14 — Open Questions & GO/NO-GO

The decision document. Consolidates findings into blocker tiers, lists open questions for human/GPT/council, and gives the GO/NO-GO recommendation for moving to canonical design patching.

14.1 Findings roll-up

• Internal contradictions found (6): B5/F4 (exception as owner-path vs separate term) · F1 (per-object identity vs container-grain) · F3/J5 (gate "covered==true" vs warning non-blocking) · I2/J3 (apply writes object edge vs CHECK-blocked + §5.4-EXT deferred) · C2 (federated multi-owner vs Đ37 §4.12 one-owner) · D2/J2 (anarchic "capacity" circular vs governed-object definition).
• Unaddressed gaps (mission-required, not in pack): A2/B1 (non-governed class) · C1 (role taxonomy / accountable-vs-support) · D1 (birth↔governance joint matrix) · E1 (exception replacement_plan + full fields) · G2 (route registry) · K1/K3 (tiered gate + waiver authority).
• Live-verified traps (substrate can't yet do what the pack assumes): I1/E2 (no grant_exception/assign_owner APR action-type) · I2 (no agency→object edge) · H1 (event names mis-cited).
• Memory-dependence re-introductions: B3 (design_ref "by convention") · G1 (L1 completeness base case) · G3 (context-change blind spot) · G4 (unsourced thresholds).

14.2 Tier-1 — BLOCKERS (must resolve before canonical design patch)

These either (a) make the deployed system noisy/unusable, (b) let real gaps hide, or (c) make a drafted clause self-contradictory or inoperable.

#	Blocker	Finding	Why it blocks
T1-1	Define the non-governed (Class 0) artifact + shared-truth test	A2/B1, M-DEF-1	Without it the scanner floods on personal pins/prefs (noise → alarm fatigue → real islands hide). Contradicts the mission's anti-over-governance instruction.
T1-2	Define accountable-owner-per-scope + role taxonomy; reconcile with Đ37 §4.12	C1/C2, M-DEF-3	The new Đ37 §4.15 otherwise contradicts the enacted §4.12; the federated model is unstatable.
T1-3	Fix the exception/owner-path contradiction (B5) and the identity grain (F1) and gate severity (F3)	B5/F1/F3/F4	Three contradictions in the invariant + gate — the keystone. If imprecise, the identity closes while gaps hide and the gate is permanently red.
T1-4	Inheritance covers owner-link ONLY (anti-hiding)	B4, M-DEF-7	Red-team #20: child policy gaps hide behind covered parents — the headline scale risk.
T1-5	Birth↔governance dedup precedence	D1, M-DEF-4	Without it, two scanners double-fire (duplicate noise — mission §7).
T1-6	Substrate prerequisites: grant_exception/assign_owner action-types (I1/E2) + reclassify §5.4-EXT object edge from "deferred" to "prerequisite for object-grain apply" (I2/B7)	I1/I2/E2/B7	Red-team #13/#14 ❌: the remediation half cannot run (PROPOSE files malformed APRs; APPLY can't assign object owners). This is the only blocker needing a future substrate change; the rest are wording.
T1-7	Route registry / detection completeness	G1/G2/G3	Red-team #1/#21/#22: the most island-prone surface (routes) is undetectable; L1 completeness base case + context-change triggers prevent silent rot.

14.3 Tier-2 — SHOULD-FIX (resolve during the design-patch phase, can run in parallel)

D2/J2 (anarchic re-base) · E1/E4/E5 (full exception record) · F2 (gate ignored) · G6 (stale/unverifiable identity term) · H1/H2/H3/H4 (event names, issue vocabulary, anti-spam) · I3 (bootstrap seed) · I4/C6 (separation of duty) · J6 (interim render-delegation) · J7 (label/taxonomy owner split) · K1/K3 (tiered gate + waiver) · A3 (QUARANTINED) · A4/K4 (emergency lane).

14.4 Tier-3 — NICE-TO-HAVE / NOTE

B3 (design_ref advisory) · A6/A7 (wording, F-ISLAND-8 split) · D3/D4 (island roll-up framing) · G7/G8 (emit ceiling, Nitro-in-scope) · J10 (slot/version drift — a P3 prerequisite check) · K5/K7 (evidence manifest, gate ordering).

14.5 Open questions for human / GPT / council

OQ	Question	Owner	Recommended default
OQ-A2	Boundary for shareable-but-personal artifacts (a personal pin that can be exported/shared)?	COUNCIL	shareable ⇒ governed at share time
OQ-A3	Standard regularization deadline for an inherited/legacy live bypass?	COUNCIL	60 days
OQ-B2	PROFILE-SURFACE-RO for read-only routes (drop rollback)?	SIV	yes
OQ-B3	Build a design_link registry, or accept design_ref permanently advisory?	COUNCIL	advisory now
OQ-B7/I2	Extend governance_relations (CHECK change) vs new governance_object_ownership table?	COUNCIL+SIV	new table (no CHECK-migration risk on live edges)
OQ-C7/E2/I1	One bundle of new APR action-types (assign_governance_owner, grant_governance_exception, delegate_authority)? Handlers or council-review-only?	COUNCIL	one bundle; owner-assign has a handler, exception/delegate council-review
OQ-D1	Does the Đ37 agency-orphan detector overlap the new object-coverage detector for laws 24/26/28/45?	SIV	keep law/agency gaps with Đ37 detector; object gaps with new detector; state boundary
OQ-D3	Which island sub-types are PG-detectable vs CI-only?	SIV	no-owner-table=PG; approval-flag/owner-constant=CI
OQ-E1/K3	Max exception/waiver renewals before replacement_plan must execute?	COUNCIL	2
OQ-F3/H3/K2	warning→high escalation deadline?	COUNCIL	30 days
OQ-G1	File/config ground truth (model-B File:… substrate, nginx)?	SIV	defined scan root on VPS
OQ-G2	Route registry: new table vs derived-on-scan from nginx+Nuxt?	MOUT/SIV	derived-on-scan (reuse-first)
OQ-H1	New governance/integrity event domain (GOV-SIV) vs reuse mother dormant rows?	SIV+Đ45	new domain; leave mother rows for factories
OQ-I2	Bring §5.4-EXT forward as Tier-1 prerequisite?	COUNCIL	yes
OQ-I5	Verify the event_outbox registry CHECK actually exists (fail-closed)?	SIV	verify in P4
OQ-J6	Provisional COUNCIL render-delegation vs fast-track GOV-MOUT activation?	COUNCIL	delegation now, activation as end-state
OQ-J7	Taxonomy/label substrate owner (GOV-KG-SYS vs dedicated taxonomy agency)?	COUNCIL	GOV-KG-SYS
OQ-J10	Resolve Đ45 (ban_hanh=false leftover) + Đ36 (v4.0 vs v5.0) authoritative status (content-only)?	NRM-SYS	resolve before P3
OQ-K6	CI blocking-in-deploy vs scheduled-detect?	SIV	blocking-in-deploy for a true gate

14.6 GO / NO-GO recommendation

RECOMMENDATION: NO-GO for canonical design patching until the Tier-1 blockers (§14.2) are folded into a hardening revision of the decision pack. After that single revision, GO.

Rationale:

• The decision pack is structurally sound, live-anchored, reuse-first, and non-islanding — its architecture is the right one and should be kept, not redesigned.
• But the red-team is decisive (doc 12): the un-hardened pack catches only the cases the existing Đ35/Đ45 machinery already enforces; 19 of 24 governance-specific attacks require a hardening fix, and 2 cannot run at all until substrate prerequisites land. Patching canonical design on top of an invariant with 3 internal contradictions and a remediation path that can't execute would propagate those defects into the design docs and then into law.
• The fixes are almost entirely wording-level and self-contained (M-DEF-1..7 + the Đ37/31/35/24/26/28/45 revisions in doc 13). Only one blocker (T1-6) needs a future substrate change (new APR action-types + §5.4-EXT object edge) — and that is design/prep work, not part of this or the next macro.

CONDITIONAL GO available: the design-patch concept docs (doc 09 P-09.1/P-09.4 — the One-Roof statement + the new 14-governance-coverage… doc) may proceed if and only if they carry the hardened definitions (M-DEF-1..7) and explicitly flag T1-6 as a substrate prerequisite. The surface-binding patches (P-09.2/3/5/6/7) should wait until T1-1..T1-7 are settled, because they encode the owner/gate/exception semantics that the blockers change.

14.7 Answers to the mission's final-response questions

• Clause-hardening verdict: the pack is a strong v1 but not yet tight enough; 6 contradictions + 6 gaps + substrate traps must be closed.
• Hardened definitions ready? — partially YES: doc 13 provides ready hardened wording for every reviewed clause + the 7 missing definitions; they need council adoption, not more drafting.
• Readiness gate ready? — NO (needs K1+K3+F3): must become tiered + severity-aware + waiver-governed before it is a usable gate.
• Red-team verdict: the un-hardened pack does not yet stop the attacks it was built to stop (doc 12); after Tier-1 it stops 22/24, with 2 (#13/#14) pending the substrate prerequisite.
• Biggest issues: non-governed class (A2), inheritance anti-hiding (B4), the invariant/gate contradictions (B5/F1/F3), the substrate-can't-do-it traps (I1/I2/E2), and route undetectability (G2).
• GO/NO-GO: NO-GO now; GO after one hardening revision (Tier-1). Conditional GO for the two concept-only design docs that carry the hardened definitions.

Next: doc 15 (prompts), doc 16 (self-review).