fix(bookkeeping): classify template audit evidence (#1594)

DECISIONS.md

@@ -934,3 +934,5 @@ One line per decision: `[YYYY-MM-DD] <decision>: <why>`. Appended by agents and
 [2026-08-13] Kontantmetoden year-end VAT supersedes the 2026-08-06 VAT-reporting premise: BAS 2618/2628/2638 and 2648 feed the final declaration, reverse-charge purchases include both VAT sides and their basis, and only the mechanical day-one reversal is excluded from later VAT periods. Skatteverket requires unpaid invoice VAT in the final period and warns against reporting it twice after year end.
 [2026-08-13] Per-account VAT treatment is class-aware and explicit values override the static BAS mapping; SIE #SRU and #KTYP never supply it because they encode tax-return fields and account class, not momsdeklaration treatment.
 [2026-08-13] Per-account VAT treatment is class-aware; explicit values extend custom accounts while canonical accounts keep their static BAS momsdeklaration mapping. SIE #SRU and #KTYP never supply it because they encode tax-return fields and account class, not momsdeklaration treatment. VMB carries no default account rate because its VAT base is the margin, not gross sales.
+[2026-08-13] Issue #1408 separates evidence classification from lock and VAT overlays, and treats import and correction sources as false positives for live-template provenance: 1,316 of 1,361 production signature matches came from those sources, so a broad account signature must not become a correction queue.
+[2026-08-13] Issue #1408 does not route locked candidates through the current correctEntry service or an unlock: the service creates its storno in the original period, so a compliant locked-period correction needs a separately tested open-period storno and replacement path that preserves the original and full rättelse trail.

docs/TEMPLATE_MISBOOKING_REMEDIATION.md

@@ -48,11 +48,13 @@ anything.
   `gnubok_correct_entry`. Never `gnubok_reverse_journal_entry` alone: the
   business event remains valid, only its classification is wrong.
 - Do not write directly to `journal_entries` or `journal_entry_lines`.
-- Any `effective_lock_status` other than `open` is a hard stop. The correction
-  path never bypasses a lock. Unlocking or reopening requires a separately
-  reviewed and explicitly approved workflow for the exact company and period.
-  If a declaration or closing has relied on the period, establish the required
-  omprövning or closing consequences before requesting that approval.
+- Any `effective_lock_status` other than `open` is a hard stop for the current
+  `gnubok_correct_entry` path. Never unlock, reopen or edit the original merely
+  to make a correction possible. A locked-period correction requires a
+  separately reviewed and tested path that records a storno and replacement in
+  an open period, references the original, and preserves who corrected it and
+  when. If a declaration or closing relied on the locked period, establish the
+  required omprövning or closing consequences before requesting approval.
 - Never run a production correction without explicit approval for the exact
   company, vouchers and replacement lines.
 - No automated bulk mutation. Ever.
@@ -63,16 +65,41 @@ Run [`scripts/audit-template-misbookings.sql`](../scripts/audit-template-misbook
 against the intended database. It is read-only, creates nothing, and returns
 only posted entries.
 
-**Output is diagnostic, never a verdict.** There is no provenance link from a
-posted entry back to the template that produced it (`template_id` lives on
-`mapping_rules`, not on journal entries), so candidates are matched by account
-signature, and both signatures have legitimate shapes:
+The audit now separates the broad signature from the evidence classification.
+There is no direct template id on a journal entry, but `source_type`, exact line
+shape, transaction context and linked underlag can rule candidates in or out.
+Both signatures still have legitimate shapes:
 
 - `5820` is the **correct** account for actual car hire.
 - Representation at 25% is lawful only when the supply is actually subject to
   25% VAT and the supplier invoice is correct.
 
-Read `review_priority` as an ordering aid only:
+Read `evidence_classification` as the remediation queue:
+
+| Classification | Meaning |
+|---|---|
+| `confirmed_correction` | Exact defective shape plus corroborating evidence. Review the underlag, then stage `gnubok_correct_entry` and request explicit approval for the exact voucher and replacement lines. |
+| `insufficient_evidence` | The shape matches, but the underlag or provenance is not decisive. Obtain evidence before proposing any write. |
+| `false_positive` | Not attributable to either defective live template. This is not a general verdict that every aspect of the voucher is correct. |
+
+`source_type` values `import` and `correction` are false positives for this
+template-remediation audit because neither live template generated them. Keep
+any independent accounting concern outside this remediation batch.
+
+Linked underlag is resolved from direct journal-entry and journal-line links,
+transactions and supplier invoices. Document identifiers and extracted VAT
+signals are returned, but storage paths and file names are not. A decisive 25%
+extraction on a representation voucher rules out the defective 12% restaurant
+template signature. It does not by itself prove that the representation cap,
+participant record or business purpose was handled correctly.
+
+Every representation row therefore carries
+`independent_representation_review = participants_purpose_and_300_sek_vat_base_cap_required`.
+`observed_6072_cost_exceeds_300_sek` is a triage signal only: the statutory cap
+is per person and occasion, so the number of participants and the underlag are
+still required.
+
+Read `review_priority` as a supporting ordering aid:
 
 | Priority | Meaning |
 |---|---|
@@ -82,14 +109,25 @@ Read `review_priority` as an ordering aid only:
 | `manual_review_5820_no_corroborating_signal` | A `5820` entry with no hotel-shaped text or 12% VAT ratio. Most likely legitimate car hire. |
 | `manual_review_6072_with_vat` | A `6072` entry with `2641` VAT whose aggregate ratio is not exactly 25%. It may be correct or may be a mixed voucher masking the defective representation leg. |
 
-The classifier is verified against read-only probes: a hotel booked to `5820`
-ranks high, a genuine car hire on `5820` falls to manual review, and an exact
-25% representation signature ranks high. Correct 12% representation and mixed
-vouchers whose aggregate ratio is not 25% remain visible for manual review,
-because unrelated `2641` lines can otherwise hide a defective representation
-component. The representation probes include multiple `6072` and `2641` lines;
-the audit aggregates them per entry and returns one candidate instead of a
-many-to-many set of line pairs.
+The classifier is verified against read-only probes for confirmed, insufficient
+and false-positive outcomes. The audit aggregates multiple `6072` and `2641`
+lines per entry and returns one candidate instead of a many-to-many set of line
+pairs. Mixed representation vouchers without decisive underlag remain
+`insufficient_evidence`; they are never cleared merely because their aggregate
+VAT ratio differs from the exact defective 25% signature.
+
+Treat lock and VAT-return impact as overlays, not evidence classes:
+
+- `effective_lock_status` identifies candidates that the current correction
+  service cannot handle. Never unlock or reopen merely to permit a correction.
+- `vat_return_impact = potential_if_confirmed` means a representation candidate
+  could affect ruta 48. Determine the filing status before requesting approval.
+- `vat_reporting_period`, `vat_deadline_statuses` and `vat_filing_status`
+  surface the in-app filing record. `not_proven_by_in_app_records` is unknown,
+  not proof that no declaration was filed externally.
+- Keep exact company and voucher identifiers in an access-controlled approval
+  manifest. Public issue updates should use aggregate batches unless disclosure
+  has been explicitly approved.
 
 ## Review each candidate
 
@@ -107,8 +145,8 @@ many-to-many set of line pairs.
    filed. If so, an omprövning is in scope and the ledger correction alone is
    not sufficient.
 6. Decide materiality. A 40 kr VAT difference on one lunch two years ago in a
-   closed period is unlikely to warrant reopening anything; a systematic error
-   across a year is different. Record the decision either way.
+   closed period may not warrant a separate correction in an open period; a
+   systematic error across a year is different. Record the decision either way.
 
 Keep the reviewed set, the evidence, the proposed replacement lines and the
 reviewer identity together as the correction record.
@@ -121,6 +159,12 @@ retain `original_lines`, copy every original line, change only what is wrong,
 verify the replacement balances, stage
 `gnubok_correct_entry`, and approve only with explicit authorisation.
 
+`gnubok_correct_entry` routes through `correctEntry()` in
+`lib/core/bookkeeping/storno-service.ts`. It retains the original, creates a
+posted storno with `reverses_id`, creates the replacement with
+`correction_of_id`, and records the acting user and timestamps. Never replace
+this audited envelope with direct journal-table writes.
+
 For the hotel defect, the only change is `account_number` on the cost line,
 `5820` to `5830`.
 

scripts/audit-template-misbookings.probe.sql

@@ -1,68 +1,106 @@
--- Read-only classifier probes for scripts/audit-template-misbookings.sql.
--- Expected result: four rows, each with actual_result = expected_result.
+-- Read-only evidence-classifier probes for audit-template-misbookings.sql.
+-- Expected result: eight rows, each with passed = true.
 
-with probe_lines(probe, account_number, debit_amount) as (
+with probes(
+  probe,
+  defect,
+  source_type,
+  review_priority,
+  line_count,
+  observed_vat_rate,
+  document_has_12pct_vat,
+  document_has_25pct_vat,
+  expected_classification
+) as (
   values
-    ('hotel_on_5820', '5820', 1000.00::numeric),
-    ('hotel_on_5820', '2641', 120.00::numeric),
-    ('car_hire_on_5820', '5820', 1000.00::numeric),
-    ('car_hire_on_5820', '2641', 250.00::numeric),
-    ('representation_25_multi_line', '6072', 60.00::numeric),
-    ('representation_25_multi_line', '6072', 40.00::numeric),
-    ('representation_25_multi_line', '2641', 10.00::numeric),
-    ('representation_25_multi_line', '2641', 15.00::numeric),
-    ('representation_25_with_extra_vat', '6072', 100.00::numeric),
-    ('representation_25_with_extra_vat', '2641', 25.00::numeric),
-    ('representation_25_with_extra_vat', '2641', 20.00::numeric),
-    ('representation_12', '6072', 100.00::numeric),
-    ('representation_12', '2641', 12.00::numeric)
-),
-probe_context(probe, transaction_description, expected_result) as (
-  values
-    ('hotel_on_5820', 'Scandic Stockholm', 'high_hotel_counterparty_on_car_hire_account'),
-    ('car_hire_on_5820', 'Hertz rental car', 'manual_review_5820_no_corroborating_signal'),
-    ('representation_25_multi_line', 'Customer dinner', 'high_vat_is_25pct_of_6072_cost'),
-    ('representation_25_with_extra_vat', 'Mixed expense voucher', 'manual_review_6072_with_vat'),
-    ('representation_12', 'Customer dinner', 'manual_review_6072_with_vat')
-),
-line_groups as (
-  select probe, account_number, sum(debit_amount) as debit_amount
-  from probe_lines
-  group by probe, account_number
+    (
+      'exact_hotel_template_shape',
+      'travel_hotel_5820', 'manual',
+      'high_hotel_counterparty_on_car_hire_account',
+      3, 0.12::numeric, false, false, 'confirmed_correction'
+    ),
+    (
+      'car_hire_with_25pct_vat',
+      'travel_hotel_5820', 'manual',
+      'manual_review_5820_no_corroborating_signal',
+      3, 0.25::numeric, false, false, 'false_positive'
+    ),
+    (
+      'hotel_shape_without_hotel_evidence',
+      'travel_hotel_5820', 'manual',
+      'medium_12pct_vat_on_car_hire_account',
+      3, 0.12::numeric, false, false, 'insufficient_evidence'
+    ),
+    (
+      'imported_representation_signature',
+      'representation_25pct_vat', 'import',
+      'high_vat_is_25pct_of_6072_cost',
+      3, 0.25::numeric, false, false, 'false_positive'
+    ),
+    (
+      'live_representation_without_underlag',
+      'representation_25pct_vat', 'manual',
+      'high_vat_is_25pct_of_6072_cost',
+      3, 0.25::numeric, false, false, 'insufficient_evidence'
+    ),
+    (
+      'live_representation_with_25pct_underlag',
+      'representation_25pct_vat', 'manual',
+      'high_vat_is_25pct_of_6072_cost',
+      3, 0.25::numeric, false, true, 'false_positive'
+    ),
+    (
+      'live_representation_with_12pct_vat',
+      'representation_25pct_vat', 'manual',
+      'manual_review_6072_with_vat',
+      3, 0.12::numeric, true, false, 'false_positive'
+    ),
+    (
+      'mixed_representation_without_underlag',
+      'representation_25pct_vat', 'manual',
+      'manual_review_6072_with_vat',
+      4, 0.18::numeric, false, false, 'insufficient_evidence'
+    )
 ),
 classified as (
   select
-    context.probe,
-    context.expected_result,
+    probe.*,
     case
-      when hotel.debit_amount is not null
-        and context.transaction_description ~* '(hotel|hotell|scandic|elite|best western|nordic choice|clarion|quality inn|radisson|booking\.com|airbnb|logi|övernattning)'
-        then 'high_hotel_counterparty_on_car_hire_account'
-      when hotel.debit_amount is not null
-        and abs(coalesce(vat.debit_amount, 0) - hotel.debit_amount * 0.12) < 0.02
-        then 'medium_12pct_vat_on_car_hire_account'
-      when hotel.debit_amount is not null
-        then 'manual_review_5820_no_corroborating_signal'
-      when representation.debit_amount is not null
-        and abs(coalesce(vat.debit_amount, 0) - representation.debit_amount * 0.25) < 0.02
-        then 'high_vat_is_25pct_of_6072_cost'
-      when representation.debit_amount is not null and vat.debit_amount is not null
-        then 'manual_review_6072_with_vat'
-      else 'absent'
-    end as actual_result
-  from probe_context context
-  left join line_groups hotel
-    on hotel.probe = context.probe and hotel.account_number = '5820'
-  left join line_groups representation
-    on representation.probe = context.probe and representation.account_number = '6072'
-  left join line_groups vat
-    on vat.probe = context.probe and vat.account_number = '2641'
+      when source_type in ('import', 'correction') then 'false_positive'
+      when defect = 'representation_25pct_vat'
+        and review_priority = 'high_vat_is_25pct_of_6072_cost'
+        and document_has_25pct_vat
+        and not document_has_12pct_vat
+        then 'false_positive'
+      when defect = 'representation_25pct_vat'
+        and review_priority = 'high_vat_is_25pct_of_6072_cost'
+        then 'insufficient_evidence'
+      when defect = 'representation_25pct_vat'
+        and observed_vat_rate between 0.119 and 0.121
+        then 'false_positive'
+      when defect = 'representation_25pct_vat'
+        and document_has_12pct_vat
+        and not document_has_25pct_vat
+        then 'false_positive'
+      when defect = 'representation_25pct_vat' then 'insufficient_evidence'
+      when defect = 'travel_hotel_5820'
+        and line_count = 3
+        and observed_vat_rate between 0.119 and 0.121
+        and review_priority = 'high_hotel_counterparty_on_car_hire_account'
+        then 'confirmed_correction'
+      when defect = 'travel_hotel_5820'
+        and line_count = 3
+        and observed_vat_rate between 0.119 and 0.121
+        then 'insufficient_evidence'
+      else 'false_positive'
+    end as actual_classification
+  from probes probe
 )
 
 select
   probe,
-  expected_result,
-  actual_result,
-  actual_result = expected_result as passed
+  expected_classification,
+  actual_classification,
+  actual_classification = expected_classification as passed
 from classified
 order by probe;

scripts/audit-template-misbookings.sql

@@ -6,11 +6,12 @@
 -- named accounts that never existed in BAS, so account-backfill could not seed
 -- them and every booking through those templates failed. Nothing was posted.
 --
--- This query is diagnostic, not a list of confirmed errors. There is no
--- provenance link from a posted entry back to the template that produced it.
--- Both signatures also have legitimate shapes: 5820 is correct for actual car
--- hire, and representation can lawfully carry 25% VAT when the supply itself is
--- subject to 25% and the invoice is correct. Review the underlag before acting.
+-- The query separates template-remediation evidence from broad account
+-- signatures. Imported entries and earlier corrections cannot have been
+-- generated by either live template in this database. The remaining signatures
+-- can still have legitimate shapes: 5820 is correct for actual car hire, and
+-- representation can lawfully carry 25% VAT when the supply itself is subject
+-- to 25% and the invoice is correct. Review the underlag before acting.
 --
 -- The query performs no writes, creates no objects, and returns one row per
 -- posted journal entry and defect. Companion procedure:
@@ -36,9 +37,11 @@ candidate_entry_ids as (
   where account_number in ('5820', '6072')
 ),
 
-transaction_refs as (
+candidate_transactions as (
   select
     t.journal_entry_id,
+    t.id as transaction_id,
+    t.document_id,
     concat_ws(' ', nullif(t.merchant_name, ''), nullif(t.description, '')) as transaction_description
   from public.transactions t
   join candidate_entry_ids candidate on candidate.journal_entry_id = t.journal_entry_id
@@ -48,6 +51,8 @@ transaction_refs as (
 
   select
     tvl.journal_entry_id,
+    t.id as transaction_id,
+    t.document_id,
     concat_ws(' ', nullif(t.merchant_name, ''), nullif(t.description, '')) as transaction_description
   from public.transaction_voucher_links tvl
   join candidate_entry_ids candidate on candidate.journal_entry_id = tvl.journal_entry_id
@@ -61,7 +66,7 @@ transaction_context as (
       distinct transaction_description,
       ' | ' order by transaction_description
     ) as transaction_description
-  from transaction_refs
+  from candidate_transactions
   group by journal_entry_id
 ),
 
@@ -77,6 +82,7 @@ hotel_candidates as (
     je.entry_date,
     je.committed_at,
     je.fiscal_period_id,
+    je.source_type,
     cost.line_ids as cost_line_ids,
     coalesce(vat.line_ids, '{}'::uuid[]) as vat_line_ids,
     '5820'::text as observed_cost_account,
@@ -117,6 +123,7 @@ representation_candidates as (
     je.entry_date,
     je.committed_at,
     je.fiscal_period_id,
+    je.source_type,
     cost.line_ids as cost_line_ids,
     vat.line_ids as vat_line_ids,
     '6072'::text as observed_cost_account,
@@ -152,6 +159,7 @@ all_candidates as (
     entry_date,
     committed_at,
     fiscal_period_id,
+    source_type,
     cost_line_ids,
     vat_line_ids,
     observed_cost_account,
@@ -177,6 +185,7 @@ all_candidates as (
     entry_date,
     committed_at,
     fiscal_period_id,
+    source_type,
     cost_line_ids,
     vat_line_ids,
     observed_cost_account,
@@ -217,18 +226,310 @@ entry_line_snapshots as (
     from all_candidates
   ) candidate on candidate.journal_entry_id = jel.journal_entry_id
   group by jel.journal_entry_id
+),
+
+candidate_documents as (
+  -- Resolve every current underlag path that can legally support the entry:
+  -- direct entry or line links, transaction links, and supplier-invoice
+  -- references. Return document ids only, never storage paths or file names.
+  select distinct
+    c.journal_entry_id,
+    d.id as document_id,
+    d.extracted_data
+  from all_candidates c
+  join public.document_attachments d
+    on d.is_current_version = true
+   and (
+     d.journal_entry_id = c.journal_entry_id
+     or exists (
+       select 1
+       from public.journal_entry_lines jel
+       where jel.journal_entry_id = c.journal_entry_id
+         and jel.id = d.journal_entry_line_id
+     )
+   )
+
+  union
+
+  select distinct
+    tx.journal_entry_id,
+    d.id as document_id,
+    d.extracted_data
+  from candidate_transactions tx
+  join public.document_attachments d
+    on d.id = tx.document_id
+   and d.is_current_version = true
+
+  union
+
+  select distinct
+    c.journal_entry_id,
+    d.id as document_id,
+    d.extracted_data
+  from all_candidates c
+  join public.supplier_invoices si
+    on si.registration_journal_entry_id = c.journal_entry_id
+    or si.payment_journal_entry_id = c.journal_entry_id
+    or exists (
+      select 1
+      from public.supplier_invoice_payments sip
+      where sip.supplier_invoice_id = si.id
+        and sip.journal_entry_id = c.journal_entry_id
+    )
+  join public.document_attachments d
+    on d.id = si.document_id
+   and d.is_current_version = true
+),
+
+candidate_document_vat_rates as (
+  -- Extraction payloads are historical data and may have missing, malformed,
+  -- or differently scaled rate values. Accept only numeric values, normalizing
+  -- both 0.25 and 25 to percent before they reach the classifier.
+  select
+    cd.journal_entry_id,
+    cd.document_id,
+    case
+      when rate.raw_rate::numeric <= 1 then rate.raw_rate::numeric * 100
+      else rate.raw_rate::numeric
+    end as normalized_rate
+  from candidate_documents cd
+  cross join lateral (
+    select vat ->> 'rate' as raw_rate
+    from jsonb_array_elements(
+      case
+        when jsonb_typeof(cd.extracted_data -> 'vatBreakdown') = 'array'
+          then cd.extracted_data -> 'vatBreakdown'
+        else '[]'::jsonb
+      end
+    ) vat
+
+    union all
+
+    select item ->> 'vatRate' as raw_rate
+    from jsonb_array_elements(
+      case
+        when jsonb_typeof(cd.extracted_data -> 'lineItems') = 'array'
+          then cd.extracted_data -> 'lineItems'
+        else '[]'::jsonb
+      end
+    ) item
+  ) rate
+  where rate.raw_rate ~ '^\s*[0-9]+(?:\.[0-9]+)?\s*$'
+),
+
+document_signals as (
+  select
+    cd.journal_entry_id,
+    array_agg(distinct cd.document_id order by cd.document_id) as document_ids,
+    count(distinct cd.document_id)::integer as document_count,
+    count(distinct cd.document_id) filter (
+      where cd.extracted_data is not null
+    )::integer as extracted_document_count,
+    max(
+      case
+        when cd.extracted_data ->> 'confidence'
+          ~ '^\s*[0-9]+(?:\.[0-9]+)?\s*$'
+          then (cd.extracted_data ->> 'confidence')::numeric
+      end
+    ) as max_extraction_confidence,
+    bool_or(rate.normalized_rate between 11.9 and 12.1)
+      as document_has_12pct_vat,
+    bool_or(rate.normalized_rate between 24.9 and 25.1)
+      as document_has_25pct_vat
+  from candidate_documents cd
+  left join candidate_document_vat_rates rate
+    on rate.journal_entry_id = cd.journal_entry_id
+   and rate.document_id = cd.document_id
+  group by cd.journal_entry_id
+),
+
+candidate_output as (
+  select
+    c.*,
+    co.name as company_name,
+    lines.original_lines,
+    jsonb_array_length(lines.original_lines) as line_count,
+    coalesce(doc.document_ids, '{}'::uuid[]) as document_ids,
+    coalesce(doc.document_count, 0) as document_count,
+    coalesce(doc.extracted_document_count, 0) as extracted_document_count,
+    doc.max_extraction_confidence,
+    coalesce(doc.document_has_12pct_vat, false) as document_has_12pct_vat,
+    coalesce(doc.document_has_25pct_vat, false) as document_has_25pct_vat,
+    case
+      when c.defect = 'representation_25pct_vat'
+        then 'participants_purpose_and_300_sek_vat_base_cap_required'
+      else 'not_applicable'
+    end as independent_representation_review,
+    c.defect = 'representation_25pct_vat'
+      and c.cost_debit_amount > 300
+      as observed_6072_cost_exceeds_300_sek,
+    reporting.vat_reporting_period,
+    coalesce(filing.deadline_statuses, '{}'::text[]) as vat_deadline_statuses,
+    case
+      when c.defect <> 'representation_25pct_vat' then 'not_applicable'
+      when reporting.vat_reporting_period is null
+        then 'unknown_reporting_period'
+      when coalesce(filing.has_confirmed_record, false)
+        then 'confirmed_in_app'
+      when coalesce(filing.has_submitted_record, false)
+        then 'submitted_in_app'
+      else 'not_proven_by_in_app_records'
+    end as vat_filing_status,
+    case
+      when fp.id is null then 'missing_fiscal_period'
+      when fp.is_closed then 'closed'
+      when fp.locked_at is not null then 'locked'
+      when cs.bookkeeping_locked_through is not null
+        and c.entry_date <= cs.bookkeeping_locked_through then 'behind_company_lock_date'
+      else 'open'
+    end as effective_lock_status
+  from all_candidates c
+  join public.companies co on co.id = c.company_id
+  join entry_line_snapshots lines on lines.journal_entry_id = c.journal_entry_id
+  left join document_signals doc on doc.journal_entry_id = c.journal_entry_id
+  left join public.company_settings cs on cs.company_id = c.company_id
+  left join public.fiscal_periods fp on fp.id = c.fiscal_period_id
+  left join lateral (
+    select case
+      when c.defect <> 'representation_25pct_vat' then null
+      when cs.moms_period = 'monthly'
+        then to_char(c.entry_date, 'YYYY-MM')
+      when cs.moms_period = 'quarterly'
+        then concat(
+          extract(year from c.entry_date)::integer,
+          '-Q',
+          ceil(extract(month from c.entry_date) / 3.0)::integer
+        )
+      when cs.moms_period = 'yearly' and fp.period_end is not null
+        then case
+          when extract(month from fp.period_end) = 12
+            then extract(year from fp.period_end)::integer::text
+          else concat(
+            extract(year from fp.period_end)::integer - 1,
+            '/',
+            extract(year from fp.period_end)::integer
+          )
+        end
+      else null
+    end as vat_reporting_period
+  ) reporting on true
+  left join lateral (
+    select
+      array_agg(distinct d.status order by d.status) as deadline_statuses,
+      bool_or(d.status = 'submitted') as has_submitted_record,
+      bool_or(d.status = 'confirmed') as has_confirmed_record
+    from public.deadlines d
+    where c.defect = 'representation_25pct_vat'
+      and d.company_id = c.company_id
+      and d.tax_deadline_type = case cs.moms_period
+        when 'monthly' then 'moms_monthly'
+        when 'quarterly' then 'moms_quarterly'
+        when 'yearly' then 'moms_yearly'
+      end
+      and d.tax_period = reporting.vat_reporting_period
+  ) filing on true
+),
+
+classified as (
+  select
+    candidate.*,
+    case
+      -- Imports and earlier corrections were not generated by either live
+      -- template in this database. They can have independent accounting
+      -- issues, but are false positives for this template-remediation audit.
+      when candidate.source_type in ('import', 'correction')
+        then 'false_positive'
+      when candidate.defect = 'representation_25pct_vat'
+        and candidate.review_priority = 'high_vat_is_25pct_of_6072_cost'
+        and candidate.document_has_25pct_vat
+        and not candidate.document_has_12pct_vat
+        then 'false_positive'
+      when candidate.defect = 'representation_25pct_vat'
+        and candidate.review_priority = 'high_vat_is_25pct_of_6072_cost'
+        then 'insufficient_evidence'
+      when candidate.defect = 'representation_25pct_vat'
+        and candidate.observed_vat_rate between 0.119 and 0.121
+        then 'false_positive'
+      when candidate.defect = 'representation_25pct_vat'
+        and candidate.document_has_12pct_vat
+        and not candidate.document_has_25pct_vat
+        then 'false_positive'
+      when candidate.defect = 'representation_25pct_vat'
+        then 'insufficient_evidence'
+      -- The defective hotel template deterministically emitted three lines:
+      -- 5820 cost, 2641 at 12 percent, and one settlement credit. The hotel
+      -- counterparty signal supplies the final discriminator from car hire.
+      when candidate.defect = 'travel_hotel_5820'
+        and candidate.line_count = 3
+        and candidate.observed_vat_rate between 0.119 and 0.121
+        and candidate.review_priority = 'high_hotel_counterparty_on_car_hire_account'
+        then 'confirmed_correction'
+      when candidate.defect = 'travel_hotel_5820'
+        and candidate.line_count = 3
+        and candidate.observed_vat_rate between 0.119 and 0.121
+        then 'insufficient_evidence'
+      else 'false_positive'
+    end as evidence_classification,
+    case
+      when candidate.source_type in ('import', 'correction')
+        then 'entry source cannot be either live template'
+      when candidate.defect = 'representation_25pct_vat'
+        and candidate.review_priority = 'high_vat_is_25pct_of_6072_cost'
+        and candidate.document_has_25pct_vat
+        and not candidate.document_has_12pct_vat
+        then 'extracted underlag confirms a 25 percent supply'
+      when candidate.defect = 'representation_25pct_vat'
+        and candidate.review_priority = 'high_vat_is_25pct_of_6072_cost'
+        then 'exact defective signature but no decisive underlag or provenance'
+      when candidate.defect = 'representation_25pct_vat'
+        and candidate.observed_vat_rate between 0.119 and 0.121
+        then 'voucher has the intended 12 percent aggregate VAT signature'
+      when candidate.defect = 'representation_25pct_vat'
+        and candidate.document_has_12pct_vat
+        and not candidate.document_has_25pct_vat
+        then 'extracted underlag confirms a 12 percent supply'
+      when candidate.defect = 'representation_25pct_vat'
+        then 'mixed representation voucher needs decisive underlag or provenance'
+      when candidate.defect = 'travel_hotel_5820'
+        and candidate.line_count = 3
+        and candidate.observed_vat_rate between 0.119 and 0.121
+        and candidate.review_priority = 'high_hotel_counterparty_on_car_hire_account'
+        then 'exact defective hotel shape plus hotel counterparty'
+      when candidate.defect = 'travel_hotel_5820'
+        and candidate.line_count = 3
+        and candidate.observed_vat_rate between 0.119 and 0.121
+        then 'exact defective hotel shape without hotel evidence'
+      else 'voucher does not have the defective hotel template shape'
+    end as classification_reason
+  from candidate_output candidate
 )
 
 select
   c.defect,
+  c.evidence_classification,
+  c.classification_reason,
+  case
+    when c.evidence_classification = 'confirmed_correction'
+      then 'explicit_write_approval_required_after_underlag_review'
+    when c.evidence_classification = 'insufficient_evidence'
+      then 'underlag_required_before_approval_request'
+    else 'no_template_remediation'
+  end as approval_status,
+  case
+    when c.defect = 'representation_25pct_vat'
+      and c.evidence_classification <> 'false_positive'
+      then 'potential_if_confirmed'
+    else 'none_from_this_defect'
+  end as vat_return_impact,
   c.company_id,
-  co.name as company_name,
+  c.company_name,
   c.journal_entry_id,
   c.voucher_series,
   c.voucher_number,
   c.entry_date,
   c.committed_at,
   c.fiscal_period_id,
+  c.source_type,
   c.cost_line_ids,
   c.vat_line_ids,
   c.observed_cost_account,
@@ -241,21 +542,27 @@ select
   c.expected_vat_rate,
   c.transaction_description,
   c.review_priority,
-  lines.original_lines,
-  case
-    when fp.id is null then 'missing_fiscal_period'
-    when fp.is_closed then 'closed'
-    when fp.locked_at is not null then 'locked'
-    when cs.bookkeeping_locked_through is not null
-      and c.entry_date <= cs.bookkeeping_locked_through then 'behind_company_lock_date'
-    else 'open'
-  end as effective_lock_status
-from all_candidates c
-join public.companies co on co.id = c.company_id
-join entry_line_snapshots lines on lines.journal_entry_id = c.journal_entry_id
-left join public.company_settings cs on cs.company_id = c.company_id
-left join public.fiscal_periods fp on fp.id = c.fiscal_period_id
+  c.line_count,
+  c.document_ids,
+  c.document_count,
+  c.extracted_document_count,
+  c.max_extraction_confidence,
+  c.document_has_12pct_vat,
+  c.document_has_25pct_vat,
+  c.independent_representation_review,
+  c.observed_6072_cost_exceeds_300_sek,
+  c.vat_reporting_period,
+  c.vat_deadline_statuses,
+  c.vat_filing_status,
+  c.original_lines,
+  c.effective_lock_status
+from classified c
 order by
+  case c.evidence_classification
+    when 'confirmed_correction' then 0
+    when 'insufficient_evidence' then 1
+    else 2
+  end,
   c.defect,
   case
     when c.review_priority like 'high%' then 0