07 — Findings & F-6 Readiness

F5-1 (load-bearing): coverage model end-to-end correct over the real F-4 substrate. 2-row seed → predicted 3-row inventory coverage and 210→207 gap. Direct (depth 0) and group-inherited (depth 1) both behave per view defs. No fixtures — containment is real collection_registry."group".

F5-2: direct accountability resolves at depth 0 independent of is_inherited_anchor; inheritance to children requires ancestor is_inherited_anchor=true. Two-mode contract confirmed empirically.

F5-3 (subtle, for reviewers): v_object_effective_owner reports the group anchor as a DIRECT owner of itself (direct=2 when only 1 collection seeded). Does NOT distort gap math because group nodes are not in inventory → gap reduction counts only inventory objects =3. Inventory-scope eo counts before comparing to gap.

F5-4: single-accountable-owner enforced by partial unique index uq_gov_obj_accountable, not by the conflict view. v_object_owner_conflict is a belt-and-suspenders detector that (given the index) can only fire on historical/non-active rows — returned 0 with a live duplicate attempted. Fail-closed at write layer.

F5-5: owner-row integrity fully constrained (FK agency, FK scope, CHECK owner_kind, chk_delegated_ttl, CHECK lifecycle). All six negatives rejected.

F5-6: ownership table has no triggers — inserting an owner has no birth/outbox/issue/audit side effect (F-83-1 trap absent). Sequence advance on rollback is the only residue, non-substantive.

F5-7 (carried): axis_registry/axis_assignment/coverage_rule remain ABSENT — axis-grain coverage still fake/test-only, out of scope. Object×scope (210 model) is the only real coverage surface.

F-6 readiness

Coverage path proven at every stage read-only AND under live transient seed: inventory→containment→direct→inherited→gap→conflict/negatives→rollback. Remaining unproven step toward activation = persistence of a real owner seed (a COMMIT of ownership rows) — exactly what the ratification gate governs.

Candidate F-6 macros (all gated, none authorized now):

• F-6 = ownership-seed COMMIT (TEST_MODE): persist a small, real, ratified ownership set. Requires L2+L4 ratification AND ospa≥1. First step that changes governed state beyond view visibility. NO-GO until ratified.
• F-6′ = larger rollback-only seed: rehearse full multi-group × multi-scope seed (e.g. anchor every group across all 6 scopes), project max gap reduction, find inheritance edge cases at scale. Still rollback-only; GO-eligible.
• Recursive/multi-level containment not exercisable (real containment strictly depth-1; deeper needs fixtures, out of scope).

The one gate (unchanged): L2+L4 sovereign ratification + a real os_proposal_approvals row remain the sole blockers to persistent owner commit/activation. NOT solved by F-5.