03 — Inventory Seam Results

Output of S1/S2 (doc 02). Read-only; collection grain over the 3-collection subset.

3.1 Subset selected (S1)

The lexically first 3 of the 35 governed+BIRTH_REQUIRED collections:

#	collection_name	governance_role	coverage_status
1	agents	governed	BIRTH_REQUIRED
2	approval_requests	governed	BIRTH_REQUIRED
3	apr_action_types	governed	BIRTH_REQUIRED

3.2 Inventory seam output (S2)

The Seam-A collection-grain expression, evaluated inline over the subset, yields exactly 3 rows:

object_type	object_ref	requires_owner	born
collection	agents	true	true
collection	approval_requests	true	true
collection	apr_action_types	true	true

3.3 Interpretation

• requires_owner=true for all 3 — correct: governance_role='governed' ∈ ('governed','locked').
• born=true for all 3 — correct: coverage_status='BIRTH_REQUIRED'.
• Row count = 3 = expected; ≤ 38 hard ceiling (doc 03 §3.3). No row explosion.
• The seam logic is sound: fed the 3-collection slice, it produces precisely the in-scope inventory rows with the correct flags. This is what v_governance_object_inventory would return at collection grain if it were wired to collection_registry (a ratified F-4 decision, not done here).

3.4 What this does NOT do

• It does not wire the live v_governance_object_inventory view — that view stays WHERE false (confirmed 0 rows, doc 05). The expression was run inline only.
• It does not touch object/birth_registry grain (1.08M rows) — collection grain only.
• It writes nothing.

3.5 Inventory seam verdict

PASS. Exactly 3 bounded, correctly-flagged inventory rows. The inventory seam logic is proven correct read-only without activating the seam.