the '"kept"' key)."""
     return _salvage_objects(text, after='"kept"')
 
@@ -216,12 +216,15 @@ def _score_stage(prefs: dict, ctx: dict | None, sifted_jobs: list[dict], s) -> l
     except Exception as e:  # noqa: BLE001
         logger.warning("score stage (Opus) failed, falling back to sift: %s", e)
         return None
+    if not content or not content.strip():
+        logger.warning("score stage LLM returned empty content — falling back to sift")
+        return None
     try:
         kept = _parse_json(content).get("kept") or []
     except Exception:  # noqa: BLE001
         kept = _salvage_kept(content)
         if not kept:
-            logger.warning("score stage unparseable — falling back to sift")
+            logger.warning("score stage unparseable — falling back to sift (content[:200]=%r)", content[:200])
             return None
         logger.warning("score stage JSON malformed; salvaged %d items", len(kept))
     # Floor check on the RAW computed score (weak never shown), THEN the transparent presentation
@@ -256,6 +259,9 @@ def _cards_stage(prefs: dict, ctx: dict | None, scored: list[dict], by_id: dict,
     except Exception as e:  # noqa: BLE001 — prose is non-critical; degrade to templated cards
         logger.warning("cards stage (Haiku) call failed, using evidence-based fallback prose: %s", e)
         return {}
+    if not content or not content.strip():
+        logger.warning("cards stage LLM returned empty content — using evidence-based fallback prose")
+        return {}
     try:
         cards = _parse_json(content).get("cards") or []
     except Exception:  # noqa: BLE001