[NOTICKET] check: read documents from analysis-scope catalog

Documents live ONLY in the analysis-scope catalog row (verified via the
`data_catalog` table: `scope_type='analysis'` rows carry `unstructured`
sources; user-scope rows have none). So the first cut — scoping only
`structured` reads to analysis and leaving documents on the user-scope reader —
made `check_knowledge` always return empty ("not listed"), because user-scope
has zero documents.

Fix: `AnalysisScopedCatalogReader` now reads analysis-scope for BOTH structured
and unstructured, falling back to the user-scope reader only when the analysis
has no catalog row. No regression risk: user-scope carried no documents anyway.

A document not bound to the current analysis still won't appear in that room —
there is no user-level document inventory in `data_catalog`.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

src/catalog/reader.py

@@ -77,10 +77,12 @@ class AnalysisScopedCatalogReader(CatalogReader):
     """Reads the analysis-scope catalog, falling back to the user-scope reader.
 
     Used by the `check` skill so "what data do I have" inside a room reflects
-    that analysis's bound sources AND their real names — a database shows as
-    "xl test" (analysis-scope) instead of the auto-generated `postgres_<hash>`
-    placeholder stored in the user-scope row. When the analysis has no catalog
-    row (legacy / not yet bound) or the read fails, it degrades to the wrapped
+    that analysis's bound sources — structured AND documents — with their real
+    names. A database shows as "xl test" (analysis-scope) instead of the
+    auto-generated `postgres_<hash>` placeholder, and documents show at all
+    (the user-scope catalog holds no `unstructured` sources, so reading them from
+    user-scope always came back empty). When the analysis has no catalog row
+    (legacy / not yet bound) or the read fails, it degrades to the wrapped
     user-scope reader, so unbound rooms behave exactly as before.
     """
 
@@ -92,13 +94,15 @@ class AnalysisScopedCatalogReader(CatalogReader):
         self._analysis_id = analysis_id
 
     async def read(self, user_id: str, source_hint: SourceHint) -> Catalog:
-        # Only STRUCTURED reads get analysis scope — that's where the naming/scope
-        # problem lives (databases named `postgres_<hash>` in user-scope vs their
-        # real name in analysis-scope). Documents pass through to the user-scope
-        # reader: they carry no scope-name gap, and we don't assume Go populates
-        # the analysis-scope catalog with unstructured sources (scoping them could
-        # wrongly hide a user's documents in a room).
-        if source_hint == "structured" and self._analysis_id:
+        # Read analysis-scope for BOTH structured and unstructured. Verified via
+        # the dedorch `data_catalog` table: the analysis-scope rows carry the real
+        # DB names AND the room's documents (`source_type='unstructured'`), whereas
+        # the user-scope rows hold only structured sources with `postgres_<hash>`
+        # placeholder names and NO documents at all — so reading documents from
+        # user-scope always returned empty ("not listed"). Fall back to the
+        # user-scope reader only when the analysis has no catalog row (legacy /
+        # unbound room).
+        if self._analysis_id:
             try:
                 catalog = await self._store.get_by_analysis(self._analysis_id)
             except Exception:  # noqa: BLE001 — never block check on the analysis read