test_doc_counts: AUTOCOUNT db-where supports *: glob for corpus-wide claims

The three "92 claim_pack docs" tags were drifting against shard 000.db's
21 rows because the harness only counted one shard, but the doc prose
("#000031 closed at 92") meant the corpus total (21+16+38+17 across
genesis shards 000-003).

Two-line fix path: extend the harness to sum across all ???.db shards
via a `*:` prefix (e.g. `*:documents?source_type=claim_pack`), then
prefix the three drifted tags. Aligns the harness scope with the
semantic scope of the claim instead of forcing the claim to shrink to
one shard.

The `*:` glob:
  - Matches `[0-9][0-9][0-9].db` basenames only (operator sidecars
    qa.db / snapshots.db / selfmodel-chain.db skipped)
  - Skips shards lacking the named table (schema-version tolerance)
  - Returns _DB_MISSING when no genesis shard exists (CI / fresh-
    checkout skip semantic preserved)
  - Returns _TABLE_MISSING when no contributing shard has the table

Documented in ticket-000044 §3.4 + a third example showing the new
syntax. Diagnosis credit to a sub-agent investigation that confirmed
zero eviction/falsification audit events on claim_packs — the data is
intact; the harness was just single-shard.

docs/tickets/ticket-000037-prometheus-sigma-recursive-falsification-controller.md

@@ -1030,7 +1030,7 @@ These are the right guardrails for the pure-function layer.
 
 ### 17.1 Warrant-promotion integration (#000031)
 
-#000031 closed at <!--AUTOCOUNT:db-where:documents?source_type=claim_pack-->92<!--/AUTOCOUNT-->
+#000031 closed at <!--AUTOCOUNT:db-where:*:documents?source_type=claim_pack-->92<!--/AUTOCOUNT-->
 atomic claim-pack records (started at 78, reached 92 by 2026-05-09
 under Phase 2.5). All 92 records now resolve to at least one
 textbook substrate via the 74 citation-aliases (#000041) + 13

docs/tickets/ticket-000044-autocount-doc-drift-discipline.md

@@ -115,8 +115,9 @@ query. SQLite connection opens `mode=ro` URI.
 ### 3.4 `db-where`
 
 ```markdown
-<!--AUTOCOUNT:db-where:documents?source_type=claim_pack-->92<!--/AUTOCOUNT-->
+<!--AUTOCOUNT:db-where:documents?source_type=claim_pack-->21<!--/AUTOCOUNT-->
 <!--AUTOCOUNT:db-where:001.db:documents?source_type=wikipedia_xml-->867695<!--/AUTOCOUNT-->
+<!--AUTOCOUNT:db-where:*:documents?source_type=claim_pack-->92<!--/AUTOCOUNT-->
 ```
 
 Live value: `SELECT COUNT(*) FROM <table> WHERE <column> = ?`
@@ -124,6 +125,16 @@ with `<value>` bound as a SQL parameter (not string-interpolated).
 Single-column equality predicate only — anything more complex is
 out of scope for v1 (write a custom assertion test instead).
 
+The `*:` prefix sums across every `???.db` shard (basename glob
+`[0-9][0-9][0-9].db`) in `~/.arborist/shards/`. Use this when the
+claim is corpus-wide rather than scoped to one shard — e.g. "92
+claim_pack docs across the corpus" (21+16+38+17 across genesis
+shards 000-003). Operator sidecar dbs (qa.db, snapshots.db,
+selfmodel-chain.db) are skipped. Shards lacking the named table
+contribute zero (different schema versions don't fail the assert);
+if NO shard has the table, the metric returns `_TABLE_MISSING`
+(skipped, like the single-shard case).
+
 ---
 
 ## 4. Skip semantics (CI / fresh-checkout safety)

docs/warrant-substrate-cookbook.md

@@ -34,7 +34,7 @@ patterns that keep the chains audit-clean.
 ## Pillar-by-pillar substrate map
 
 Eight pillars in the claim-pack source (`#000029`);
-<!--AUTOCOUNT:db-where:documents?source_type=claim_pack-->92<!--/AUTOCOUNT-->
+<!--AUTOCOUNT:db-where:*:documents?source_type=claim_pack-->92<!--/AUTOCOUNT-->
 records total. As of 2026-05-10, every record resolves to at
 least one substrate in `~/.arborist/crawl/textbook_*.db`.
 
@@ -346,7 +346,7 @@ Full cascade order, tightest → loosest:
 
 ## Honest tier breakdown
 
-The <!--AUTOCOUNT:db-where:documents?source_type=claim_pack-->92<!--/AUTOCOUNT--> chains have three quality tiers:
+The <!--AUTOCOUNT:db-where:*:documents?source_type=claim_pack-->92<!--/AUTOCOUNT--> chains have three quality tiers:
 
 - **Direct primary (~25 records)** — citation matches the literal
   textbook ingested. Hilbert pillar IV (18), Goldstein → Newton

tests/test_doc_counts.py

@@ -173,8 +173,14 @@ def _live_db_where(target: str) -> int:
 
     Target syntax::
 
-        <table>?<column>=<value>
-        <shard>:<table>?<column>=<value>
+        <table>?<column>=<value>            # default shard 000.db
+        <shard>:<table>?<column>=<value>    # one named shard
+        *:<table>?<column>=<value>          # sum across ALL ???.db shards
+
+    The ``*`` form is for claims that span the corpus (e.g. "92
+    claim_pack docs" is 21+16+38+17 across genesis shards 000-003).
+    Globs match basenames ``[0-9][0-9][0-9].db`` only — operator
+    sidecar dbs (qa.db, snapshots.db) aren't summed.
 
     Same sentinel returns as ``_live_db_rows``. ``<value>`` is bound
     as a SQL parameter (no string interpolation), so even author-
@@ -195,6 +201,8 @@ def _live_db_where(target: str) -> int:
 
     if ":" in head:
         shard_name, table = head.split(":", 1)
+        if shard_name == "*":
+            return _live_db_where_sum_shards(table, column, value)
         db = _DEFAULT_SHARDS_DIR / shard_name
     else:
         db = _DEFAULT_SHARD
@@ -225,6 +233,50 @@ def _live_db_where(target: str) -> int:
         return _DB_ERROR
 
 
+def _live_db_where_sum_shards(table: str, column: str, value: str) -> int:
+    """Sum rows matching ``<column>=<value>`` across all genesis shards
+    (``???.db`` basename pattern) in ``_DEFAULT_SHARDS_DIR``.
+
+    Skips shards where the table is absent (different schema versions);
+    returns _DB_MISSING when no shard matches the glob at all (operator
+    state, not source state). Identifier validation matches the single-
+    shard path.
+    """
+    import sqlite3
+
+    if not re.fullmatch(r"[A-Za-z_][A-Za-z0-9_]*", table):
+        return _DB_ERROR
+    shards = sorted(_DEFAULT_SHARDS_DIR.glob("[0-9][0-9][0-9].db"))
+    if not shards:
+        return _DB_MISSING
+    total = 0
+    saw_any_table = False
+    for db in shards:
+        try:
+            c = sqlite3.connect(f"file:{db}?mode=ro", uri=True)
+            try:
+                row = c.execute(
+                    f"SELECT COUNT(*) FROM {table} WHERE {column} = ?",
+                    (value,),
+                ).fetchone()
+                if row:
+                    total += int(row[0])
+                    saw_any_table = True
+            finally:
+                c.close()
+        except sqlite3.OperationalError as exc:
+            if "no such table" in str(exc).lower():
+                continue  # shard predates the table — skip, don't fail
+            if "no such column" in str(exc).lower():
+                return _DB_ERROR
+            return _DB_ERROR
+        except sqlite3.Error:
+            return _DB_ERROR
+    if not saw_any_table:
+        return _TABLE_MISSING
+    return total
+
+
 def _strip_fenced_code_blocks(text: str) -> str:
     """Replace contents of triple-backtick fenced code blocks with newlines.