KB-F4D4

dot-iu-cutter v0.5 First Snapshot MARK Dry-run Output Review — Manifest Structure Review (2026-05-19)

5 min read Revision 1
dot-iu-cutterv0.5first-snapshot-mark-dryrunoutput-reviewmanifest-reviewdieu442026-05-19

dot-iu-cutter v0.5 — First Snapshot MARK Dry-run Output Review · Manifest Structure Review

basis: existing manifest-summary-and-digests-2026-05-18.md + review_evaluation.json + verification-result (V-4,V-5,V-11,V-12)
note: full manifest.json (84,157 B) is not stored verbatim in KB — sanctioned summary+digests form per command-package large-artifact ruling; deterministically reproducible (re_run_equal:true)

3. Manifest structure review

3.1 Emitted levels

levels_present (review_evaluation.json): [DIEU, KIEN_TRUC_SECTION, NGUYEN_TAC]
required:                                 NGUYEN_TAC, KIEN_TRUC_SECTION, DIEU
result: PASS — all three required levels present, no foreign/extra level
level_census: NGUYEN_TAC=15, KIEN_TRUC_SECTION=3, DIEU=42  → 15+3+42 = 60 = candidate_count ✅

Verification V-5 (levels_present ⊇ {NGUYEN_TAC, KIEN_TRUC_SECTION, DIEU}) and V-12 (hierarchy, no orphan/foreign) PASS.

3.2 Canonical addresses

  • All 60 candidate addresses begin with ICX-CONST/ (full 60-unit index inspected in manifest summary).
  • Address uniqueness: review_evaluation.address_unique = true; verification V-11 PASS.
  • Status excluded from address: V-11 explicitly observes "all addresses ICX-CONST/<path>, unique; status never in address". Confirmed by inspection — no enacted/draft/controlled_draft token appears in any address path.
  • Address key shapes observed: NT-1..NT-15, KT-A/KT-B/KT-C, and DIEU keys DIEU-TERMINOLOGY, DIEU-0, DIEU-0-B/G/H/S-M-L, DIEU-1..DIEU-43 (non-contiguous).

result: PASS — address prefix, uniqueness, and status-exclusion all satisfied.

3.3 Provenance / version binding

source_document_version_id (manifest_header): icxconst-008a06ace23a96ea6cd456146e805c97
verification V-4: 100% of candidates+excluded carry provenance.source_document_version_id
                  = icxconst-008a06… (single distinct value)
snapshot binding (manifest_header): snapshot_region_sha256 17660443…cae80c, length 17522,
                  artifact tests/fixtures/constitution-normalized-17660443e0f23e99.md
result: PASS — version + snapshot identity bound on every row

3.4 Manifest schema sanity

  • manifest_header carries: generated_for, source_document_version_id, snapshot path/sha/length, marker_census_observed {enacted:19,controlled_draft:1,draft:1,obsolete:1}, grammar_profile incomex-architecture-constitution-v4, address_template at.icx.const.v4, parser_refimpl nuxt-incomex-portal-constitution-v1.refimpl.r1, docprefix ICX-CONST, scope_policy enacted_only, mode mark-manifest-only, db_write NONE, candidate_count 60, excluded_count 4, noncontent_count 69, manifest_digest.
  • Candidate index + excluded index are internally consistent with the header counts (60 / 4) and with coverage_proof line accounting.
  • result: PASS — schema is well-formed; header counts reconcile with indices and coverage proof.

Observations (non-defect — carry-forward to CUT command-review)

  • N-1 (transport, not content): Full manifest.json (84,157 B, sha256 7d56f3ce…012179) is not stored verbatim in KB; only the summary+digests form is. This was sanctioned by the command-package large-artifact ruling and the manifest is byte-reproducible from the pinned deterministic command. CUT command-review will require the full per-unit manifest body (full provenance objects, span text, per-unit fields). Recommendation for the next phase: regenerate the full manifest deterministically into ephemeral scratch and validate sha256 == 7d56f3ce…012179 before designing the CUT row-set (or persist a content-addressed/base64 full-manifest blob). This is a logistics precondition, not a manifest defect and not a blocker for the readiness decision.
  • N-2 (heterogeneous DIEU keys): DIEU address keys are non-contiguous and include compound/named forms (DIEU-0, DIEU-0-B, DIEU-TERMINOLOGY). This is intentional (catalog Điều numbers are not all enacted content rows) and fully reconciled by reconstruction_ok:true. CUT command-review should treat manifest addresses as canonical and stable — do not renumber or normalise DIEU keys at CUT.
  • N-3 (marker-vs-candidate ratio): 19 enacted marker tokens promulgate 60 enacted candidates via the ratified 3-tier status cascade (doc→group→row). This is by design (one group marker promulgates many units), confirmed by V-2/V-6; not an anomaly.

Manifest review verdict

manifest_structure: PASS (levels, addresses, uniqueness, status-exclusion, provenance, schema all correct)
defects_found: NONE
carry_forward_notes: N-1 (full-manifest transport for CUT), N-2 (canonical DIEU keys), N-3 (cascade ratio is by-design)
Back to Knowledge Hub knowledge/dev/laws/dieu44-trien-khai/v0.5-first-snapshot-mark-dryrun-output-review/dot-iu-cutter-v0.5-first-snapshot-mark-manifest-review-2026-05-19.md