10 — Anti-Forgetting Ledger v2

Per mission §5: for each domain — what was cross-checked this pass, first-pass coverage verdict, possible misses, confidence. This ledger is built on independent live re-verification, not re-reading the first pass.

Newly-surfaced prior reports (the first pass's main completeness gap)

The first pass (doc 05) listed only 2026-05-28/29/30 packs. This pass surfaced the foundational April family: S150+ Đ37/38 bootstrap (NRM 39→47), S164 (Đ38 14 normative DOTs, total 265), S165 (Đ37 v3.3 restore), S165-KB (Đ35/Đ36 file restore + reference-table/config-table anti-hardcode pattern), S166-KB (KB Protection Phase 2 version history), S178 (DOT Repair Governance → Đ35 v5.2). None introduce a missed law or a capability that overturns reuse-first. Two are worth explicit capture: S165-KB's reference/config-table pattern (doc 07) and S178 as the DOT-governance origin (doc 06).

Domain ledger

#	Domain	Cross-checked this pass	1st-pass verdict	Possible misses	Confidence
1	Pivot/counting	live counts (37/126); PIV-001/007/008/104/207 filter_spec read; v_registry_counts 8 CATs; trigger status	VERIFIED + 1 correction (trigger enabled, not disabled)	exact LỆCH formula (count_b recompute) not yet characterized — needs read-only test_counting_contract()	0.96
2	DOT governance	dot_tools 9-field breakdown; CAT-006 vs CAT-DOT; Đ35 v5.2 title + S178	VERIFIED + sharper dual-count	dot_iu_command_catalog sub-fields not re-counted; scanner fn bodies not read	0.92
3	Birth/collection	birth_registry live; meta_catalog/collection_registry/entity_species/table_registry counts	VERIFIED + 1 stale literal (959,372)	fn_pre_birth_check/fn_birth_gate bodies not re-read (high prior confidence)	0.93
4	IU/profile	information_unit unit_kind distinct; count 219	VERIFIED	Đ44 DRAFT → target profile schema not final	0.85
5	KG/relation	iu_relation 60; universal_edges 2199; v_kg_edges_all 2259 (=union)	VERIFIED	edge semantics by type not re-aggregated	0.90
6	Event/queue	event_outbox live; route 15/15 dry_run; DLQ 0/0; event_type_registry 40	VERIFIED + 1 stale literal (166,422)	retry/idempotency policy contents not read	0.90
7	Approval/governance	approval_requests 211; apr_approvals 28+14; wcr 3; governance_registry 9	VERIFIED	apr_request/action_types not re-listed	0.92
8	Workflow/MOWD	workflows 2 / steps 70; inline-not-IU (carried)	VERIFIED (carried)	iu_ref absence not re-queried by column this pass	0.85
9	UI Preview official lists	review-log.md read (stale 19/4); ui_pages 37; CAT-007 LỆCH (37 vs 52); manifest 61 (first-pass fetch)	VERIFIED + new CAT-007 LỆCH	manifest not re-fetched live this pass (read_file not allowlisted for static dir)	0.88
10	Law/normative registry	full NRM dump (47 rows, article set); law_catalog (5 rows); Đ45 enacted evidence; Đ23 KB	VERIFIED + sharper (dieu28 = wrong law, not drift)	articles 20/21/25/27/40/42 status unconfirmed (likely unallocated)	0.95
11	Scanner/detector	named fns carried; trigger pg_trigger query	PARTIAL (carried)	most scanner bodies not read; same 0.78 residual as first pass	0.80
12	Rollback/audit	BEGIN..ROLLBACK reversibility carried from prior G1/MOWD packs; no live re-run	CARRIED	not re-exercised this read-only pass (would mutate)	0.85

Possible misses still open (honest)

• Exact LỆCH cause — characterized as record_count↔recompute mismatch, not the trigger; precise formula needs a read-only test_counting_contract() run (P1).
• Scanner fn bodies (~150 fn_iu_* + scanners) enumerated by signature, not all read — same residual the first pass declared; no evidence of a hidden capability, but not exhaustively proven.
• manifest.json not re-fetched live this pass — relied on first-pass HTTP record (61/36/25/16) + KB; read_file allowlist excludes the static web dir. Low risk (first pass fetched it 2026-05-30, same day).
• Articles 20/21/25/27/40/42 absent from NRM — not on mandatory list; not proven whether reserved/unwritten vs hidden. Recommend a one-line check before any "all laws" enumeration ships.

Gate

Ledger v2 complete → decision gate (doc 11) authorized. No domain dropped below 0.80 except scanner (0.80) — same as first pass. Net confidence in the reuse-first thesis: HIGH.