04 — Step A: Ownership Seed Commit on Clone

Verdict: PASS — FIRST persistent ownership seed in project history. gap 210→0. SQL sql/A_ownership_seed_commit_clone.sql sha256 07c7dda62a260486812d9ab36b91882c68f56bb409ee17c6cf71a71304fbf305.

Committed 30 group-level anchors = 5 groups × 6 active scopes, owner_kind=accountable, is_inherited_anchor=true, lifecycle=active, created_by=CLONE_TEST_F6_SEED, approval_ref=CLONE-TEST-NO-RATIFICATION. Mapping: GRP-GOVERNANCE→GOV-COUNCIL, GRP-REGISTRY→GOV-KG-SYS, GRP-AI→GOV-DOT, GRP-BUSINESS→GOV-MOIT, GRP-WORKFLOW→GOV-DOT.

Difference from production script (deliberate, safe): production commit_ready_f6_ownership_seed.sql self-aborts unless os_proposal_approvals>=1 (sovereign gate protecting PRODUCTION). Clone twin removes ONLY that gate (isolated disposable test DB, no sovereignty) while keeping every structural guard: abort unless current_database() matches directus_gov_test_%; abort unless pre-state gap=210/own=0/conflict=0; abort unless post-insert exactly 30 rows/gap=0/conflict=0; provenance unambiguously clone-test.

Execution (persistent COMMIT): ENTRY own=0 gap=210 conflict=0 → NOTICE CLONE GATE OPEN → INSERT 0 30 → NOTICE PRE-COMMIT OK new=30 gap=0 conflict=0 → COMMIT → POST own=30 gap=0 conflict=0 gov_emit=0.

Why 30 anchors close 210 gaps: containment is depth-1, every one of 35 governed objects is a direct child of exactly one of 5 groups; a group accountable anchor is inherited by all children across all 6 scopes; 5×6=30 anchors resolve 35×6=210 cells. Group-anchoring is maximally efficient — headline result now COMMITTED, not just rollback-only (F-5/F-6).

Side-effect: governance_object_ownership has 0 triggers ⇒ insert has no write path; gov_emit=0 after. Significance: first time the seed PERSISTED and gap actually driven to 0 in a committed DB; de-risks production rollout (production commit = identical op minus gate-removal).