00 - T1-XHigh FIX7 Refactor Blueprint Independent Review - Readme First

Date: 2026-06-08 Reviewer: T1-XHigh (independent high-rigor reviewer for Agent Data) Macro: PROGRAM_REVIEW_XHIGH_FIX7_REFACTOR_BLUEPRINT_INDEPENDENT_AUDIT_AND_DIRECT_REVISION Mode: READ-ONLY production. Direct-revision of blueprint KB docs allowed. No production mutation.

What this is

An independent XHigh adversarial review + direct-revision pass over the FIX7 Existing-System Refactor Execution Blueprint (t1-fix7-existing-system-refactor-execution-blueprint-2026-06-08/). The blueprint's own embedded "XHigh/Max" self-review was treated as evidence only, NOT as final authority. Live production evidence and document consistency decided each verdict.

Final status

FIX7_REFACTOR_BLUEPRINT_XHIGH_REVIEWED_AND_REVISED_READY_FOR_MAX_REVIEW

Not a blocker: no hardcode/PG-native FAIL, read path open, no invariant changed. Real defects were found and directly fixed in the blueprint docs; this report records what and why.

Live evidence gathered (read-only, DB directus, 2026-06-08)

• Legacy/control functions proacl = NULL for fn_dot_birth_qt001_apply, sp_dot_birth_qt001_apply, fn_qt001_plan_v5, fn_dot_birth_qt001_plan_v2, fn_birth_registry_auto. In PostgreSQL a NULL ACL means the default EXECUTE to PUBLIC - i.e. the legacy apply/writer entrypoints are PUBLIC-executable right now. This concretely substantiates the "legacy blocked, not merely unreachable" requirement and the gateway-overwrite risk; the blueprint's neutralization step is genuinely load-bearing.
• Relations matching qt001%/birth% (no v_ prefix) owned by directus = 26 (= 20 qt001_* tables + 6 birth* relations); the 196 legacy views carry the v_qt001_* prefix. Consistent with the blueprint inventory; the "262 control objects" figure (cited from FIX6) is the broader functions+views+triggers count, not these 26 relations.

Findings summary (10 findings; all fixed in-blueprint; 0 blockers)

id	track	severity	one-line	fixed in
XHI-01	A integrity	P2	guard count stale (26 stated, 28 actual after embedded review)	doc 06, 12
XHI-02	A integrity	P1	G-LEGACY-FROZEN referenced (doc 04 S17, doc 07 PKG-H) but never defined	doc 06
XHB-01	C bypass	P1	no guard that frozen/old DOT or non-owner can overwrite gateway/control-plane	doc 06/04/07/08
XHB-02	C/F bypass	P1	S15 rollback restores legacy EXECUTE - read as silent bypass re-open	doc 05/04
XHD-01	D directus	P2	PKG-G no-go omitted app-table authority + unverified snapshot	doc 07
XHO-01	E order	P2	regclass source_relation binding order implicit; S19 precond wrong (S16->S18)	doc 04
XHH-01	I hard-block	P1	missing hard-block rows: birth-gateway-modification, registry-pivot-repoint	doc 08
XHL-01	L cross-layer	P2	no explicit OUT-OF-SCOPE for Đ43 / QT-006 / registry-pivot / raw-birth-as-truth	doc 08
XHJ-01	J hardcode	P3	S00 classification could read as manual-inventory-as-authority	doc 04
XHM-01	B mapping	P2	#26 mapping could read as replacing the birth gateway	doc 02

Document map

Doc	Track
00	This readme
01	A - blueprint source integrity
02	B - refactor-vs-greenfield
03	C - legacy neutralization / bypass
04	D - Directus read-vs-authority cutover
05	E - dependency-safe construction order
06	F - rollback
07	G - test / guard
08	H - package split
09	I - hard blocks / do-not-touch
10	J/K - hardcode / PG-native
11	L - cross-layer impact
12	M - direct revisions applied
13	final verdict

Boundary

This is XHigh review only. Next step after PASS is Max review. Implementation, Stage 2.6B, permit, REAL_RUN, QT001 apply, manifest activation, owner/ACL cutover all remain BLOCKED. No production object was created, altered, owned, granted, revoked, or executed.