fix(store): guard store_findings against None engagement_score on update path

CHANGELOG.md

@@ -11,6 +11,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - Native Grok Build (xAI) plugin and marketplace lane: `.grok-plugin/plugin.json` + `.grok-plugin/marketplace.json` so `grok plugin install mvanhorn/last30days-skill` and `grok plugin marketplace add mvanhorn/last30days-skill` work as first-class install paths. The self-hosted catalog uses a bare Git URL source (tracks HEAD); submitting to the official `xai-org/plugin-marketplace` remains a post-merge SHA-pinned outbound PR documented in `AGENTS.md`.
 
+### Fixed
+
+- `store_findings` no longer raises `TypeError` when a re-sighted finding carries `engagement_score: None`. `.get("engagement_score", 0)` only substitutes `0` for an *absent* key, so a present-but-`None` value reached `max(None, …)` on the update branch — `new_engagement` is now guarded with `or 0`, symmetric with the existing-row side. Hardens the public `store_findings(List[Dict[str, Any]])` boundary for arbitrary callers; the standard `findings_from_report` pipeline already coerces `None`→0 at the producer, so this closes the API boundary rather than a reachable pipeline crash. Regression test added in `tests/test_store.py`.
+
 ## [3.11.0] - 2026-07-05
 
 ### Added

skills/last30days/scripts/store.py

@@ -424,7 +424,7 @@ def store_findings(
 
         for url, f in with_urls:
             existing = existing_by_url.get(url)
-            new_engagement = f.get("engagement_score", 0)
+            new_engagement = f.get("engagement_score") or 0
             if existing:
                 update_rows.append((
                     max(new_engagement, existing["engagement_score"] or 0),

tests/test_store.py

@@ -848,5 +848,48 @@ def test_get_new_findings_filters_by_date(temp_db, sample_report):
 
     assert len(new_findings) == 4
 
+
+def test_store_findings_none_engagement_on_update_does_not_crash(temp_db):
+    """Regression: store_findings must not TypeError when an update-path finding
+    carries engagement_score=None.
+
+    .get("engagement_score", 0) only substitutes 0 for an *absent* key, so a
+    present-but-None value reached max(None, existing) on the update branch.
+    store_findings accepts arbitrary List[Dict[str, Any]] callers, so the
+    boundary must stay null-safe (engagement_score is float | None in schema.py).
+    """
+    topic_id = store.add_topic("None Engagement Topic")["id"]
+    url = "https://example.com/none-engagement"
+    base = {
+        "source": "reddit",
+        "source_url": url,
+        "source_title": "T",
+        "author": "",
+        "content": "",
+        "summary": "",
+        "relevance_score": 0.5,
+    }
+
+    # First sighting carries a real engagement score.
+    run1 = store.record_run(topic_id, source_mode="test", status="running")
+    store.store_findings(run1, topic_id, [{**base, "engagement_score": 5.0}])
+
+    # Re-sighting the same URL with engagement_score=None takes the UPDATE branch;
+    # before the fix this raised TypeError from max(None, 5.0).
+    run2 = store.record_run(topic_id, source_mode="test", status="running")
+    counts = store.store_findings(run2, topic_id, [{**base, "engagement_score": None}])
+    assert counts["updated"] == 1
+
+    con = sqlite3.connect(str(temp_db))
+    try:
+        stored = con.execute(
+            "SELECT engagement_score FROM findings WHERE source_url = ?", (url,)
+        ).fetchone()[0]
+    finally:
+        con.close()
+    # max(None -> 0, existing 5.0) keeps the higher real score.
+    assert stored == 5.0
+
+
 if __name__ == "__main__":
     pytest.main([__file__, "-v"])