a652dcae1ae30209cd75dbdf1de98ec2d265ed88
142
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
|
|
a652dcae1a |
Skv/e2e overview (#515)
* feat(agi): refactor AGI XML generation and data handling - Remove deprecated AGI field codes from field-codes.ts. - Update generate-declaration.ts to include new employee fields and handle absence data with stable specification numbers. - Enhance XML generation in xml-generator.ts to support new flags for housing benefits and adjusted benefits. - Introduce new database migrations to support: - `removed_from_agi` flag for tombstoning individuppgifter. - `benefits_adjusted` flag for tracking adjustments to benefits. - `franvaro_specifikationsnummer` for stable absence event identification. - `housing_benefit_type` to differentiate between housing benefit types. * feat: Implement strict validation for AGI employee data and introduce pre-flight validation schemas - Added Zod schemas for validating employee data in AGI declarations to ensure all required fields are present and correctly typed, preventing silent errors during processing. - Introduced AGI pre-flight validation schemas for Skatteverket endpoints to validate individual and head unit submissions before sending to the API. - Created a new audit log table for tracking all outbound calls to Skatteverket, ensuring compliance and traceability for AGI and moms submissions. - Implemented advisory locks in the database to manage concurrent updates to absence specification numbers, enhancing data integrity. - Added compliance documentation for GDPR processing activities related to AGI and moms submissions, detailing data handling and retention policies. * feat: Extend DELETE RLS policy to protect 'declined' signatures in årsredovisning * fix: Update date handling in salary absence migrations to use immutable year-month key * fix: Update SELECT policy in skatteverket_api_audit_log to use IN clause for company_id * feat: Add skatteverket_api_audit_log and salary_absence_franvaro_audit tables with RLS policies and immutable triggers |
||
|
|
831920fede |
fix(arcim): allow Fortnox re-sync to replace prior SIE import per fiscal year (#512)
* fix(arcim): allow Fortnox re-sync to replace prior SIE import per fiscal year A user reported a sync failure when retrying Fortnox after adding more verifications: Import failed: Failed to create pending import record: duplicate key value violates unique constraint "sie_imports_company_id_file_hash_active_idx" Root cause: Fortnox embeds the export-time #GEN date in every SIE export, so the file hash always differs between syncs. The wizard's hash-based duplicate detection treated each sync as a brand new file, but the engine still rejected the insert because the per-period import slot was held by the prior 'completed' row. This change reframes Fortnox re-sync as a replace operation rather than a fresh import: - New executeSIEImport option `onExistingPeriod: 'block' | 'replace'`. Manual SIE upload at /api/import/sie keeps default 'block' (current behavior, no regression). The Fortnox /import-sie endpoint passes 'replace', which runs replaceSIEImport on any overlapping completed import before insert. Imported journal entries from the prior import are cancelled per BFL 5 kap 5§; user-created entries (manual, transaction, invoice) are untouched. - /sie-data switches from hash-based to period-based duplicate detection and returns previousImport metadata per fiscal year. - Wizard drops the alreadyImported skip filter, surfaces an amber callout in the confirm dialog listing fiscal years that will be replaced, and shows "ersatte N tidigare importerade verifikationer" per year. - createPendingImportRecord translates 23505 partial-index violations to a clear Swedish recovery message instead of leaking the raw constraint name. - cleanupStaleImportRecords drops the 1-hour age gate and also cleans status='mapped' orphans. SIE imports are single-flight per company so the gate just made legitimate retries fail. - After replace, the fiscal_periods row's opening_balances_set and opening_balance_entry_id are cleared (only when they pointed at the cancelled prior IB entry), so the new IB import isn't skipped. Schema-drift migration captures the partial unique index sie_imports_company_id_file_hash_active_idx that already exists in production (added out-of-band) and drops the now-superseded plain sie_imports_company_id_file_hash_key constraint. Both statements are idempotent — verified no-op against production. Tests: new pg-real test covers the partial index admit-replaced semantics, source_type='import'-only cancellation in replace_sie_import, and the post-replace insert path. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(arcim): address PR review — restore 5-min cleanup gate, fix test source_type Greptile P2: `cleanupStaleImportRecords` was deleting `pending` rows unconditionally, which could wipe a concurrent in-flight import in another tab/session. Restored a 5-minute age gate (long enough for any normal interactive import, short enough that legitimate retries after a crash still succeed). Also dropped `mapped` from the cleanup — it is defined in SIEImportStatus but no code path writes it, so including it was both unnecessary and added the concurrent-session risk Greptile flagged. pg-real test: insertPostedEntry used `source_type='transaction'` which is not a valid value per the journal_entries_source_type_check constraint (migration 20260516060000). Switched to `'bank_transaction'` — the actual source_type emitted when a user categorizes a bank transaction in gnubok. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
fd4274787c |
feat(bokslut): Phases 4-7 — accruals + årsredovisning + EF + MCP tools (#509)
* feat(bokslut): Phases 4-7 — accruals + årsredovisning + EF + MCP tools Builds on the Phase 1-3 PR (#508). Same K2-AB scope decisions hold (iXBRL deferred, K3 deferred, koncernredovisning out of scope) — what ships: PHASE 4 — Periodiseringar - lib/bokslut/accruals/accrual-detector.ts: auto-proposes the vacation- liability change (delta on 2920 against 7090 + 31,42 % avgifter on 7519/2940) by reading lib/reports/vacation-liability.ts. Manual prepaid (17xx) / accrued (29xx) / audit-fee builders for entries the heuristic can't derive — supplier-invoice service_period detection is deferred until the data model grows the field. - /api/bookkeeping/fiscal-periods/[id]/accruals (GET + POST) — posts each accrual as a separate manual-source journal entry with the next-day reverse date embedded in the description. - New AccrualsStep wizard step between Preflight and Dispositions. - Auto-reversal cron is follow-up infra; for now reverses_on is metadata + a visible UI badge. PHASE 5 — Årsredovisning PDF + signing - lib/bokslut/arsredovisning/{types,build-data,arsredovisning-pdf,signature -service}: pre-fills flerårsöversikt from prior 3 fiscal periods, eget- kapital-förändring from journal data, K2-minimum noter with avskrivningstider auto-derived from the asset register and medelantal anställda from the employees table. PDF via @react-pdf/renderer mirroring the income-statement/pdf pattern. - Migration 20260516170000: arsredovisning_signature_requests with RLS, signed-immutability trigger, and DELETE policy that blocks signed rows. Signature-service exposes list/create/markSigned/isFullySignedOff. - BankID call itself is not wired here — the table + service make the request layer available so a follow-up can hook lib/auth/bankid.ts to the sign action without rework. - /api/.../arsredovisning (data + pdf + signatures) endpoints. - /bookkeeping/year-end/arsredovisning page with editable narrative, flerårsöversikt table, signer slots, PDF download, and Bolagsverket Mina Sidor link. Explicit warning about the FY2026 iXBRL mandate. PHASE 6 — Enskild firma NE-bilaga UI - lib/bokslut/enskild-firma/{egenavgifter,rantefordelning,periodiseringsfond -ef,expansionsfond}-calculator.ts. All declaration-only — never produce a journal entry. Egenavgifter 28,97 % / pensionärssats 10,21 % / passive SLP 24,26 %, schablonavdrag 25/10/20 %. Räntefördelning SLR+6 / SLR+1 with the -500 000 negative threshold. P-fond EF cap 30 % (vs 25 % för AB). Expansionsfond 125,94 % av kapitalunderlag, 20,6 % skatt. - EfDeclarationSection mounted inside DispositionsStep when entity_type is enskild_firma — live recompute as the user adjusts kapitalunderlag, prior-year amounts, p-fond desired, expansionsfond change. Each card shows the NE-bilaga ruta the number lands in. NE-bilaga preview link. PHASE 7 — Agent-native MCP tools - gnubok_propose_dispositioner: read-only AB dispositions proposal - gnubok_propose_accruals: read-only accruals proposal - gnubok_propose_annual_depreciation: read-only depreciation per asset - gnubok_post_annual_depreciation: stages depreciation commit (high-risk) - gnubok_preview_arsredovisning: structured K2 ÅR preview - gnubok_preview_ef_declaration: EF skattemässiga justeringar preview - All ≤280-char descriptions, additionalProperties:false, conform to STAGED_OPERATION_SCHEMA for write tools. Read tools return the same shapes as their HTTP counterparts so agents and the UI share a contract. - The existing GET /bokslutsdispositioner endpoint now calls the shared buildDispositionsProposal helper that the MCP tool also uses, removing the duplicate logic that lived in both. Verification - 149 unit tests pass (was 125 on Phase 1-3; +24 across accruals, EF calculators, expansionsfond, periodiseringsfond-ef) - Zero lint or typecheck errors on any new file - Migration 20260516170000 applied to remote Supabase via MCP Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(bokslut): address PR #509 round-1 — CI blocker + 3 P1s + compliance bundle CI BLOCKER (output-schema test) - The 5 new read-only MCP tools (gnubok_propose_dispositioner / _accruals / _annual_depreciation, _preview_arsredovisning / _ef_declaration) were missing outputSchema, which the existing strict-schemas guard reads as a hard failure on core-only CI. Added a permissive { type: 'object', additionalProperties: true } outputSchema to each — the return shapes are the same as their HTTP counterparts and trying to mirror them inline would duplicate the type tree across two boundaries. P1 — vacation accrual: two real bugs in one entry - Delta was anchored on the OPENING balance of 2920, so any mid-year postings (partial accruals, reversals) were ignored. Now anchors on the current closing balance via tb.rows.find(2920).closing_credit - closing_debit. Updated the computation field label to current_2920. - More importantly: 2920 is a balance-sheet carry-forward (semesterlöneskuld persists until the actual vacation is paid). The original implementation set reverses_on to Jan 1 of the next year, which would zero the liability on day 1 of the new year — a known Swedish bookkeeping error. The vacation proposal now ships with reverses_on = '' to suppress the reversal badge, the API route emits a "Bokslutsjustering" description instead of "Periodisering (vänds …)", and AccrualsStep renders "Rullas vidare (ingen vändning)" so the user knows the liability carries forward. P1 — signature_requests immutability gap (security) - The existing trigger only guarded role / signer_name / signed_at / status on signed rows, leaving bankid_signature_data, signer_personnummer_encrypted, and signer_personnummer_hash mutable. An UPDATE on a signed row could silently alter the BankID proof. New migration 20260517090000 replaces the trigger function to cover the full audit-critical column set (plus fiscal_period_id and company_id as belt-and-braces). P1 — narrative edits never reached the PDF - The /bookkeeping/year-end/arsredovisning page let the user edit description / important_events / resultatdisposition but the download link pointed at a plain GET that regenerated boilerplate. Wired the PDF endpoint to accept description / events / disposition as query params (length-capped) and the page now constructs the download URL with the current narrative state — only fields the user actually changed are included, keeping the URL short for the unchanged-defaults case. Compliance quick wins - Added period_lock check to gnubok_post_annual_depreciation MCP tool (matches the existing accruals POST guard). - Added explicit fiscal-period ownership pre-check to the signatures POST route (RLS would reject anyway; the route layer just makes the 404 envelope cleaner). - Replaced free-text role on the signatures schema with an enum allowing only Styrelseledamot / Styrelseordförande / VD / Verkställande direktör. - Added Cache-Control: no-store + Pragma: no-cache to the ÅR PDF response so the document (officer names + financials = personal data) isn't cached by any intermediary. - Sanitized period_end in the PDF Content-Disposition header to dodge header-injection via stray chars (defensive — period_end is a date, but the cost is one regex). - Softened the iXBRL warning text on the ÅR page: digital filing is proposed by Bolagsverket but not yet enacted; PDF is still valid today. False positives I'm intentionally not chasing on this round - Greptile P2 `sate` typo in the address city field — the rename would touch the type and every consumer; defer. - Greptile P2 "computation field label" — already addressed as part of the P1 vacation fix above. - Compliance V2.2 "silent skip" in accruals POST — the silent skip is in the UI's empty-row filter, not in the server route. Server validation already returns 400 via Zod for any invalid item. Verification - 154 unit tests pass (was 149; +5 from re-running the MCP strict-schemas + output-schema suites that now include the new tools) - Zero typecheck errors on any touched file - Zero lint errors on any touched file - Migration 20260517090000 applied to remote Supabase via MCP Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(bokslut): address PR #509 round-2 — 3 real bugs + 5-item polish 3 real bugs from the round-1 Swedish review + Compliance Swarm re-eval: - Soliditet inflated by obeskattade reserver. build-data.ts filtered equity with startsWith('20') || startsWith('21') — but 21xx (periodiseringsfonder, överavskrivningar) are partially deferred tax, not eget kapital. Splitting them out per K2 / ÅRL. Filter is now startsWith('20') only. - Resultaträkning omitted bokslutsdispositioner + skatt rows. K2 RR must include 88xx (dispositioner) and 89xx (skatt) before "Årets resultat" per ÅRL 3:2 — without them, the printed RR doesn't reconcile to BS 2099 and the document is non-compliant for any AB that posted bolagsskatt or periodiseringsfond. flattenIncomeStatement now splits the financial sections on title (Bokslutsdispositioner / Skatter och årets resultat) and emits the K2-required intermediate subtotals: "Resultat efter finansiella poster" → dispositioner → "Resultat före skatt" → skatt → "Årets resultat". - Accruals POST had no idempotency. Re-running the wizard (or a retried POST after a flaky network) would create duplicate accrual entries that distort both the balance sheet and trial balance. New findExistingAccrualEntry helper queries the period for an existing posted entry whose description matches the kind's stable prefix (or for manual prepaid/accrued, the user-supplied description). Duplicates land in a new `skipped` array in the response with a reference to the existing entry id, rather than producing a second posting. 5-item compliance polish: - signer_name on the signatures schema now has .max(200) per GDPR Art.25.2 data-minimization. Swedish personal names are well under that — the bound is a defense against an unbounded-string injection. - Audit-fee accrual: 6420 is BAS-specific to lagstadgad revision. Bokslut fees for a non-revisionspliktigt bolag (liability_account = 2991) now debit 6590 (övriga externa tjänster) instead — Skatteverket may query a 6420 debit when there is no revisor i bolaget. - Räntefördelning ne_ruta label: was 'R30 / INK1 kapital' (confusing — INK1 is a separate form). Positive now reads 'R30 (avdrag i näringsverksamhet)' with the INK1 T4 cross-reference moved to the description; negative reads 'R30 (tillägg till resultat)'. - gnubok_post_annual_depreciation MCP tool now checks the caller's company_members role and throws on viewer. RLS would reject the underlying INSERT anyway; failing fast here produces a cleaner error than the cascaded RLS rejection. Mirrors the HTTP route's { requireWrite: true } guard. - Signature DELETE policy now blocks both 'signed' AND 'declined' rows. A declined signature is auditable evidence (board member refused) and is material under ABL 8 kap. New migration 20260517100000. Bot-flagged items I'm deliberately not chasing on this round: - V8.2.1 × 2 cross-tenant findings on ÅR routes — same false-positive class I've responded to repeatedly: buildArsredovisningData internally filters by company_id, bot can't see past the route handler. - V4.5 / V2.2 MCP arg Zod redundancy — MCP server's central handler validates against each tool's inputSchema. - Narrative-in-URL GDPR concerns — proper fix is POST + body or server-side persistence; tracked as follow-up. - `sate` typo (Greptile P2) — type-wide rename, deferred. - Vacation avgifter age-tier rate split — inherited from upstream generateVacationLiability which doesn't expose age; needs upstream work. - Medelantal anställda proper monthly average — needs salary-run aggregation across the year, follow-up. - K2 noter aktiekapital + fastställelseintyg blocks — real K2 gaps, tracked as follow-up. Verification - 94 tests pass (subset for bokslut + MCP suites; full Phase 1-7 suite unchanged in scope) - Zero typecheck errors on any touched file - Zero new lint errors on any touched file (the 2 server.ts warnings are pre-existing) - Migration 20260517100000 applied to remote Supabase via MCP Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
c06395f633 |
feat(mcp): agent-native API sprint — quick wins (items 8/10/38/39/50) (#505)
* feat(mcp): agent-native API sprint — quick wins (items 8/10/38/39/50)
Five Tier-S items from dev_docs/api_ai_architecture/PLAN.md, picked for highest
impact-per-day on a solo budget. ~7.5 engineer-days of work.
Item 38 — gnubok_reverse_journal_entry MCP tool. Wraps the existing
reverseEntry() engine function (lib/bookkeeping/engine.ts) as a staged
high-risk operation. Description distinguishes pure makulering (use this) from
rättelse (use gnubok_correct_entry) per BFL 5 kap 5§ guidance — leaving a real
affärshändelse unbooked is itself a BFL violation, so agents must understand
which storno pattern to apply. New operation_type 'reverse_entry' wired through
PendingOperationType, risk-tiers (high), commit.ts executor, and TOOL_SCOPE_MAP
(bookkeeping:write). Six executor cases + three staging-gate cases cover the
new tool.
Item 39 — period_status threading. New helper resolvePeriodStatusForDate() in
lib/core/bookkeeping/period-service.ts returns { period_id, status, lock_date }
using the same two-layer logic as the v1 REST check (company-wide
bookkeeping_locked_through + fiscal_period flags). Threaded through
stagePendingOperation via a new dateForPeriodCheck option so agents and widgets
can detect locked/closed periods without round-trips. Applied to seven
bookkeeping-touching tools: categorize_transaction, create_transactions,
create_voucher, approve_supplier_invoice, mark_invoice_as_paid, correct_entry,
reverse_journal_entry. Resolution failure is non-fatal — DB triggers stay
authoritative.
Item 50 — gnubok://company/current expansion. Replaces the metadata-only
resource with per-company working memory: active fiscal period status, lock
dates, counts (customers, suppliers, open AR/AP, uncategorized transactions),
voucher series state across open periods, recency signals (last categorization,
last invoice sent, last bank sync), and the next five approaching deadlines.
All queries parallelized via Promise.all; payload stays well under 8 KB.
Mirrors the context.md pattern from Shipper+Claude's agent-native architecture
guidance and prevents the context-starvation anti-pattern.
Item 8 — schema strictness. additionalProperties: false on every one of the 67
inputSchemas in extensions/general/mcp-server/server.ts. New
strict-schemas.test.ts guards against regression on newly authored tools.
CLAUDE.md documents the tool-authoring contract (strict input schemas,
description ≤280 chars, STAGED_OPERATION_SCHEMA + next as the
completion-signal pattern — do NOT introduce a parallel S/H/C/O envelope).
Payload-size ceiling raised from 20K → 25K tokens with a comment pointing at
item 15 (Tool Search + defer_loading) as the long-term answer rather than
relaxing the watchdog further.
Item 10 — prompt cache groundwork. The only Anthropic SDK call site in the
codebase is the invoice-inbox extension's Bedrock-backed extractor; tagged the
~3.5 KB SYSTEM_PROMPT with cache_control: { type: 'ephemeral' } and added
usage logging (cache_read_input_tokens / cache_creation_input_tokens) so the
hit ratio is measurable. The plan's 1h TTL is direct-Anthropic-only;
documented the constraint and the MCP-side determinism contract (tool
definitions must be byte-stable across requests) in the new mcp-server
README.md.
Carry-over: includes a small untracked migration
(20260516060000_journal_entries_source_type_inbox_item) and its pg test guard
that fix a production CHECK-constraint gap for source_type='inbox_item' —
unrelated to the sprint but bundled per request.
Tests: 3615/3615 pass across 252 files. TypeScript build clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(mcp): address PR #505 review — cross-tenant leaks, company-wide lock, PII
Five reviewer findings on PR #505 addressed:
1. Cross-tenant leak — voucher_sequences (OWASP V8.2.1, SOC 2 CC6.3).
Resource query filtered by user_id only; switched to company_id since the
table has both (added in the 2026-03 multi-tenant refactor migration).
2. Cross-tenant leak — deadlines (OWASP V8.2.1, GDPR Art.5(1)(f), ISO A.8.3).
Same fix; the deadlines table also gained a company_id column in the
multi-tenant refactor and the RLS policies enforce it. With the company_id
filter active, the userId parameter is no longer needed in the resource —
removed from the destructure.
3. Compliance gap — commitReverseEntry and commitCorrectEntry only checked
fiscal_periods.is_closed, not company_settings.bookkeeping_locked_through.
Agents could stage a reversal with period_status: locked warning (caught
by resolvePeriodStatusForDate at staging time), have the user approve,
and the commit would slip through. Both executors now run
resolvePeriodStatusForDate at commit time so the gate matches the
staging-time signal. Pre-existing gap on commitCorrectEntry also fixed.
4. Schema mismatch — period_status was spread into both `preview` and the
top-level response, but STAGED_OPERATION_SCHEMA only declares it at the
top level. Removed the preview-nested copy to match the schema and avoid
ambiguous reads.
5. Tool description — swedish-compliance bot flagged that "pure makulering
(storno)" conflates two distinct Swedish accounting terms: storno
preserves the original; makulering voids it entirely. Code does storno;
description now says so plainly and cites BFL 5 kap.
6. Input hardening — added ^\d{4}-\d{2}-\d{2}$ pattern to reversal_date in
inputSchema plus a runtime regex check in execute(), so a malformed date
never reaches the pending_operations payload.
7. GDPR — ai_extraction_usage and the two pre-existing fileName log
emissions in extract-invoice-fields.ts replaced raw fileName with a
12-char SHA-256 prefix. Raw invoice file names (e.g.
"faktura_Sven_Andersson.pdf") can constitute personal data; hashing
preserves operator correlation without exposing PII to log destinations
that may lack documented retention controls.
Notes on findings NOT addressed:
- Double-reversal guard (Greptile/swedish-compliance): false positive.
reverseEntry() flips the original's status to 'reversed' (engine.ts:538)
and the staging tool already rejects anything not 'posted'. Engine also
has a CAS guard at lines 541-551.
- Staging vs commit TOCTOU re-validation: pre-flight + DB triggers remain
authoritative; the window is narrow enough that adding executor-side
re-checks isn't load-bearing this sprint.
- Runtime Zod validation of args inside execute(): codebase doesn't do
this for any MCP tool today; cross-cutting refactor deferred.
New test: voucher-executors.test.ts adds a case for the company-wide lock
branch on reverse_entry (verifies the new resolvePeriodStatusForDate gate
fires when bookkeeping_locked_through covers entry_date).
Tests: 3616/3616 pass. TypeScript build clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(mcp): address second-round PR #505 review — locked_at, reason cap, log
Re-review by compliance-swarm and swedish-accounting-compliance bots after the
first fixes raised three more legitimate findings:
1. Per-period `locked_at` not directly checked from the fetched row
(swedish-accounting-compliance). Both commitCorrectEntry and
commitReverseEntry already call resolvePeriodStatusForDate which covers
locked_at, but a transient DB blip in the resolve helper would silently
skip that gate. Now reading locked_at directly from the inner-join row and
checking it alongside is_closed before the resolve helper runs — same
pattern, two defense-in-depth layers instead of one.
2. `reason` field had no maxLength (OWASP V4.5). Added maxLength: 500 to the
inputSchema and a runtime length check; an adversarial agent could
otherwise push an arbitrarily large string into pending_operations.
3. periodStatus resolution failure was silently swallowed (ISO 27001 A.8.15).
Now logging via console.warn with operationType, companyId,
dateForPeriodCheck, and error so a systematic outage (missing
company_settings row, dropped query) is observable in audit logs rather
than degraded silently.
Findings deliberately NOT addressed (pushed back to the bots):
- gnubok_reverse_journal_entry needs per-operation role check (V8.2.1) and
narrower 'bookkeeping:reverse' scope (CC6.3) — cross-cutting refactor; no
MCP tool in gnubok enforces per-operation roles today. Introducing it just
for one tool would be inconsistent. Will surface as a separate item.
- Reduce line_description in reverse_entry preview (A.8.3, Art.5(1)(c)) —
the preview is shown to the human approver who needs to see what they're
approving under BFL 5 kap. Aggregate-only previews would harm the
approval workflow.
- Audit company-current fields for PII (A.8.12, Art.25(1)) — vat_number,
org_number, etc. are intentionally part of working memory; agents need
them to make compliant booking decisions.
- Payload-size ADR reference (A.8.9) — the test comment already cites plan
item 15 (Tool Search) as the long-term answer.
- mime_type classification label (CC7.2) — theoretical concern;
ai_extraction_usage events are already operator-only.
- False positive: commitReverseEntry already has the closed-period check
(V2.3); bot was hallucinating.
Tests: 3616/3616 pass. TypeScript build clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(mcp,env): structured logger + description trim + env alias support
Two further follow-ups on PR #505:
1. resolvePeriodStatusForDate catch now uses the structured logger
(createLogger from @/lib/logger) instead of console.warn. Three
reviewers (compliance-swarm V16.1.1, ISO 27001 A.8.15, SOC 2 CC7.2)
independently flagged that console.warn bypasses the centralized log
aggregation pipeline used elsewhere, so systemic outages of the
period-status resolver were invisible to the SIEM. log.warn now routes
through the same sink as other server events.
2. Tool description for gnubok_reverse_journal_entry now routes the refund
case explicitly to gnubok_credit_invoice. The Swedish accounting
compliance bot flagged that the previous "cancelled credit invoice"
example was ambiguous — a real credit invoice flow goes through
gnubok_credit_invoice, not this tool. Description stays under 280 chars.
3. lib/init.ts: REQUIRED_EXTENSION_VARS now models each entry as a list of
acceptable aliases instead of a single required name. The fallback in
extensions/general/enable-banking/lib/jwt.ts already accepts the
_PRODUCTION-suffixed variants (used by Vercel prod) as equivalent to
the base names, but the env validator at boot didn't, so every cold
start in prod warned about missing ENABLE_BANKING_APP_ID even though
ENABLE_BANKING_APP_ID_PRODUCTION was set and the runtime was healthy.
Each entry now satisfies if ANY listed alias is present; missing
entries print all acceptable names so operators can pick either form.
Tests: 3616/3616 pass. TypeScript build clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(mcp): staging tools reject locked_at periods too, not just is_closed
Swedish accounting compliance bot flagged that gnubok_reverse_journal_entry
and gnubok_correct_entry pre-flight checks only rejected closed periods —
locked-but-not-closed periods passed staging and only got rejected at
commit time. The commit-time gate was correct (both executors check
is_closed AND locked_at AND resolvePeriodStatusForDate), but the
staging-time signal was confusing: agent saw staged:true with
period_status:"locked" in the same envelope.
Now the staging pre-flight reads locked_at from the same inner-join and
rejects on either flag, matching the commit-time pattern. The error
message updated to "locked or closed" since both branches reach the same
throw. BFL 5 kap 5§ alignment is unchanged — both paths still block
mutations to locked/closed periods; only the layer at which the rejection
fires changes.
Findings pushed back (response in PR thread, not addressed here):
- companyId/mimeType in log.warn flagged as PII (overreach; tenant IDs
are operational identifiers, not personal data, and the codebase logs
them consistently elsewhere).
- HMAC-keyed file_name_hash instead of plain SHA-256 prefix (overreach;
48 bits already addresses the immediate GDPR Art. 5(1)(f) concern).
- 'title' field in deadlines may contain PII (overreach; would require
redacting every text field in every read resource).
- RLS regression test for voucher_sequences/deadlines (legitimate but
pg-test scope; tracked for a follow-up sprint).
- Payload-size ADR record (comment already cites plan item 15).
- company-current data minimisation (already pushed back; agents need
the fields for compliant booking decisions).
- Error message conflates "locked" and "closed" — minor UX nit not
worth distinguishing here since the remediation step (unlock / omprövning)
is the same for the user.
- reversal_date period attribution & voucher series integrity flagged as
unverifiable from diff — false positives, both already handled by the
engine (period_id from original, atomic voucher number).
Tests: 3616/3616 pass. TypeScript build clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(mcp): address Swedish-accounting compliance round 4 — BFL invariant + VAT warning
Three legitimate findings from the swedish-accounting-compliance bot acted on
(out of five total; two pushed back as theoretical/false positive):
1. BFL 5 kap 5§ invariant assertion (finding 1). The engine guarantees that
reverseEntry() posts the storno to original.fiscal_period_id (engine.ts:492
— verified by reading the code), but the executor previously took that on
faith. commitReverseEntry now asserts reversal.fiscal_period_id ===
original.fiscal_period_id after the call and returns a 500 with an
explicit "BFL invariant broken" error if the engine ever drifts. New
executor test covers this. The reversal_date parameter is unchanged —
it's used as the storno's entry_date (operational date), not for period
attribution, per BFL practice (entry_date can differ from period_id's
range for a rättelse made later).
2. resolvePeriodStatusForDate unhandled-rejection path (finding 2). Both
commitCorrectEntry and commitReverseEntry now wrap the resolve call in
try/catch, returning a clean Swedish 500 instead of letting the
dispatcher surface a raw Postgres error message. Matches the
log-and-degrade pattern already used at staging time in
stagePendingOperation.
3. VAT-period warning in the reverse preview (finding 4 — swedish-vat).
When the original entry contains 2610–2670 BAS accounts, the staged
preview now includes a Swedish warnings[] field telling the approver
that a storno is legally insufficient if the moms period has been
filed with Skatteverket — they must use omprövning per ML 2023:200
instead. Soft warning (not a hard block) since gnubok doesn't track
per-VAT-period filing status today; the human decides at approval.
Pushed back:
- Finding 3 (TOCTOU between staging and commit on fiscal_period_id):
posted entries are immutable per the enforce_journal_entry_immutability
trigger (migration 20240101000017). fiscal_period_id can't change
between staging and commit. Status change is already caught by the
status !== 'posted' check.
- Finding 5 (migration 20260516060000 not wrapped in BEGIN/COMMIT):
Supabase migration tooling runs each migration file in an implicit
transaction. PostgreSQL DDL is transactional. The DROP/ADD pair is
atomic in practice. The bot acknowledges this as low severity.
Tests: 3617/3617 pass (one new — BFL invariant assertion). Build clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
||
|
|
7175daee87 |
Bug/skv konto numbers (#498)
* feat(skattekonto): add overdue transactions handling and split logic * feat: enhance transaction handling and loading states - Update BalanceHero component to display last synced date and additional information about Skatteverket updates. - Refactor BookDirectlyDialog to simplify transaction linking logic and improve UI for transaction selection. - Revamp InvoiceInboxWorkspace layout for better responsiveness and user experience, including improved skeleton loading states. - Introduce new loading states for ExtensionWorkspace to match the live layout and improve user feedback during data fetching. - Implement exchange rate fetching in QuickReviewDialog, ensuring transactions are always processed in SEK with error handling for rate fetching. - Add structured error handling for unavailable exchange rates in the transaction API. * feat(skattekonto): add 'Skattekonto – saldo & transaktioner' scope and update authorization checks * feat: implement reverse charge handling in supplier invoice calculations and UI |
||
|
|
b0890c7c79 |
Add/docs skv mcp (#494)
* feat: add "book directly" functionality for invoice inbox items - Extend JournalEntrySourceTypeSchema to include 'inbox_item'. - Introduce BookInboxItemDirectlySchema for direct journal entry creation. - Update InvoiceInboxItem type to include matched_transaction_id and created_journal_entry_id. - Implement BookDirectlyDialog component for user interaction. - Create API route for booking directly from inbox items with appropriate validations. - Add SQL migration to support new journal entry references in the invoice inbox items table. - Implement tests for the new booking functionality and ensure proper error handling. * fix(invoice-inbox): update status handling for resolved inbox items * feat: enforce unique journal entry constraint for invoice inbox items |
||
|
|
722b22972d |
feat(enable-banking): per-account ledger mapping for multicurrency setups (#443) — rebase of #487 (#488)
* feat(enable-banking): per-account ledger mapping for multicurrency setups (#443) Today every bank account routes to BAS 1930. Multicurrency setups (Wise, SEB foreign-currency sub-accounts) get pooled into a single SEK ledger account, making year-end FX revaluation a mess. This change lets each bank account under a PSD2 consent map to its own BAS account (SEK→1930, EUR→1932, USD→1933, etc.). The mapping engine already honors IngestOptions.settlementAccount (lib/bookkeeping/mapping-engine.ts:55-57) so the wiring is small: - StoredAccount gains an optional ledger_account field (no migration — bank_connections.accounts_data is already JSONB). - syncAccountTransactions passes account.ledger_account through as settlementAccount so the bank-side leg routes to the right BAS account. - PATCH /accounts accepts account_mappings, validates 4-digit BAS format, and verifies each ledger_account exists in chart_of_accounts before persisting. Selection edits without account_mappings preserve existing values for back-compat. - AccountPickerDialog shows a per-account "Bokför till konto" combobox populated from the company's 19xx accounts, with currency-based defaults (SEK→1930, EUR→1932, USD→1933, GBP→1934) and a non-blocking warning when two enabled accounts route to the same BAS account with different currencies. Reconciliation still scans 1930 only — foreign-currency accounts skip auto-matching until follow-up PR 4 (filed in the plan). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(enable-banking): address review feedback on PR #487 Two real bugs flagged in review: 1. AccountPickerDialog row wrapped a Radix <Checkbox> (which renders as its own <button role="checkbox">) inside <button onClick={toggle}>. Browsers silently flatten nested interactive elements, the Checkbox lost its onCheckedChange handler, and the toggle interaction was effectively broken. Reverted to <label> + <Checkbox onCheckedChange> with the <Select> as a sibling outside the label so clicking it doesn't also toggle. 2. BAS_ACCOUNT_PATTERN was /^[0-9]{4}$/ — accepted 3001 (revenue), 2640 (input VAT), or any 4-digit account that happens to exist in the chart. Direct API calls would bypass the UI's 19% picker filter and silently misroute every bank-side journal-entry leg into the wrong BAS class, corrupting both the ledger and momsdeklaration. Tightened to /^19[0-9]{2}$/ (kassa/bank only). Tests updated to assert non-19xx rejection. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(enable-banking): validate account_mappings UIDs against accounts_data A typo'd UID in account_mappings was silently dropped — the entry never landed in the resulting accounts_data while the response was still 200, leaving the client to believe the mapping was applied. Mirror the existing enabled_uids guard: reject unknown UIDs with 400 + unknown_uids in the body. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(enable-banking): catch syncPromise rejections to prevent process crash When the 60s timeout wins the Promise.race, the underlying Promise.all keeps running. A subsequent bank-API rejection has no registered handler, which surfaces as an unhandledRejection — Node 22 (the self-hosted Docker runtime) terminates the process by default on those. The cron self-heals via initial_sync_completed_at IS NULL, so a no-op catch is the right policy: drop the late rejection, let the cron retry. Caught by Greptile review on PR #488. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
04f902fe8f |
fix(enable-banking): reliable initial backfill, no more silent ~30-day windows (#443) (#486)
* fix(enable-banking): reliable initial backfill, no more silent ~30-day windows (#443) PSD2 first-sync was a compound bug: the cron runs once daily so users got no data for up to 24h after activation; the "Sync now" button defaulted to 30 days and set last_synced_at, permanently locking the cron into 7-day incremental mode and discarding the 90-day backfill window. ASPSPs also truncate history below requested ranges, but the discrepancy was only logged. This change: - Runs the initial backfill inline when the user finishes account selection (PATCH /accounts), so data is available the moment they finish onboarding. - Tracks initial_sync_completed_at separately from last_synced_at; the cron now gates first-sync 90-day window on that, so manual syncs no longer clobber the backfill path. - Surfaces the actual returned date range to the UI ("Initial historik: X → Y (begärde Z)") with a warning when the bank truncated history. - Defaults manual /sync to 90 days (was 30) — matches user intent. - AccountPickerDialog uses SpeedLedger's SIE-anchor pattern when an SIE import covers prior periods (auto-defaults lookback to "day after last SIE entry"), with Bokio-style PSD2 disclosure on the standard path. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(enable-banking): address review feedback on PR #486 Two fixes from review: 1. Memoise the browser Supabase client in AccountPickerDialog. createClient() in the component body returned a new reference every render, and `supabase` was in the SIE-fetch effect's dep array — every checkbox tick or parent re-render re-fired the SIE-imports query. 2. Drop `accounts_data` from the second supabase update inside the activation backfill. The first update already wrote it; including it here races with any concurrent writer (e.g. cron firing in the sub-60s window) and would silently overwrite. Only initial_sync_* metadata + last_synced_at need to be persisted in the second update. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(enable-banking): check metadata-update error after inline backfill The second supabase.update() inside the activation backfill block didn't check its error return. Supabase client methods don't throw on DB errors; they return { data, error }. If the metadata write failed (network blip, RLS quirk, etc.), the handler still populated initialSyncSummary and returned success — UI saw "imported N transactions" while the DB had initial_sync_completed_at = NULL, causing the cron to schedule another full 90-day backfill the next morning. Capture { error } from the metadata update. On failure, surface as initial_sync_error with a metadata_update_failed: prefix and skip the initialSyncSummary population. The cron's gate (initial_sync_completed_at IS NULL) still self-heals on the next run; this just keeps the UI honest about which path got us there. New test stub: SupabaseStub.updateErrorByCall lets a test succeed the first update and fail the second. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
c8461397c8 |
Bug/accounting ps eu (#474)
* feat(api): implement commit functionality for journal entries * fix(extensions): make ExtensionSettings.clear() a real delete so disconnect flows work The 2026-03-30 multi-tenant refactor dropped all RLS policies on extension_data and recreated only SELECT/INSERT/UPDATE. Combined with `value jsonb NOT NULL`, every extension that called `settings.set(key, null)` to clear stored state (cloud-backup disconnect, skatteverket OAuth/AGI cleanup, arcim-migration consent reset) silently failed — the upsert hit the NOT NULL constraint and the error was swallowed, leaving users stuck with stale connection rows. Adds an `extension_data_delete` RLS policy, a `clear(key)` method backed by a real DELETE, switches the four affected handlers, and makes `set()` throw on Supabase error so this class of silent failure can't recur. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(journal-entries): add draft saving functionality to journal entry form * feat: add periodisk sammanställning report generation and CSV export - Implemented period date helpers in `period-dates.ts` for calculating start and end dates based on period type (monthly, quarterly, yearly). - Created `periodisk-sammanstallning.ts` to generate the periodisk sammanställning report, including data fetching, validation, and warning handling. - Developed CSV serializer in `periodisk-sammanstallning-csv.ts` for exporting the report in SKV574008 format. - Added new columns to `company_settings` for storing periodisk sammanställning settings and tax contact information via migration. - Introduced a new migration to add a `paid_with_private_funds` flag to `supplier_invoices` for tracking out-of-pocket expenses. - Updated journal entries to include the new source type for privately paid supplier invoices. * feat(migrations): add paid_with_private_funds flag to supplier_invoices and expand journal_entries.source_type CHECK * fix(ai_requests): drop existing policies and trigger before creating new ones * fix(migrations): ensure extension_data has a proper DELETE policy for ExtensionSettings.clear() * fix(supplier-invoices): update error handling for invalid input in POST request * fix: correct capitalization in project title * fix(migrations): resolve duplicate version 20260513120000 Two migrations shared the same timestamp prefix, causing schema_migrations_pkey collision on Supabase preview branches. Bump extension_data_delete_policy to 20260513120001. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
3fa871c742 |
Bug/accounting suggestion (#456)
* feat: add bike benefit handling and optional vacation accrual - Introduced bike benefit (cykelförmån) with calculations for annual market value and monthly taxable value. - Updated schemas to include new benefit types and validation rules. - Implemented API routes for creating, updating, and deleting employee benefits. - Enhanced salary calculation logic to accommodate new vacation rule options, including a 'none' option for no accrual. - Added UI components for managing employee benefits, including input for bike benefit specifics. - Created database migrations for employee benefits and updated salary line items to support new benefit types. * chore: remove Langfuse env var checks Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat: enhance OAuth callback URL handling and update default scopes for Visma integration * feat: remove trade_name field and simplify company naming in invoices * refactor: destructure canWrite from useCanWrite for consistency across components * feat: enhance PATCH endpoint to validate existing benefits and handle bike benefit updates * feat: add missing label for bike benefit in salary line item types --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
980f29dae8 |
Bug/momsdeklaration skv (#449)
* fix(salary): show birthdate in masked personnummer, hide the 4-digit suffix Flip the personnummer display format from XXXXXXXX-NNNN to YYYYMMDD-XXXX so the sensitive 4-digit suffix is hidden while the (public) birthdate stays visible. Affects the employees list/detail, salary run, payslip PDF, payslip email, and the MCP server tools (list_employees, get_salary_run). Each call site now decrypts the stored personnummer before masking. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(transactions): allow deleting unbooked transactions from "Alla transaktioner" The history list only let users delete via the inbox card; once a category or mall was picked but the verifikation hadn't been created, the row showed "Ej bokförd" with no way to remove it. The API already permits delete while journal_entry_id is null, so the gap was purely a missing UI affordance. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(vat): populate ruta 20-24 for reverse charge + dishonest "Validera OK" Three connected issues caused Skatteverket to reject momsdeklarationer with FK004 even after our local "Validera"-knapp returned OK. 1. supplier-invoice-entries booked fiktiv moms (2614/2624/2634 + 2645/2647) on reverse-charge invoices but never the underlying basbelopp on 44xx/45xx. Ruta 30-32 filled up at SKV while ruta 20-24 stayed at 0 — SKV's FK004 ("silent netting prohibited", ML 13 kap kräver båda sidor). Fix: generateReverseChargeBasisLines in vat-entries.ts emits parallel 45xx/44xx debit + 4598 motkonto credit per rate group. Engine calls it from registration, cash, and credit-note paths. Skipped when the user booked the expense directly on a basis account to avoid double-counting. 4598 added to BAS reference (no migration needed; account_number is plain text on journal_entry_lines). 2. rutorToMomsuppgift rounded each ruta independently but computed summaMoms from the unrounded ruta49. SKV recomputes the sum from integer rutor on their side, so fractional öres caused ±1 SEK drift and SKV rejected with FK009. Fix: derive summaMoms from the already-rounded VAT-amount rutor. 3. "Validera"-knappen only confirmed SKV's internal arithmetic — a declaration with ruta 30-32 populated and ruta 20-24 empty validated fine until /utkast hit FK004. Users got a false green light. Fix: vat-declaration-checks.ts runs locally before the SKV call, blocks Validera/Spara when ERROR-level findings exist, and surfaces them in a separate "Lokala kontroller"-section. Success message reworded so SKV's OK is no longer presented as filing-ready. Tests: 4535/4536/4531/4425 lines + 4598 motkonto on EU/non-EU/byggtjänster RC, credit-note reversal, fractional-öres summaMoms, all four pre-flight codes (RC_BASIS_MISSING, RC_OUTPUT_MISSING, RC_INPUT_VAT_MISMATCH, SUMMA_MOMS_DRIFT). Backfill for already-posted entries follows in the next commit. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat: add skattekonto matching functionality - Enhance TransactionInboxCard to display a warning for potential 1930↔1630 transfers. - Implement match suggestions for skattekonto transactions in the backend. - Create SkattekontoMatchDialog component for linking skattekonto rows to existing journal entries. - Develop SkattekontoInboxCard component to handle skattekonto transactions in the inbox. - Introduce skattekonto-match utility functions for candidate matching and linking. - Update types to include match suggestions and enriched transaction responses. * refactor: reorganize skattekonto types and implement bank counterpart matching logic * docs: update CLAUDE.md to streamline integrations and clarify architecture details * refactor: enhance reverse charge logic to handle non-basis accounts and prevent double-counting --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
eb77ad50b5 |
feat(mcp): add create_voucher + correct_entry MCP tools (#448)
* feat(mcp): add create_voucher + correct_entry MCP tools The MCP toolset had no way to post a journal entry outside the preset workflows (categorize_transaction, create_invoice, …). That blocks legitimate flows the engine already supports — K3 capitalization to BAS 1010, period-end accruals, FX adjustments, prepayments, and rättelseposter for foreign reverse-charge VAT that landed on 2641 instead of 2614/2645. create_voucher exposes the existing createJournalEntry() primitive: arbitrary balanced lines, optional fiscal-period auto-resolution, staged for human approval. correct_entry exposes correctEntry() (storno + new corrected entry per BFL 5 kap 5§) so part of a posted verifikation can be fixed without losing the legs that were right. Both are HIGH risk in OPERATION_RISK_TIERS — the arbitrary account/amount/ period inputs make them compliance-critical despite being structurally similar to uncategorize_transaction (medium). Approval flow unchanged; no auto-commit, regardless of trust level. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(mcp): address PR #448 review — voucher tools hardening Greptile P1 + compliance bot findings, all in one pass. commitCreateVoucher (commit.ts): - Hardcode source_type to 'manual' instead of reading from params. A future direct-staging path or hand-inserted pending_operations row could otherwise inject 'bank_transaction'/'invoice_created'/etc. and corrupt the audit-trail origin. - Re-validate balance defensively before reaching the engine, so a tampered params row surfaces a clean Swedish 400 instead of an opaque engine error. gnubok_create_voucher (server.ts): - Validate the explicit fiscal_period_id when supplied: confirm it exists, is open (is_closed = false), and that entry_date falls within its span. Without this, a closed/locked period was only caught at commit-time with a generic DB-trigger error. - Throw at staging when any line targets an account that's missing from chart_of_accounts or marked inactive, rather than relying on the approver to spot the advisory flag. - Remove source_type from the staged params blob entirely — the executor ignores it anyway, no point letting it travel through. - Add a comment that the staging-time period-lock check is advisory and the executor is the authoritative guard, so future cleanup doesn't remove either as 'redundant'. Descriptions: - gnubok_correct_entry now explicitly notes that the storno + corrected entries land in the original period (defends against compliance bot's speculative "different period" concern recurring on future reviews). - Both tools' tax_code field gets a note that the BAS account number drives momsdeklaration ruta mapping, not tax_code — guards against an LLM treating tax_code as the VAT-routing dial. commitCorrectEntry (commit.ts): - Add a comment pointing at storno-service.ts:99,102,195,198 to make the "uses original period and date" invariant explicit in this file. Tests: - +2 voucher-executors cases: source_type tampering is ignored, unbalanced params return 400. - +10 new voucher-tools tests (MCP layer): unbalanced, closed explicit period, missing explicit period, entry_date outside period, unknown account, inactive account, happy path + correct_entry registration + unbalanced replacement. 720 tests pass in the impacted suites; full suite 2998/2998. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
17c67fece0 |
Inbox UX overhaul + cross-currency supplier-invoice fixes (#444)
* feat(kpi): expense mix and top suppliers charts Replace the single monthly-trend chart with two additional compact visuals on /kpi: expense composition donut (BAS class 4-7) and top suppliers bar (supplier_invoices sum_sek over the fiscal period). KPIReport gains expenseComposition and topSuppliers fields, computed from the trial balance and supplier_invoices rows already fetched in the API. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(nav): swap Deadlines sidebar slot for Dokumentinkorg Sidebar main-menu slot now points to the invoice-inbox extension. The /deadlines page stays accessible via dashboard widgets and direct links — only the prominent nav entry changes. Most users open gnubok to act on incoming documents, not to read tax deadlines. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(supplier-invoices): cross-currency totals, FX residual, review SEK display Five fixes around foreign-currency supplier invoices: - Form layout: move Valuta / Växelkurs / Reverse charge from collapsed "Övrigt" into a visible row above the line-item table. Auto-fetch the Riksbanken rate when switching to a non-SEK currency; never clobber a user-typed rate; clear it when switching back to SEK. - Form submit: reset() the form on successful submit so the useUnsavedChanges hook detaches its beforeunload listener before the router.push, killing the "Are you sure you want to leave?" prompt that fired during Turbopack-mediated navigations. - BankTransactionPicker: drop the strict currency filter that hid every SEK transaction when the invoice was in EUR/USD. Cross-currency rows fall to the bottom with an "Annan valuta" hint instead of producing a meaningless numeric diff. - match-supplier-invoice route: when the bank transaction currency differs from the invoice currency, compute the FX diff against the AP-booked SEK and pass it to createSupplierInvoicePaymentEntry so 7960/3960 catches the residual instead of leaving a permanent stub on 2440. Fix also covers the "EUR transaction paying a SEK invoice" case that the first iteration missed. - Review dialog: buildJournalPreview now multiplies amounts by the exchange rate so the "Verifikation som bokförs" table shows the actual SEK numbers that hit the DB, not the EUR magnitudes labelled with no unit. Header gains an "(i SEK)" hint when foreign currency. Test coverage for the FX residual path covers SEK-SEK (no diff), SEK-into-EUR-invoice (loss), SEK-into-EUR-invoice (gain), foreign-tx- into-SEK-invoice, and the no-rate fallback. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(inbox): rate limits, multi-file UX, onboarding, retry, supplier autolink Big workspace pass on /e/general/invoice-inbox. Highlights: Backend - New table inbox_rate_counters + RPC check_and_increment_inbox_quota. Postgres-backed (no Upstash dep) per-company limit: 30/min, 500/day. Applied at /upload, /inbound, and /items/:id/retry-extraction. - POST /items/:id/retry-extraction — re-runs the deterministic extractor on a stored document when the previous attempt errored. - POST /items/:id/match-supplier — links a freshly-created supplier back to the inbox item so the next action prefills correctly. - POST /api/transactions/create-from-document — creates an uncategorized manual transaction from an inbox item for the "I have a receipt, no bank transaction" case. The user categorizes through the normal flow. - /inbound caps email at 20 attachments/email; truncated count goes to processing_history as AttachmentsTruncated. Rate-limit drops emit RateLimitedDropped and return 200 so Resend doesn't retry. - attach-document side effect: when the document came from an inbox item, the inbox row's matched_transaction_id is updated so the UI can flip it to "Kopplad till transaktion" without a round-trip. New migration: re-introduces matched_transaction_id on invoice_inbox_items as a plain FK (the AI metadata that the previous migration stripped doesn't come back). Workspace UI - Onboarding card replaces the thin empty-state with a 3-step checkmark guide (Aktivera adress → Ladda upp → Matcha eller bokför). Auto-hides when all three steps are done; localStorage-backed dismiss. Beta badge + link to gnubok.se/priser. - Responsive layout: 3-pane at lg, 2-pane at md, master-detail toggle on phone (list xor detail with a back button). - Filter pills (Alla / Behöver åtgärd / Bearbetade / Fel) + search input above the list — client-side over the existing items list. - Multi-file upload queue with "Laddar X av N…" progress counter on the button. Sequential to avoid hammering pdfjs. Selection stays put during a batch (only single-file drops auto-jump the detail pane). - Bulk select + delete with sticky action bar. Items linked to a supplier invoice are skipped with a count toast. - Retry button in the FieldsRail error branch. - "Skapa transaktion från underlag" CTA in the match dialog when no unmatched bank transactions exist. Prefills date/amount/description from the extracted data; user picks the sign. - "Skapa leverantör" inline CTA when the extractor caught a supplier name with no match against existing suppliers. POSTs /api/suppliers with the extracted fields, then auto-links via /items/:id/match-supplier. - Matched-state CTA renamed to "Bokför transaktionen" with link to /transactions?highlight=<id> so the categorize panel auto-opens. Tests - lib/rate-limits/__tests__/inbox.test.ts — RPC wrapper happy/error/scope - app/api/transactions/create-from-document/__tests__/route.test.ts — auth, validation, 404/409/200/500, inbox-link failure tolerated - extensions/general/invoice-inbox/__tests__/retry-extraction.test.ts — auth, rate limit, 404, 409, 400 no-doc, success, extraction failure - attach-document tests extend coverage to the new inbox-link side effect (both success and best-effort failure paths) - inbound-webhook test mocks the rate-limit module so the queued-mock sequence in each existing test doesn't have to know about it CLAUDE.md gains a row for lib/rate-limits/ so the new helper is discoverable. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(transactions): paperclip indicator and highlight-row param Close the feedback loop after a user attaches a receipt to a transaction from the inbox: the row in /transactions now shows a paperclip icon when transaction.document_id is set, with a click handler that fetches a signed download URL and opens the document in a new tab. Works for both uncategorized and history views. When the inbox sends a user to /transactions?highlight=<id>, the page now scrolls that row into view and auto-opens the categorize panel if the transaction is still uncategorized. Behind a double-rAF so the row DOM exists when scrollIntoView fires. QuickReviewDialog no longer prompts to upload underlag when the transaction already has a doc attached (which it does after the inbox match flow). Shows "Underlag bifogat — Visa" instead, opening the existing doc in a new tab. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(pr-444): address review feedback (Greptile + compliance bots) Migration rules - New migration 20260512092423: adds updated_at trigger on inbox_rate_counters (CLAUDE.md rule 2) and explicit USING (false) RLS policies for the four DML verbs to make the SECURITY DEFINER-only intent explicit (rule 1). - New pg-real test inbox-rate-limit.pg.test.ts covering happy path, minute-cap rejection, day-cap rejection, per-company isolation, and the updated_at trigger firing. CLAUDE.md mandates *.pg.test.ts for every new RPC because mocks pass on broken PL/pgSQL. Bugs - Stale exchange rate on currency switch (Greptile P1) — userTouchedRateRef was scoped per session, not per currency. Switching EUR (with a hand-edited rate) → USD kept the EUR rate. Now tracks the last fetched currency in a ref and resets the touched flag on currency change while still honoring manual edits within a single currency. - topSuppliersResult.error silently swallowed (Greptile P2) — failed queries used to render an empty chart matching the no-data state. Logged now. - Currency from extracted_data not validated (GDPR Art.25(2), OWASP V4.5, Swedish compliance bot) — extracted PDF currency was inserted into transactions.currency without sanitisation. Allowlisted against the six supported ISO 4217 codes; coerce to SEK otherwise. - Idempotency gap on create-from-document (OWASP V2.3) — two concurrent POSTs with the same inbox_item_id could each pass the matched_transaction_id IS NULL read and insert duplicate transactions. UPDATE now includes .is('matched_transaction_id', null) as an optimistic-lock release and returns 409 with an orphan-transaction rollback when the predicate doesn't match. - FX residual on cash-method match path (Swedish compliance bot) — createSupplierInvoiceCashEntry has no exchange_rate_difference path, so a cross-currency match would silently leave a 1930 reconciliation gap. Added a guard that returns MATCH_SI_CASH_FX_UNSUPPORTED (400) before the JE is created. Users on cash method can switch to accrual or book the FX diff manually. Design system - gap-y-1.5 / gap-1.5 in KPIExpenseMixChart — replaced with gap-y-2 / gap-2 (CLAUDE.md design tokens; 2.5/1.5/5/hardcoded pixels are forbidden spacing values). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore(migrations): rename to match applied versions The mcp__plugin_supabase_supabase__apply_migration tool stamps its own timestamp when it applies a migration to the live project, so the version recorded in supabase_migrations.schema_migrations differs from my local generation-time filenames. Renaming the local files so a production CD run sees the migrations as already-applied (matching versions) instead of trying to re-apply them — which would fail for the trigger/RLS migration (CREATE TRIGGER and CREATE POLICY don't support IF NOT EXISTS). Follows the pattern from d854efcd ("chore(migration): rename to match applied version"). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(create-from-document): scope orphan rollback DELETE by company_id Defence in depth on the inbox-link race rollback. newTx.id is a fresh UUID from a company-scoped insert two statements above, so the existing single-key DELETE is already safe, but adding .eq('company_id', companyId) makes the cross-company invariant explicit on every write — addresses the OWASP ASVS V2.3 finding from the compliance swarm on PR #444. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(nav): mark Dokumentinkorg with Beta badge Same signal we use for Löner and Anställda — the inbox flow (AI extraction, supplier autolink, manual transaction creation) is in end-to-end customer testing. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
a53a119a2e |
Fix/vat parent accounts (#438)
* feat(enable-banking): add support for account selection and syncing - Updated StoredAccount interface to include an 'enabled' flag for account syncing preferences. - Enhanced ensureFiscalPeriod function to handle overlapping fiscal periods with posted entries and opening balances. - Added tests for fiscal period validation and account syncing logic. - Implemented AccountPickerDialog component for user account selection. - Created API routes for PATCH /accounts and POST /sync to manage account syncing. - Introduced 'pending_selection' status for bank connections to allow user account selection before syncing. - Updated database migration to support new connection status and backfill existing accounts with enabled=true. * feat(enable-banking): implement account selection and consent event logging --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
ed9bdc4c00 |
Bug/transaction import (#435)
* fix(invoices): drop UTKAST banner on numbered invoices and preserve logo aspect ratio Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * revert(invoices): restore UTKAST banner for drafts; keep logo aspect ratio fix Numbered drafts intentionally surface UTKAST until manually marked sent. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * revert(invoices): drop logo objectFit change Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(enable-banking): implement transaction fetch strategy and update related logic --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
75043632de |
Fix/compliance and invoices (#433)
* feat(invoices): implement inline membership checks for invoice-number RPCs and enhance error handling * fix(invoices): enhance unpaid amount calculation to support currency-specific rounding * fix(api-client): handle response.text() error for 403 status in skvRequest |
||
|
|
1876b3b402 |
fix(agi): update AGI submission logic to reflect SKV's auto-persistence and adjust error handling (#432)
fix(pdf): align company logo and document title in invoice PDF layout |
||
|
|
c40e1114ca |
Bug/skv connection (#431)
* feat(invoices): implement öresavrundning logic and next invoice number preview - Added `getDisplayTotal` utility to handle rounding for SEK invoices based on company settings. - Updated `InvoicesPage` to utilize the new rounding logic when displaying totals. - Introduced `peek_next_invoice_number` function to allow previewing the next invoice number without consuming the sequence. - Modified invoice number generation to remove the year prefix and prevent truncation of numbers exceeding three digits. - Enhanced tests for invoice number generation and rounding functionality to ensure correctness. - Updated PDF template to reflect new rounding logic for totals and display appropriate values. - Adjusted company switcher to hide options in sandbox mode. - Improved error handling and logging in sandbox seeding process. * fix(skatteverket): remove unused scope labels from SCOPE_LABELS and DEFAULT_SCOPES * feat(skatteverket): add token revocation handling and disconnect functionality * feat(skatteverket): implement PKCE support for OAuth2 flow to enhance security |
||
|
|
d7a7ddae2c |
Bug/skv connection (#430)
* feat(invoices): implement öresavrundning logic and next invoice number preview - Added `getDisplayTotal` utility to handle rounding for SEK invoices based on company settings. - Updated `InvoicesPage` to utilize the new rounding logic when displaying totals. - Introduced `peek_next_invoice_number` function to allow previewing the next invoice number without consuming the sequence. - Modified invoice number generation to remove the year prefix and prevent truncation of numbers exceeding three digits. - Enhanced tests for invoice number generation and rounding functionality to ensure correctness. - Updated PDF template to reflect new rounding logic for totals and display appropriate values. - Adjusted company switcher to hide options in sandbox mode. - Improved error handling and logging in sandbox seeding process. * fix(skatteverket): remove unused scope labels from SCOPE_LABELS and DEFAULT_SCOPES * feat(skatteverket): add token revocation handling and disconnect functionality |
||
|
|
0ee5219b6c |
feat(invoices): implement öresavrundning logic and next invoice numbe… (#429)
* feat(invoices): implement öresavrundning logic and next invoice number preview - Added `getDisplayTotal` utility to handle rounding for SEK invoices based on company settings. - Updated `InvoicesPage` to utilize the new rounding logic when displaying totals. - Introduced `peek_next_invoice_number` function to allow previewing the next invoice number without consuming the sequence. - Modified invoice number generation to remove the year prefix and prevent truncation of numbers exceeding three digits. - Enhanced tests for invoice number generation and rounding functionality to ensure correctness. - Updated PDF template to reflect new rounding logic for totals and display appropriate values. - Adjusted company switcher to hide options in sandbox mode. - Improved error handling and logging in sandbox seeding process. * fix(skatteverket): remove unused scope labels from SCOPE_LABELS and DEFAULT_SCOPES |
||
|
|
c110d3ef99 |
Bug/stale skv connection (#428)
* fix(dashboard): enable salary features with "Beta" badge for testing * fix(errors): enhance Swedish error message patterns for better user feedback fix(salary): update Nordea Personkonto handling in account encoding logic * fix(agi): implement feature flag for AGI transmission and update button states * fix(swedish-payroll): update youth rate eligibility criteria and enhance documentation * fix(agi-panel): remove feature flag for AGI transmission and simplify button states * fix(agi-panel): handle stale session errors and improve reconnect flow * fix(api-client): enhance error handling for 401 responses and clarify subscription issues |
||
|
|
efdd3a50ba |
Bug/stale skv connection (#427)
* fix(dashboard): enable salary features with "Beta" badge for testing * fix(errors): enhance Swedish error message patterns for better user feedback fix(salary): update Nordea Personkonto handling in account encoding logic * fix(agi): implement feature flag for AGI transmission and update button states * fix(swedish-payroll): update youth rate eligibility criteria and enhance documentation * fix(agi-panel): remove feature flag for AGI transmission and simplify button states * fix(agi-panel): handle stale session errors and improve reconnect flow |
||
|
|
7e81f661b2 |
Add/skv salary agi (#423)
* feat: add Bankgirot LB-fil support for salary payments and tax payments - Implemented `generateBgLb` for salary batch payments, producing opening, payment, and closing records. - Added tests for `generateBgLb` to ensure correct record generation and validation. - Created `generateBankgiroPaymentBgLb` for single tax payments to Skatteverket, including validation and formatting. - Added tests for `generateBankgiroPaymentBgLb` to verify record structure and data integrity. - Introduced `generateSkattekontoOcr` for generating valid OCR references for Skattekonto payments, with tests for various input formats. - Updated database schema to track payment file formats and timestamps for salary runs and AGI declarations. - Created a new table for logging salary payslip deliveries to ensure compliance with audit requirements. * feat: add write permission check and company ID validation for payment file generation * feat: add write permission check for salary payment file generation |
||
|
|
8a1c7f0714 |
feat(invoice-inbox): loading state + review fixes + env-driven config (#416)
* feat(invoice-inbox): make Bedrock model + max_tokens env-overridable Read BEDROCK_MODEL_ID and BEDROCK_MAX_TOKENS from process.env so ops can swap models or raise token caps without a code deploy. Defaults align with production: eu.anthropic.claude-sonnet-4-6 in eu-north-1, 8192 tokens (up from 4096 — enough headroom for 20+ line-item invoices that were still occasionally truncating mid-JSON). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): guard onFieldsUpdated, tighten env parsing, hoist afterAll Three fixes from Greptile review on #416: 1. Stale-closure bug in onFieldsUpdated. setSelected used an unconditional spread, so a PATCH that completed after the user had navigated to a different item would silently overwrite the new selection's extracted_data with the old item's payload. Guard with prev?.id === selected.id (capturing selected.id outside the updater so it's stable through both setSelected and setItems). 2. BEDROCK_MAX_TOKENS env parser used `Number(...) || 8192`, which silently treats a deliberate `0` the same as unset. Use Number.isFinite(n) && n > 0 so an invalid configuration is not masked by the fallback. 3. Hoist `afterAll` import in the test file to the top alongside other vitest imports. Was working via ES-module hoisting but reads weird to humans and linters. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
9c13586b13 |
feat(invoice-inbox): AI extraction (Bedrock Sonnet 4.6) + editable fields (#415)
* feat(invoice-inbox): AI extraction via Bedrock + editable fields Two big changes that go together: 1. Replace the regex extractor with Claude Sonnet 4.6 via AWS Bedrock. The PDF or image is sent directly to the model — no unpdf, no DOM stubs, no worker bundling. Sonnet handles English receipts (Anthropic, AWS, Stripe), USD/EUR currency symbols, scanned PDFs and image receipts, and Subtotal vs Total disambiguation that the regex layer couldn't. Output is JSON, validated with Zod; anything that doesn't parse falls back to an empty result so the inbox row still lands. 2. Make the extracted fields editable inline in the workspace. Each field is an Input bound to local draft state with debounced auto-save (800ms) to a new PATCH /items/:id/fields route. The route refuses edits once the item is converted to a supplier invoice. The parent updates both the selected-item view and the list rail when a field saves so the summary stays in sync. Drops `unpdf`, adds `@anthropic-ai/bedrock-sdk`. Requires three env vars in Vercel prod (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_REGION) — not yet set; when missing, the extractor logs a warning and returns empty so the upload still works. All 43 invoice-inbox unit tests pass. tsc clean. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(invoice-inbox): loading state + #415 review fixes Loading state for AI extraction: - Optimistic placeholder row inserted into the workspace list the moment an upload starts. Shows the file name + a "Tolkar dokument med AI…" spinner instead of timestamp/amount. - Document preview pane shows a centered loader + same caption while the placeholder is selected. - Fields rail shows 6 skeleton inputs + the same caption. - Action buttons are hidden until the real row arrives. - Placeholder is removed on success or failure (toast on error). PR #415 review follow-ups (Greptile + Swedish compliance bot): - Bump max_tokens 1500 → 4096 so multi-line invoices don't silently truncate mid-JSON and land with all-null fields. - Tighten NullableDate to range-checked regex + Date.parse refine — rejects 2026-13-45 / 2026-02-30 with a clean 400 instead of a 500. - Make vatRate representation consistent: percent integer (25, 12, 6, 0) for both lineItems[].vatRate AND vatBreakdown[].rate. Previously the AI was instructed to emit decimals for one and integers for the other. - EditableFieldsList re-seeds drafts when the parent passes a new data snapshot, but only on fields where the local draft still matches the previous server value — so a server-normalised currency upper-cases cleanly without clobbering an in-progress edit. - 409 conflict (item already linked to a supplier invoice) shows the server's specific Swedish message ("Posten är låst") instead of the generic "Kunde inte spara". Out of scope (separate tickets if needed): live updates for email-arrived items, AWS_SESSION_TOKEN plumbing, 6 % livsmedel rate transition validation, server-side totals consistency check. All 43 invoice-inbox unit tests pass. tsc clean. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): tighten extraction schema per compliance bot review Three follow-ups from the Swedish compliance bot's second pass on #415: 1. lineItems[].vatRate and vatBreakdown[].rate now refine to 0–100, blocking AI hallucinations like 5000% or negative rates while still accepting non-Swedish rates (UK 20, DE 19) since gnubok stores foreign invoices for reference. The strict Swedish [0, 6, 12, 25] allowlist is not enforced here on purpose — that check belongs in the supplier- invoice-creation step where the data hits the ledger. 2. accountSuggestion is now coerced to null at parse time via .transform, eliminating the brief intermediate window where a hallucinated string could appear in the parsed object before the post-validation .map() nulled it. Removes the redundant .map() afterwards. 3. PATCH /items/:id/fields currency now requires ISO 4217 format (^[A-Z]{3}$). Previously accepted any 3–8 char string, which would flow into supplier-invoice creation and produce a faktura with an invalid currency. Test fixture vatRate updated 0.25 → 25 to match the percent-integer convention introduced in the previous commit. All 43 invoice-inbox unit tests pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
9e1e13d388 |
Fix/bank and onboarding (#413)
* fix(onboarding): validate org number at step 2 instead of final submit Run normalizeOrgNumber (Luhn + 10/12-digit check) inside the Step 2 Zod schema and gate the duplicate-check and TIC lookup effects on it, so users get an inline error on the field they just typed instead of filling out two more steps and being bounced back from the server. Server-side check stays as defense in depth. Also fixes the old regex incorrectly rejecting valid 12-digit org numbers. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(reconciliation): pass real userId to transaction.reconciled events runReconciliation and manualLink fell back to companyId when no userId was provided, causing FK violations on event_log.user_id (which references auth.users). Make userId a required arg in both functions, drop the fallback, and forward a real user id from every caller (user.id from authed routes, connection.user_id from the cron path). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * refactor(enable-banking): use local companyId instead of reinlining ctx fallback Three sites in the enable-banking sync handler reinline `ctx?.companyId ?? user.id` instead of using the local `companyId` declared at the top of the block. Collapse all three to the local for consistency and to remove drift risk if the fallback expression ever changes. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
8676f10fdd |
fix(invoice-inbox): switch from pdfjs-dist to unpdf (#409)
* fix(invoice-inbox): switch from pdfjs-dist to unpdf for PDF text extraction After three rounds of fighting pdfjs on Vercel (#407 stubbed DOM globals, #408 tried to ship the worker file via outputFileTracingIncludes), text extraction still failed in prod with "Setting up fake worker failed" — Next's tracer can't reliably include pdfjs-dist's worker file when the package is marked as a server external. unpdf is a serverless-first wrapper around pdfjs (by unjs) that ships its own bundled pdfjs build with no canvas/worker dependencies. Drop-in replacement: extractText returns merged page text directly. - Remove DOM stubs, serverExternalPackages, outputFileTracingIncludes - Replace pdfjs-dist with unpdf (no transitive deps) - Update test mock from getDocument → extractText All 45 invoice-inbox unit tests pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): make unpdf import static, rename pdfjs test labels Per Greptile review on #409: 1. CRITICAL: tryExtractPdfText was still using await import('unpdf'), a dynamic import. CLAUDE.md forbids dynamic imports in extensions precisely because Next.js bundling can't reliably trace them — which is the same class of failure that caused the pdfjs prod bug. unpdf bundles statically (no canvas/worker), so a top-level static import is safe and correct. 2. NIT: two test descriptions still said "pdfjs" after the mock rename. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
5f9f72959f |
fix(invoice-inbox): stub DOM globals so pdfjs runs on Vercel (#407)
* fix(invoice-inbox): stub DOM globals so pdfjs text extraction works on Vercel pdfjs-dist's legacy build references DOMMatrix/ImageData/Path2D at module load. On Vercel's Node runtime those globals don't exist, so the dynamic import threw "DOMMatrix is not defined" before getDocument() ran — every inbox item came back with empty extracted_data. We only call getTextContent (no rendering), so empty-class stubs are sufficient. Also mark pdfjs-dist as a server external package so Next.js doesn't try to bundle it (the bundling step pulls in @napi-rs/canvas references that aren't installed and fails silently with the same downstream error). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * refactor(invoice-inbox): hoist DOM stubs to module scope Per Greptile review on #407: stubs run once at module load instead of per-call. Same effect, clearer intent. Wrapped in a block to scope the local variable. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
e77423d099 | feat(transactions): add tool to list transactions without documents and update scope map (#403) | ||
|
|
4131db2894 |
chore: MCP intent-tools, BankID enrichment table, multi-tenant fixes (#402)
* chore: MCP intent-tools, BankID enrichment table, multi-tenant fixes MCP server gains six intent-shaped tools that collapse multi-call agent flows into one: vat_close_check, query_journal, auto_match_period, create_supplier_invoice_from_inbox, audit_package, year_end_readiness. Tools wired into TOOL_SCOPE_MAP and OPERATION_RISK_TIERS as appropriate (create_supplier_invoice_from_inbox at medium tier — reversible until approve, but stages a leverantörsskuld). BankID enrichment now persists to a dedicated bankid_enrichment table keyed by user_id. extension_data has been company-scoped (NOT NULL company_id) since the multi-tenant refactor, so every BankID signup has silently been failing the enrichment upsert. Select-company picker reads from the new table. delete_last_voucher (BFNAR 2013:2) needs to clear document_attachments.journal_entry_id before deleting the entry, but the new document immutability trigger blocks that UPDATE. Added the same gnubok.allow_delete transaction-scoped bypass pattern used by the journal-entry/line/retention triggers. pg-real tests cover the happy path, the unauthorized direct UPDATE, and the swap-to-different-entry attempt under the bypass flag. fiscal_periods.no_overlapping_fiscal_periods exclusion was scoped to user_id from before multi-tenant — rebound to company_id so the same user can have overlapping fiscal years across companies they own/are member of. Also adds scripts/seed-demo-account.ts for end-to-end demo seeding (two companies, full FY2025, active FY2026 with mixed state). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(pr-402): address review feedback Migrations - Drop 20260506140000_document_journal_entry_immutability_delete_bypass.sql: redundant with 20260506140000_document_journal_entry_immutability_bypass.sql that landed on main while this branch was open. Both share the same gnubok.allow_delete pattern; main's version is what the DB actually has. - Rename 20260506150000_bankid_enrichment_table.sql → 20260506160000_bankid_enrichment_table.sql to clear the timestamp clash with 20260506150000_protect_document_journal_link.sql on main (Supabase branch preview was failing on schema_migrations PK collision). Tests - Drop the swap-under-flag test from delete-last-voucher.pg.test.ts: main's bypass returns NEW unconditionally when gnubok.allow_delete='true', so the swap is permitted. Drop the duplicate happy-path test (already covered by 'clears journal_entry_id on attached documents and deletes the voucher'). Keep the unauthorized-direct-UPDATE test. - Add bankid-enrichment.pg.test.ts covering the SELECT RLS policy: user reads own row, cannot read another user's row, INSERT denied for authenticated. gnubok_query_journal - amount_min/amount_max is applied post-fetch (PostgREST can't OR abs(debit) and abs(credit) cleanly), but PostgREST's count is computed pre-filter. Reporting that as total_lines mislead agents into paginating a tail that was already filtered out. When the amount filter is applied, anchor total_lines and truncated to the filtered set and surface db_matched_pre_amount_filter + amount_filter_applied_post_fetch separately. - Escape `_` in the free-text LIKE filter so a search for "2_441" doesn't match "2X441". VAT close check - Reverse-charge blocker no longer fires on ruta 30 (seller-side domestic omvänd skattskyldighet) — the seller books no VAT, the buyer does, so missing ruta 48 is expected. Now scoped to ruta 31/32 (EU acquisition) where the buyer must book both calculated output (2615) and matching ingående moms (2645). - High-value receipt threshold no longer reads journal_entries.total_amount (column doesn't exist; check silently never fired). Sums debits across the entry's lines, which equals the gross for ordinary purchase entries — comparing a gross figure against the BFL/ML 4 000 SEK threshold per ML 17 kap 26–28 §. seed-demo-account.ts - Require an explicit email argument; refuse to run with the previously hardcoded fallback that would silently target a real user. Ensure email is non-undefined for downstream typing. - Type the supabase fiscal_periods insert result locally so tsc no longer reports 'fp implicitly any' from the loose untyped client. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(test): adjust fiscal-period-start-day pg test for per-company overlap The pg-real failure on PR #402 was a latent bug surfaced by this branch's fiscal_periods exclusion constraint flip from user_id to company_id (migration 20260506140100). The test was inserting periods that overlapped seedCompany's default 2026-01-01..2026-12-31 period; the previous constraint slipped past it because the test's INSERT didn't set user_id (NULL escapes the WITH = match), so two same-company overlapping periods silently coexisted. Now that the constraint correctly fires per company, pick years that don't overlap with the seeded 2026 period. The trigger's behavior under test (allow mid-month start when no earlier period exists, allow back-dated SIE imports, reject mid-month start when an earlier period exists) is unchanged. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(vat-close-check): correct reverse-charge/import blocker rutor Rutor 30/31/32 are the buyer's calculated utgående moms on reverse- charge purchases (domestic byggtjänster/electronics → 2614 → ruta 30; EU goods → 2624 → ruta 31; EU services → 2634 → ruta 32). The buyer must also book matching ingående moms (2647 inhemskt / 2645 utlandet → ruta 48). The previous fix removed ruta 30 on the basis that it was seller-side; that's incorrect — domestic-RC sellers book no VAT at all (they report only beskattningsunderlag on ruta 41), so 2614 only sees buyer-side entries. Restore ruta 30. Also extend the check to import rutor 60/61/62 (non-EU import VAT declared via momsdeklaration since 2015 — 2615/2625/2635). Same mechanic: importer books output VAT on these rutor and deducts the input side via ruta 48. SaaS-from-AWS / OpenAI / Vercel companies hit this path; without including 60/61/62 the blocker would silently miss their misbookings. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(mcp): expose ruta 60/61/62 (import VAT) on the local VatReportResult The vat-close-check fix referenced vatReport.rutor.ruta60/61/62 but the MCP server's local VatReportResult type only carries ruta 05-49. Build broke on tsc. Extend the MCP server's slim VAT report to also project import VAT — 2615 → ruta 60 (25%), 2625 → ruta 61 (12%), 2635 → ruta 62 (6%) — and fold those into ruta 49 (att betala/återfå). Mirrors the BAS-to-Ruta mapping in lib/reports/vat-declaration.ts. Output schema and required list updated accordingly. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
b9a2ce522b |
chore: post-bankid redirect, recapt feedback, TIC SPAR enrichment (#400)
* chore: post-bankid redirect, recapt feedback, TIC SPAR enrichment - BankID login + register now redirect to /select-company so the picker shows freshly enriched CompanyRoles from the current session. - New lib/support/submit-feedback util prefers window.recapt feedback widget when present, falls back to /api/support/contact. SupportLink uses it and hides itself in sandbox companies via new isSandbox flag on CompanyContext (+ useCompanyOptional hook). - TIC enrichment re-requests SPAR alongside CompanyRoles now that both types are enabled on the tenant; enrichment shape logged PII-free (booleans/counts only). Tests cover the SPAR+CompanyRoles path. - Skatteverket api-client: 15s AbortSignal timeout on outbound requests. - Swedish compliance review CI: bump REVIEW_MODEL to claude-opus-4-7. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore: keep compliance review model on sonnet-4-6 Reverts the opus-4-7 bump from the previous commit per request. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(tic): don't persist SPAR PII to extension_data The previous commit started requesting SPAR alongside CompanyRoles and wrote the full enrichment payload (incl. personnummer, full name, home address, birth date, gender) verbatim to extension_data.value — a plain JSON column. Personnummer is already hashed + encrypted in bankid_identities, so the extension_data row was an unencrypted PII duplicate exposed to anyone with read access to the table. No consumer (middleware, /select-company, createCompanyFromTicRole) reads any SPAR field today; they only read companyRoles. Persist a sanitized blob of { companyRoles, enrichedAtUtc } instead. SPAR is still requested from TIC (and its shape logged PII-free) so enrichment completes; if address pre-fill ships later, those fields should be encrypted before storage. Also drop the dead `null` branch from SubmitFeedbackResult.channel — every code path returns 'recapt' or 'email'. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
5725c25bf1 |
Logs/improved logging (#398)
* feat(mcp): add create_transactions tool with /pending approval gate New MCP tool gnubok_create_transactions stages 1–10 transactions per call as pending_operations of type create_transaction (risk: medium). Each item becomes its own card on /pending; on confirm, the executor inserts the row into transactions with import_source='mcp' so MCP-staged ingestion is distinguishable from PSD2 sync. Designed for skill workflows that pull external data (e.g., Airtable) and want the user to gate the writes. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(bas): strip concatenated group headers from corrupted account names A chart-data import bug had glued the next group's header onto the last account in each preceding group across all eight bas-data class files (e.g. account 2670 read "Utgående moms på försäljning inom EU, OSS 27 PERSONALENS SKATTER, AVGIFTER OCH LÖNEAVDRAG"). The corrupted names surface in transaction dropdowns, ledgers, SIE exports and årsredovisning, and risk VAT miscategorization on the OSS (2670) and blandad-verksamhet (6999) accounts specifically. - Cleans 69 account_name and 64 description fields across class-1..8 files - Adds a regression test asserting no name contains a concatenated header - Ships an idempotent safety-net migration that updates already-seeded chart_of_accounts rows, gated on the corrupted string so user customizations are preserved Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(errors): add structured error codes and handling for various operations - Introduced a new structured error registry in `structured-errors.ts` to standardize error handling across the application. - Added Swedish and English messages for various error scenarios, including validation, authorization, and bookkeeping errors. - Implemented a client-side error toast in `use-error-toast.ts` to display user-friendly error messages with remediation hints. - Created a wrapper for recording operation outcomes in `record-operation.ts`, enhancing audit capabilities for operations. - Developed a provider call wrapper in `with-provider-call.ts` to handle external HTTP calls with structured logging and error mapping. - Added a new SQL migration to extend the processing history with new event types and aggregate types for better operational telemetry. * Refactor supplier API routes to use context-based logging and error handling - Replaced direct Supabase client usage in GET and POST routes with context-based approach using `withRouteContext`. - Enhanced error handling to provide structured error responses for supplier creation and listing. - Updated logging to include request IDs for better traceability. - Introduced new error codes for supplier-related operations. - Refactored tax deadlines cron job to utilize context and improved error handling. - Updated ESLint configuration to enforce logging practices across API and lib directories. - Enhanced arcim migration extension with structured error handling and logging. - Added classification for provider errors to improve user-facing error messages. - Introduced request ID in extension context for better log correlation. * fix(route-context): update DynamicParams type for improved type safety in route handlers * feat(transactions): add 'create_transaction' operation to PendingOperationType * fix(route): ensure companyId is non-nullable in loadAndDeriveAbsence function * fix(route-context): ensure companyId is always non-null by short-circuiting with COMPANY_CONTEXT_MISSING --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
94f15b9c6c |
feat(invoice-inbox): pin documents to bank transactions + MCP tools (#397)
* feat(invoice-inbox): pin documents to bank transactions + MCP tools Adds a first-class flow for attaching unmatched inbox documents to bank transactions, separate from the existing supplier-invoice convert path: - new transactions.document_id FK → document_attachments (ON DELETE SET NULL) - POST/DELETE /api/transactions/[id]/attach-document - categorize route propagates the link to journal_entry_id on commit - three new MCP tools: gnubok_list_unmatched_documents, gnubok_get_document_content (5-min signed URL), gnubok_attach_document_to_transaction (staged via pending_operations) - InvoiceInboxWorkspace gains a "Koppla till transaktion" picker dialog ranked by amount-match, plus a "Bilaga" badge in SwipeCategorizationView - regex extraction unchanged; supplier-invoice convert flow unchanged Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): address PR #397 review findings - categorize: destructure { error } from the document-link update so Supabase-level failures are logged instead of silently dropped (BFL 5 kap 6 § receipt-on-verifikation contract). - list_unmatched_documents: emit next_cursor whenever the inbox query may have more rows, not only when the post-filter slice was full; switch to composite (created_at, id) cursor to avoid same-second collisions. - DELETE /attach-document: return 404 when the tx isn't in the company; return 409 when the linked document already has journal_entry_id set (räkenskapsinformation immutability). - risk tier: attach_document_to_transaction medium (was low) — link becomes part of verifikation underlag once categorize propagates it. - pg-real test: stop reusing $2 across uuid + text-concat contexts (Postgres couldn't deduce the parameter type). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(migrations): break duplicate version 20260505120000 (Supabase Preview) Two migrations on main share filename version 20260505120000: - 20260505120000_api_keys_refresh_token.sql (PR #392) - 20260505120000_drop_agent_auto_commit.sql (PR #394) The schema_migrations primary key is (version), so any fresh DB doing `supabase db push` over both files conflicts on the second insert. This is why every PR with a migration since #394 has had Supabase Preview either fail or skip. Renaming _drop_agent_auto_commit to 20260505190027 — that matches the timestamp recorded in prod schema_migrations from when apply_migration was called for it, so future `db push` against prod sees the file as already- applied (no re-run). The migration body is fully idempotent (IF EXISTS on every drop) so a re-run would be a no-op anyway. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): address PR #397 round-2 compliance review Two BFL gaps the compliance bot flagged on the round-1 fix commit: 1. commitAttachDocumentToTransaction silently broke verifikation→underlag if the transaction was categorized between staging and approval. Now reads transactions.journal_entry_id at commit time and, if non-null, also writes document_attachments.journal_entry_id in the same commit so BFL 5 kap 6 § is satisfied regardless of order. 2. Application-layer DELETE check was racy (SELECT then UPDATE) and the FK ON DELETE SET NULL path could null transactions.document_id even for a document that is räkenskapsinformation. Added a BEFORE UPDATE OF document_id trigger on transactions that raises check_violation when the previously-attached document has document_attachments.journal_entry_id set. The app-layer guard stays for friendly Swedish messaging; the trigger is the DB-level safety net. pg-real test extended to cover both directions of the trigger (block detach + block swap) and the happy-path detach when there's no JE link yet. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): address PR #397 round-3 compliance review Four findings from the round-2 update of the compliance bot. The first three are genuine compliance gaps; the fourth (preview metadata distinguishing pre- vs post-categorization overwrites) is a UX nicety left for follow-up. 1. transactions.document_id FK switched from ON DELETE SET NULL to RESTRICT (migration 20260506100000). Removes the "trigger ordering" concern: a doc that's pinned to any tx now cannot be deleted at all without explicit detach first. Belt-and-braces with block_document_deletion. 2. commitAttachDocumentToTransaction now does: - pre-check that mirrors the DELETE route's 409 when the existing pinned doc is räkenskapsinformation, so the same Swedish message is returned in both paths; - UPDATE…RETURNING journal_entry_id so the propagation decision uses the post-update state, closing the read-then-write race with concurrent categorize. Either ordering of attach-then-categorize or categorize-then-attach now lands at the same correct final state. 3. Both DELETE /attach-document and the MCP commit path catch the trigger's check_violation (SQLSTATE 23514) and translate to 409 with the Swedish underlag message. The trigger remains the DB-level safety net; the app layer is responsible only for friendly UX. pg-real test rewritten for ON DELETE RESTRICT (blocks deletion of pinned doc; detach-then-delete works). Unit coverage added for commitAttach: 404, two distinct 409 paths (pre-check + trigger-translation), happy-path uncategorized, and propagation when tx was categorized between staging and commit. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): address PR #397 round-4 compliance review Three of five round-3 findings actioned: 1. commitAttachDocumentToTransaction: surface propagation failure rather than logging-and-continuing. If document_attachments.journal_entry_id can't be set after the transaction has been categorized, the op fails (status 500) with a Swedish message instructing retry. Retry is idempotent — same document_id on the tx, same propagate target. 2. Replace check_violation (23514) matching with a stable "BFL_DOCUMENT_IMMUTABILITY:" message prefix. The trigger now uses default P0001 + tagged message; both the route handler and the executor match on the prefix instead of the generic SQLSTATE. Future unrelated CHECK constraints on transactions can no longer accidentally surface as the räkenskapsinformation message. 3. gnubok_list_unmatched_documents now returns invoice currency alongside amount so an agent can FX-normalise before comparing to transactions.amount. Description updated to make the requirement explicit. Mirrored in the UI: AttachToTransactionDialog ranks same-currency rows by amount distance and pushes cross-currency rows to the bottom of the list. Skipped: - Two-migration window for FK action change is acknowledged as resolved by the bot; deploy-atomicity is an ops concern, not code. - Period-lock check in attach/detach: realistic compliance concern is already covered by the existing immutability trigger (post-categorize) and by the engine's period-lock enforcement (categorize itself). A dedicated period check on pre-categorize attach would only guard against pinning a doc to a tx in a closed period — defensible defense-in-depth, but no active BFL violation. Left for a follow-up. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): address PR #397 round-5 compliance review Three of four findings actioned. The fourth (block_document_deletion verification) is already covered by 20240101000017_enforcement_triggers.sql which raises when document_attachments.journal_entry_id IS NOT NULL on a posted/reversed entry — confirmed via grep, no code change needed. 1. categorize/route.ts: propagation no longer fires-and-forgets. If document_attachments.journal_entry_id can't be set after the JE has been committed, the response now carries a document_link_warning field with a Swedish retry message. The JE is already committed so we can't roll back, but the client can no longer mistake a partial attach for a clean categorize. 2. Rättelse audit trail (BFL 5 kap 5 §): both the REST POST handler and the MCP commit executor now append a TransactionDocumentReplaced event to processing_history whenever a non-null document_id is overwritten, with previous_document_id and new_document_id in the payload. Best-effort — logging failure must not roll back the (compliant) attach. The previous doc id is also returned in the response so callers see what was displaced. 3. MCP staging preview now exposes the existing doc's identity (existing_document_id, existing_document_file_name) plus an explicit existing_document_is_rakenskapsinformation flag, so a human approver sees "replaces X.pdf with Y.pdf" rather than just a will_overwrite_existing boolean. Mirrors BFL 5 kap 5 § informed-rättelse intent. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): close trigger SELECT race (PR #397 round-6) The enforce_transactions_document_immutability trigger SELECTed document_attachments.journal_entry_id without a row lock. A concurrent UPDATE setting journal_entry_id on that row could commit between the trigger's SELECT and its RAISE, letting a detach slip through against a document that just became räkenskapsinformation. Add FOR SHARE to the SELECT inside the trigger. A concurrent journal_entry_id write blocks on our share lock until our transaction commits, so either we observe the propagation and raise, or we run first and the propagation observes our committed detach (which is fine because journal_entry_id was still null at that point). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(invoice-inbox): bidirectional immutability + richer staging preview (PR #397 round-7) Two of six round-6 findings actioned. The other four are recurring architectural recommendations (atomic audit-log writes, background reconciliation jobs, migration consolidation, anti-join via materialized view) that are properly scoped as follow-up work. 1. document_attachments side of the immutability link (BFL 5 kap 6 § works in both directions). New trigger enforce_document_journal_entry_immutability blocks UPDATE OF journal_entry_id when going from non-null to NULL or to a different uuid. The original null→uuid path (initial propagation in the categorize / commitAttach flows) still works. Migration 20260506130000. 2. gnubok_attach_document_to_transaction staging preview now joins on invoice_inbox_items.extracted_data and surfaces vendor/amount/currency/ invoice_date alongside the existing doc filename/mime metadata. Gives the human approver the same hints the agent saw before choosing the attachment. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
b05f0e59b9 |
fix(mcp): correct fiscal_periods column ref + remove agent auto-commit (#394)
* fix(mcp): correct fiscal_periods column ref + remove agent auto-commit The MCP `gnubok_list_fiscal_periods` tool selected a non-existent `fiscal_periods.status` column, causing `column fiscal_periods.status does not exist` errors when agents called it. Fixed by selecting the real columns (`is_closed`, `locked_at`, `closed_at`, `opening_balances_set`) and deriving `status` in code. Also removes the agent auto-commit feature (settings card, gating logic, DB columns, tests). In its current shape only `create_customer` was auto-commitable, so the toggle changed nothing meaningful in practice while implying a level of agent autonomy that wasn't actually granted. The risk-tier infrastructure on `pending_operations` (actor model, risk_level) is kept since it's still used by the /pending UI filters. Migration `20260505120000_drop_agent_auto_commit.sql` drops: - pending_operations.auto_commit_eligible - pending_operations.auto_committed_at - company_settings.agent_auto_commit_enabled - company_settings.agent_auto_commit_max_amount Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(mcp): explicitly pick fields in gnubok_list_fiscal_periods response Greptile flagged that `...p` spreads raw DB columns (`is_closed`, `locked_at`, `closed_at`) into the tool response alongside the derived `status`. Drop the spread for an explicit field list so the tool contract is the derived status only — agents don't need to reason about raw columns, and future SELECT additions won't silently leak. Also drops `closed_at` from the SELECT since it wasn't read. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
5c52f24a49 |
feat(mcp): agent-native improvements — progressive discovery, widgets, skills, telemetry (#393)
* feat(mcp): agent-native improvements — progressive discovery, widgets, skills, telemetry
Four coordinated streams of MCP server improvements that move gnubok toward
agent-first design, grounded in Anthropic's Nov 2025 "Code execution with MCP"
article and the May 2026 MCP conference talk.
Context budget — minimize tools/list payload
- New gnubok_search_tools: progressive discovery with name|summary|full detail
levels and scope filtering. Agents pull only the schemas they need.
- Trimmed all 50 tool descriptions from multi-paragraph blocks (avg ~500-1000
chars) to one-sentence summaries (avg ~120 chars). Args/Returns/Examples
blocks dropped — they duplicated inputSchema.
- outputSchema declared on every tool; structuredContent emitted on every
successful tools/call (was previously only widget-tagged tools).
- protocolVersion bumped to 2025-06-18 (negotiates back to 2024-11-05).
- Workflow examples consolidated into initialize.instructions.
- Net effect: tools/list payload ~43 KB / ~10.8K tokens for 51+ tools, with
headroom guard at 20K tokens.
MCP applications — server-shipped UI widgets
- New widgets/ directory with typed UiWidget contract; receipt-matcher moved
out of widget-html.ts (which was deleted) into widgets/receipt-matcher.ts.
- New gnubok_vat_review_widget tool + interactive momsdeklaration widget
(all 8 rutor with summary card, theme-aware light/dark, copy buttons).
- resources/list and resources/read iterate uiWidgets dynamically — adding
the next widget is a single file drop.
Skills over MCP — domain-knowledge primitive
- 5 user-facing SKILL.md-style workflow guides authored from existing
.claude/skills/swedish-* development skills:
• month-end-close — book → reconcile → VAT (monthly filers) → lock
• quarterly-vat-review — ruta-by-ruta map, deadlines, common errors
• year-end-close — bokslut, bokslutstransaktioner, lock → year-end
→ opening balances → close (irreversible)
• invoicing-rules — ML 17 kap. 24 §, customer types, ROT/RUT, Peppol
• payroll-monthly — salary run → calculate → review → AGI XML
- gnubok_list_skills (with optional tag filter) + gnubok_load_skill(slug).
Both unscoped — available to any authenticated key.
- Each skill also exposed as MCP resource at gnubok://skill/<slug>
(text/markdown) for forward compatibility with a future native
skills/list primitive.
Tool-call telemetry — measure before optimizing further
- Three new CoreEvent types (mcp.tool_called, mcp.tools_list_called,
mcp.resource_read), all persisted to event_log (30-day TTL, RLS-scoped).
- Fire-and-forget emission from the dispatcher — never blocks JSON-RPC
response, double-guarded against handler failures.
- tools/call instrumented at all four exit points (success, execution
error, scope denied, unknown tool). Latency measured tightly around
tool.execute() — excludes dispatcher overhead.
- tools/list logs returned tool count (informs progressive-discovery
adoption); resources/read logs URI + kind discriminator (widget /
skill / data / unknown).
- No PII or secret material in payloads — only metadata.
Out of scope (explicitly deferred):
- Code-mode SDK (no production code-mode hosts to consume it yet).
- Elicitations (require streamable HTTP transport — bigger architectural lift).
- DB lockdown / RPC funnel (foundational; should follow once telemetry tells
us where writes actually flow).
- CRUD → intent endpoints (frontend coupling — multi-PR effort).
Tests: +37 new unit tests across search-tools, output-schema, payload-size,
vat-review-widget, skills, telemetry. Existing receipt-matcher test updated
for the new structuredContent contract. 2,615 unit tests passing.
Production build green. No new lint warnings or errors in changed files.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(mcp): address PR #393 review findings
Greptile (P1 + P2) and the Swedish accounting compliance bot flagged 9 issues
across security, data correctness, and skill content. All addressed:
Security (P1)
- gnubok_search_tools: scope filter now fails closed when __keyScopes is
absent. The earlier permissive default leaked the full tool inventory if
the dispatcher's hard-coded name check ever silently broke. Marker presence
is part of the contract — explicitly empty array also hides scoped tools.
Two new test cases pin the fail-closed behaviour.
Compliance — VAT (data correctness)
- get_vat_report now aggregates 2614/2624/2634 (reverse-charge output VAT)
and exposes them as ruta30/ruta31/ruta32 per SKV 4700. ruta48 also picks
up 2647 (missing before). ruta49 formula corrected to
(10+11+12+30+31+32) − 48. The widget renders the new rutor between the
Utgående and Ingående sections.
- Widget ruta 05 sub-label updated from "3001+3002+3003" to "all momspliktig
försäljning oavsett skattesats" — ruta 05 covers all domestic taxable
supplies, not just direct-rate sales.
- Refactored: extracted computeVatReport() helper used by both
gnubok_get_vat_report and gnubok_vat_review_widget. Removes the
rename-fragile tools.find() lookup at runtime.
Compliance — payroll
- payroll-monthly skill: replaced "born 1958 or earlier = 10.21%" (the 2024
formulation) with the statutory rule "age 66+ on 1 January of the income
year (67+ from income year 2026)". Removed the unsourced "age 16–18:
11.78%" row in favour of a current växa-stöd description with explicit
Prop. 2025/26:34 reference and a "verify against current Skatteverket
tables" caveat.
Compliance — skills text
- invoicing-rules: added explicit footnote on the 1 April 2026 livsmedel
rate change. Restaurang/servering stays at 12 %; livsmedel sold in other
forms drops to 6 %. Per Prop. 2025/26:55.
- year-end-close: clarified periodiseringsfond cap as "25 % of överskott
before this year's avsättning" (IL 30 kap.), removing the ambiguous
"skattemässigt resultat" phrasing that could be misread as a circular
after-fond computation.
Schema correctness
- STAGED_OPERATION_SCHEMA.required gains "staged" — every path through
stagePendingOperation returns the field, so the schema now matches the
contract that MCP clients validate against.
Tests: 2,617 passing (+2 for the search-tools fail-closed cases).
Production build green.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(mcp): address PR #393 round-2 compliance review
Round 2 of the swedish-compliance bot ran after the previous fix-push and
flagged four substantive items + a recommendation. All addressed:
VAT computation (data correctness)
- ruta05 expanded beyond 3001/3002/3003 to cover the common BAS taxable-
revenue accounts (3001-3008, 3041-3048, 3051-3058, 3071-3078). Companies
that book to 30xx alternates were previously under-reporting taxable
turnover; now all standard BAS taxable-revenue numbering contributes.
- One-sided reverse-charge warning: when output VAT is booked on
2614/2624/2634 (rutor 30/31/32 > 0) but the matching calculated input
VAT (2645) is zero, computeVatReport now returns a Swedish-language
warning string. ruta49 is inflated in this case — the warning surfaces
the most common reverse-charge error per the swedish-vat skill. The
widget renders warnings in a terracotta panel above the summary card.
- computeVatReport exported and a focused unit test added — exercises 2647
inclusion in ruta48, reverse-charge balanced/unbalanced cases, and the
expanded ruta05 mapping. Fills the gap that prior tools/call integration
tests couldn't reach.
Skill content
- payroll-monthly Step 6: BAS journal-entry example no longer hard-codes
31.42 % on the 7510/2730 lines. The avgift line is now described as
"avgift_base × applicable_rate per employee" with explicit aggregation
semantics for runs that mix full-rate and reduced-rate employees.
Aligns with the Step-4 reduced-rate caveats already in place.
- invoicing-rules ROT/RUT block: replaced the bare "30 % / max 50 000 SEK"
text with the full year-by-year picture — RUT 50 % / 75 000 max, ROT
baseline 30 % / 50 000 max, 2024 H2 doubled ceiling, 2025 May–Dec
enhanced 50 % rate. Defaults to "verify against current Skatteverket
table" rather than a single hard-coded rate.
False positives in the round-2 review (no fix needed; documented for the
record):
- "Old vat_report path still computes ruta48 without 2647" — the old path
was replaced by computeVatReport in the previous push; the bot was
reading the diff hunk and conflating it with current behaviour.
- "Widget prose says ruta49 = (10+11+12) - 48" — no such prose exists in
widgets/vat-review.ts. The skill body has the correct
(10+11+12+30+31+32)-48 formula.
- "Including 'reversed' status entries in VAT aggregation may over-count
cross-period storno" — current behaviour is correct per Skatteverket
period-aligned filing: the reversed original stays in its own period,
the matching storno (status 'posted') lands in the reversal period,
and they net to zero across the full year. Adding code comment to
document.
Tests: 2,623 passing (+6 for computeVatReport unit tests). Production
build green.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(mcp): address PR #393 round-3 compliance review
Compliance bot re-ran after the round-2 push and flagged three items:
VAT computation
- One-sided reverse-charge warning previously only checked 2645 (EU
acquisitions). For domestic reverse charge per ML 16:13 (byggtjänster,
electronics > 100k SEK, etc.) the matching input lands on 2647 — a
correctly-balanced 2614+2647 booking would have falsely fired the
warning. Fixed: warning now triggers only when *both* 2645 and 2647
are zero. Updated message text mentions both accounts. Added a test
case asserting the no-warning path for 2647-only-input.
Skill documentation drift
- quarterly-vat-review skill body still showed ruta05 source as
"3001 + 3002 + 3003" while the runtime computeVatReport sums 32 BAS
taxable-revenue accounts. Updated the skill table to read
"3001–3008, 3041–3048, 3051–3058, 3071–3078" so the auditor-facing
docs match the implementation.
outputSchema upgrade
- gnubok_get_vat_report and gnubok_vat_review_widget previously declared
outputSchema as the bare { type: 'object' }. Replaced with a shared
VAT_REPORT_OUTPUT_SCHEMA constant declaring period, period_label, all
11 rutor (with descriptions referencing source accounts), summary,
and warnings. Modern MCP clients that validate structuredContent
against outputSchema now have an accurate contract. Added a test
asserting the schema is non-trivial and declares every ruta the
runtime returns.
Tests: 2,625 passing (+2 for the 2647 warning path and the
outputSchema shape assertion). Production build green.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(mcp): address PR #393 round-4 compliance review
The compliance bot re-ran after the round-3 push with a fresh batch.
Real findings fixed; false positives documented.
VAT computation
- Removed 3004 (Försäljning inom Sverige, momsfri / VAT-exempt) from
RUTA_05_ACCOUNTS — round-2's expansion accidentally included it. Ruta 05
is the *taxable* base; exempt sales must NOT contribute. New test pins
the exclusion.
- Added 3106 (taxable EU goods supply, momspliktig) to RUTA_05_ACCOUNTS.
Used when EU buyer's VAT number is invalid or buyer is private.
- Added ruta35 — EU intra-community goods supplies, momsfri (account
3108). Previously omitted entirely from the rutor schema; SKV 4700
has it as a distinct box separate from ruta 39 (services) and ruta 40
(export outside EU). VatReportResult, VAT_REPORT_OUTPUT_SCHEMA, the
widget table, the copy-summary block, and the quarterly-vat-review
skill table all updated. New test covers 3108 → ruta35 mapping.
- Strengthened the comment on the posted+reversed status filter to
document why current behavior is correct per ML 2023:200 and
faktureringsmetoden (the bot's cross-period storno concern is a false
positive — see commit message rationale below).
Skill content
- invoicing-rules: added explicit BFL 5 kap. 6–7 § / ML 17 kap. 22–23 §
note that the kreditfaktura itself consumes a sequential number from
the same (or dedicated KR-) fakturaserie. The KR- prefix is a display
convention; the underlying löpnummer must be unbroken just like the
regular series.
- payroll-monthly: added the missing "born 1937 or earlier → 0 %"
cohort to the rate breakdown. Previously could lead a payroll run to
over-pay avgifter on the oldest cohort.
False positives in the round-4 review (verified, not changed)
- 2644/2648 as reverse-charge inputs: verified against gnubok's actual
BAS chart (lib/bookkeeping/bas-data/class-2-equity-liabilities.ts).
2644 does not exist; 2648 is "Vilande ingående moms" (dormant input
VAT for cash method), not RC at 6 %. Canonical RC inputs are 2645 (EU)
and 2647 (domestic) — both already covered.
- Cross-period storno over-count: per ML 2023:200 + Skatteverket
faktureringsmetoden, the original sale's VAT belongs to the invoice-
date period; the kreditfaktura's reduction belongs to the storno-date
period. Including 'reversed' status entries (which still have their
original date) is therefore correct. *Excluding* them would
under-report the original period and over-credit the reversal period.
Added a multi-line comment in computeVatReport documenting this.
Tests: 2,627 passing (+2 for ruta35 mapping and 3004 exclusion).
Production build green.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
||
|
|
432a8b60dc |
feat(skatteverket): rewrite AGI flow against real Skatteverket RAML (#391)
* feat(skatteverket): rewrite AGI flow against real Skatteverket RAML The previous AGI client posted JSON to URL paths that don't exist on Skatteverket's gateway and used invented field names. POST /underlag actually accepts application/xml, and the lock/kvittenser operations live on the separate hanteraredovisningsperiod API. Verified against dev_docs/arbetsgivardeklaration-inlamning(1.7.7) and arbetsgivardeklaration-hantera-redovisningsperiod(1.2.8) RAMLs. - Replace fictional types with real schemas (kontrollresultat, granskningsunderlag, kvittenser, error envelope) - Rewrite agi-client into 9 functions matching the documented flow: /underlag (XML) -> kontrollresultat -> spara -> skapaGranskningsunderlag -> kvittenser, plus las/lasUpp on the hantera API - Drop agi-mappers entirely; lib/salary/agi/xml-generator.ts already produces schema-valid XML, so the extension just feeds agi_declarations.xml_content to POST /underlag - Extend skvRequest with a contentType option so AGI can post XML - AGIPanel state machine: underlag_submitted -> awaiting_signing -> signed, with kontrollresultat polling and normalized findings - Add the agd OAuth scope (confirmed from SKV's Tjanstebeskrivning Arbetsgivardeklaration inlamning v1.7, section 4.1.2.2) - Add Skatteverket connect step to NewUserChecklist alongside the existing SIE/old-system import and bank steps; track hasSkatteverketConnected in OnboardingProgress - Update orchestrator route + tests to point at the new /agi/submit endpoint - Declare new optional base-URL env vars in the manifest Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(skatteverket): address PR review findings on AGI flow - Surface INCORRECT_DATA felrapport link in AGIPanel skapaGranskningsunderlag returns 409 with a felrapport URL when SKV rejects the underlag. The link was persisted as `signeringslank` with status `underlag_rejected`, but the render condition only fired for `awaiting_signing`, leaving the link unreachable. Add a distinct destructive-styled block so the user can open the felrapport in Mina Sidor. - /agi/underlag DELETE clears local submission state Add optional `period` query param. When supplied, clear `agi_submission_{period}` directly. When not, fall back to scanning recent agi_submission_* keys for the matching inlamningId. Without this, an aborted underlag left a stale `underlag_submitted` entry in extension_data and the UI couldn't progress. - Re-add salary-run status guard inside loadAGIXml The orchestrator at app/api/salary/runs/[id]/agi/submit/route.ts has this check, but the extension endpoint is also reachable directly from AGIPanel and must enforce it itself. Per BFL 5 kap and SFL 26 kap, AGI must reflect finalised payroll data; submitting from a draft/cancelled run would emit incorrect figures. - Move agi_declarations.status='exported' from /agi/submit to /agi/spara Setting status on underlag-ingest was wrong because a DONE_REJECTED kontrollresultat would leave the row falsely marked as exported. The transition now happens only after the spara call commits the underlag to Eget utrymme. /agi/spara accepts salaryRunId in the body for the fast path and falls back to scanning agi_submission_* state otherwise. - Move salary_runs.agi_submitted_at stamp to kvittenser observation The orchestrator was stamping at underlag-ingest, but no later code updated the column on signing. Removed the orchestrator stamp; the /agi/kvittenser handler now stamps salary_runs.agi_submitted_at to kvittens.signeradTid (mirroring SKV's own timestamp) when it pins the receipt to the matching agi_declarations row. - Tighten misleading JSDoc in agi-client.ts taBortSparadInlamning is on the inlämning API, not hantera; the old layout grouped it under a "hantera API" heading and tripped an automated reviewer. Restructured into separate "period management" and "cleanup" blocks. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(skatteverket): address Swedish compliance review on AGI flow Follow-up to review on https://github.com/erp-mafia/gnubok/pull/391. - Migration adds 'pending_signature' to agi_declarations.status Reusing 'exported' for the spara→kvittens interval misstated the filing outcome — Eget utrymme is a staging area, not a filing — which conflicts with BFNAR 2013:2 kap 8 / BFL 5 kap 5§ behandlingshistorik faithfulness. /agi/spara now sets 'pending_signature'; /agi/kvittenser later promotes to 'submitted' when a uuidKvittens is observed. - AGIPanel auto-polls /agi/kvittenser at 30s, 2 min and 5 min after the signing link is created Previously the kvittens (and therefore salary_runs.agi_submitted_at) was only stamped if the user manually returned to the panel and clicked "Hämta kvittens". Without that follow-up the audit trail showed a NULL submitted-at for an AGI that had actually been filed. Background polls capture the kvittens for the common case where the user signs in Mina Sidor and never returns to gnubok. Cleanup on unmount via useRef + useEffect. - Distinct MISSING_SCOPE error code on 403 invalid_scope Existing tokens lack the new 'agd' scope and surface as a generic ACCESS_DENIED today. The compliance reviewer pointed out that operators may interpret this as a data error and submit a corrected AGI with altered figures. New SkatteverketAuthError code maps SKV's invalid_scope body to a clear "reconnect via Inställningar → Skatteverket" message; routes to 401 (token-level remediation). - Refine deadline copy in AGIPanel The standard AGI deadline is the 12th regardless of company size; the 17th only applies in January and August for employers with turnover ≤ 40 MSEK. Surface that nuance instead of saying just "12:e". Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(skatteverket): server-side kvittens reconciliation cron Round 2 of compliance review on https://github.com/erp-mafia/gnubok/pull/391. - /api/extensions/skatteverket/agi/kvittenser/cron Walks every agi_declarations row in 'pending_signature' status, fetches kvittenser via the matching token, and on a hit promotes the row to 'submitted' + stamps salary_runs.agi_submitted_at. Authoritative source for the audit trail per BFNAR 2013:2 kap 8 / BFL 5 kap 5§ — the AGIPanel client-side timers from the previous round remain as the fast-path UX, but no longer carry the audit-trail responsibility on their own. Per-row errors are skipped, not abort-the-run. 50s budget. Scheduled every 2 hours in vercel.json. - AGIStatus union now includes 'pending_signature' Without this update, downstream code reading the union would have rejected the new status as unknown. The migration extending the DB CHECK constraint shipped in the previous commit; this brings the type layer into sync. - Stale comment update in AGIPanel.tsx Referred to status='exported' from before the rename. Now reads 'pending_signature', matching the actual handler behavior. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(skatteverket): close audit-trail gaps from compliance round 3 - Cron now writes submitted_by from the token-owning auth.users row Previously left NULL with a "system actor" comment. The token row was created when the operator authenticated with BankID, and the kvittens' signeradAv refers to the same person — so writing the user_id from skatteverket_tokens captures actor traceability without inventing a system identity. Closes the BFL 5 kap 6§ / BFNAR 2013:2 kap 8 gap on cron-reconciled rows. - /agi/spara monotonicity guard Adds .in('status', ['generated', 'exported']) to the row update so a delayed /agi/spara call after the cron (or interactive /agi/kvittenser) has already promoted the row to 'submitted'/'accepted' won't silently regress it back to 'pending_signature'. behandlingshistorik must advance only. - DONE_REJECTED / DONE_FAILED → status='rejected' /agi/kontrollresultat handler now flips the matching agi_declarations row to 'rejected' on a terminal SKV failure, using the same cached-submission-state lookup pattern /agi/spara already uses. Without this the row sat at 'generated' indefinitely even though SKV considered the underlag failed. Same monotonicity guard prevents regressing a successfully-filed row. - Deadline criterion: lönesumma, not omsättning AGIPanel pendingText. SFL 26 kap's relaxed-deadline criterion (17:e in Jan/Aug) is the employer's total taxable wages, not turnover. Internal reference (.claude/skills/swedish-payroll/references/agi-filing.md) used the colloquial "turnover"; statutory wording is "lönesumma". Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(skatteverket): close round-4 audit-trail and UX gaps - agi_submitted_at: NULL when signeradTid absent Both /agi/kvittenser handler and the kvittens cron previously fell back to new Date().toISOString() if SKV's kvittens lacked signeradTid. Substituting wall-clock now() falsifies the filing moment in behandlingshistorik (BFNAR 2013:2 kap 8 / BFL 5 kap 6§). Now leaves the column NULL and logs a warning. Status flip to 'submitted' still happens — the audit gap was timing only. - Proactive missing-agd-scope banner SkatteverketConnectPanel and AGIPanel now warn when the stored token lacks the agd scope. Tokens issued before the agd rollout would otherwise 403 with invalid_scope at submission time, often too close to the AGI deadline. SkatteverketConnectPanel mirrors the existing "skattekonto saknas" pattern; AGIPanel surfaces a banner in the connected state and links to /settings/skatteverket. - Granskningsunderlag isError keys on tillstand only Previous check mixed HTTP 409 with the INCORRECT_DATA tillstand string. A future SKV addition like RECEIVING returned with HTTP 200 would have slipped through as awaiting_signing. Now keys solely on tillstand: only LOCKED_FOR_SIGNING / UNLOCKED are treated as signable; everything else (INCORRECT_DATA, RECEIVING, CALCULATING, SIGNING) routes to underlag_rejected. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(skatteverket): close round-5 audit-trail and recovery gaps - agi_submitted_at: stamp with reconciliation time when signeradTid absent Round 4 left the column NULL on missing signeradTid to avoid falsifying the signing moment. Round 5 pointed out that NULL hides that the filing *occurred* — also a behandlingshistorik integrity violation. Resolution: presence of uuidKvittens proves SKV signed and accepted the AGI, so we stamp with signeradTid || now() and warn-log when fallback is used. Both /agi/kvittenser handler and the kvittens cron. - Persist signeradAv + full kvittens in agi_declarations.response_data submitted_by is the auth.users UUID we have on hand (the polling / reconciling user). The legally load-bearing signer identity is kvittens.signeradAv (a personnummer) — which the token user_id does NOT necessarily match (e.g. bookkeeper vs deklarationsombud). The existing response_data jsonb column now holds the full kvittens record, preserving signeradAv for the audit trail (BFL 5 kap 6§ / BFNAR 2013:2 kap 8) without a schema change. Cron path also marks reconciledBy='cron'. - /agi/spara monotonicity: allow recovery from 'rejected' Previously .in('status', ['generated', 'exported']) excluded rejected rows, so a successful re-submission after a prior rejection couldn't promote the row to pending_signature — it silently stayed rejected. The xml-route reuses the same agi_declarations row when re-generating XML, so this is the realistic recovery path. Added 'rejected' to the allowed-from list. 'submitted'/'accepted' still blocked (no regression from filed states). - Fix misleading agi-client.ts comment Claimed users could "fix the errors in Mina Sidor" after a DONE_REJECTED save. Mina Sidor doesn't expose in-place editing; the correct recovery is to regenerate XML and resubmit. Updated the agiSparaUnderlag JSDoc to describe the actual flow. - Deadline copy: "vars sammanlagda lönesumma understiger 40 MSEK" Reads more cleanly than "≤ 40 MSEK" and matches the phrasing the compliance reviewer suggested. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(skatteverket): tighten /agi/spara guard and clarify deadline copy (round 6) - Drop 'exported' from /agi/spara allowed-from states Audit confirmed no code path writes status='exported' today; the value is preserved in the schema (and union) for the legacy manual-download path that no longer has a writer. Allowing the spara handler to flip an 'exported' row to 'pending_signature' would conflate two distinct filing attempts on a single row, weakening the chain of custody (BFL 5 kap 6§). Tightened to .in(['generated', 'rejected']) — same recovery path for re-submission after rejection, no path for the dormant state. - Deadline copy: explicit "per år" qualifier The 40 MSEK threshold is annual lönesumma, not per-payment. Adding "per år" closes the (admittedly thin) misread the compliance reviewer flagged. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
c03582b5c7 |
Fix/percistent mcp connection (#392)
* feat(oauth): add support for refresh tokens in OAuth flow and update database schema * feat(prompts): add MCP prompts and corresponding functionality for prompt retrieval * feat(auth): enhance error handling for refresh token operations and validation |
||
|
|
fa7d4075cf |
Supp/invoice bfl errors (#390)
* feat(accounting): update accounting method validation and messaging for aktiebolag and enskild firma * Remove AI subsystem and related code - Deleted AI proposals and requests persistence logic from `lib/ai/proposals/persist.ts`. - Removed re-validation logic for proposals in `lib/ai/proposals/re-validate.ts`. - Cleaned up schemas related to AI flows in `lib/api/schemas.ts`. - Removed AI-related fields from bookkeeping engine in `lib/bookkeeping/engine.ts`. - Eliminated AI event types from `lib/events/types.ts`. - Updated tests to reflect the removal of AI-related functionality in `lib/extensions/__tests__/sectors.test.ts`. - Adjusted initialization logic in `lib/init.ts` to exclude AI proposal handler registration. - Cleaned up transaction ingestion logic in `lib/transactions/ingest.ts` to remove AI flow checks. - Updated helper functions in `tests/helpers.ts` to remove AI-related settings. - Removed AI-related types and interfaces from `types/index.ts`. - Added migration script to drop AI-related tables and settings from the database. * fix(migrations): ensure foreign key constraint is dropped before removing AI tables * feat(invoice-inbox): implement deterministic invoice field extraction and inbox provisioning - Added `extract-invoice-fields.ts` for extracting fields from PDF invoices using regex and pdfjs-dist, replacing the previous AI classifier. - Introduced `inbox-provisioning.ts` to manage company inbox addresses and rotation of inboxes using Supabase RPCs. - Created `resend-inbound.ts` for handling inbound email events and attachments via the Resend API. - Defined the extension manifest for the invoice inbox, specifying required environment variables and descriptions. - Migrated database schema to remove AI-related columns and tighten the status enum in `invoice_inbox_items`. * feat(invoice-inbox): remove AI-specific columns and tighten status enum * fix(skattekonto): remove manual entry creation reference from transaction input * fix(schemas): remove accounting method validation for aktiebolag in UpdateSettingsSchema |
||
|
|
f3fd4c0822 |
feat(salary, skatteverket): per-day absence + AGI Frånvarouppgift + skattekonto + hardening (#388)
* feat(salary): per-day absence tracking with calendar UX Replace aggregated-day absence counts with per-day records so payroll calculations can correctly enforce Swedish legal rules that depend on actual dates: karensavdrag once per sjuklöneperiod, återinsjuknande within 5 calendar days, allmänt högriskskydd cap of 10 karensavdrag per rolling 12 months, day-8 läkarintyg flag, day-15 transition to Försäkringskassan. Adds: - salary_absence_days table (RLS, dedup unique on employee+date+type) - /api/salary/employees/[id]/absence CRUD route - deriveAbsenceLineItems helper that walks per-day records into sjuklöneperioder and emits correctly-classified line items, with the existing absence-calculator formulas reused for VAB / parental - Per-employee pay-spec detail page with month-grid AbsenceCalendar - Calculate route now derives line items from the calendar before running the salary engine, replacing the prior sumQuantity model - Salary run GET surfaces the formatted Skatteverket arbetsgivare ID so downstream UI can build extension URLs without a second round-trip - GET /salary/runs/[id]/employees/[employeeId] for the detail page Tests: 15 new unit tests covering segment merge, återinsjuknande within 5 days, högriskskydd cap, FK transition flag, läkarintyg flag, VAB/parental semesterlönegrundande ceilings. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(skatteverket): harden API client + add NEXT_PUBLIC_SKATTEVERKET_ENABLED feature flag Three hardening fixes from the prior audit, plus a runtime extension toggle for phased rollout. api-client.ts: - Map 429 to a new SkatteverketAuthError code RATE_LIMITED with a Swedish user message. The 4 req/sec local rate limiter normally prevents this, but the per-consumer gateway quota can still hit. - Extend the error union with TOKEN_CORRUPTED for the token-store fix below. token-store.ts: - Surface decryption failures instead of silently returning null. A rotated key or tampered ciphertext used to look like "not connected"; callers now get TOKEN_CORRUPTED with a clear "anslut igen med BankID" message and a structured log line for ops. Extension dispatcher (app/api/extensions/ext/[...path]/route.ts): - Per-extension feature flag table. When NEXT_PUBLIC_SKATTEVERKET_ENABLED is not exactly "true", the dispatcher returns 503 with code EXTENSION_DISABLED, letting ops disable a single integration mid- rollout without redeploying or removing it from extensions.config.json. UI panels (SkatteverketPanel, AGIPanel) detect the 503 and render an empty state. Tests: 7 api-client cases (401/403/403-Behörighet/429/5xx/200/auth-error codes) + 2 token-store cases (no-row → null, corrupted → TOKEN_CORRUPTED). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(salary): emit AGI Frånvarouppgift per SKV 4785, add AGIPanel for one-click submission AGI XML upgrade: - Emit <gem:Franvarouppgift> top-level blocks for VAB and parental leave events sourced from salary_absence_days, per SKV 4785 + technical doc. Element order matches the spec example file. TILLFALLIG_FORALDRAPENNING for VAB / FORALDRAPENNING for parental, with FranvaroTimmarTFP (FK825) or FranvaroTimmarFP (FK827) for hours. Stable 1-based specifikationsnummer per (employee, period), date-sorted. Skipped entirely for periods before 202501. - Sick days are NOT emitted (they go to Försäkringskassan). - FK499 TotalSjuklonekostnad now derived from sick_day2_14.quantity × dailyRate × 0.80 instead of Math.abs(amount). The line-item amount is the net deduction (lostPay − sjuklon), not the cost, so the prior formula understated by a factor of four. AGI submission UI: - New AGIPanel mirroring SkatteverketPanel's validate → draft → lock → BankID-sign → poll-submitted flow. Detects 503 EXTENSION_DISABLED and renders a clear empty state. Replaces the bare "Skicka till Skatteverket" button on /salary/runs/[id], keeping the AGI XML download as a sibling for archival / manual upload fallback. - Salary run rows now link to the per-employee detail page added in the previous commit. Tests: 14 new agi-xml cases covering element order, type↔hour-field mapping, specifikationsnummer ordering, fractional-hour formatting, range clamping (0.01-24.00), period guard at 202501 boundary, placement after Blankett blocks, multi-employee date ordering, required-fields invariant, omission when no events. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(skatteverket): skattekonto integration — read-only saldo + transactions, daily sync, per-row bokför Adds read-only Skattekonto v2.1 access via the existing BankID OAuth flow (extends the OAuth scope with `skattekonto`). Daily background sync pulls saldo + transactions, dedupes on (company_id, dedup_key), and surfaces the data in a /skattekonto dashboard plus a settings panel for connection management. Backend: - skattekonto-client.ts: GET /skattekonton/{omfragad}/saldo and /transaktioner. Felkod 1–5 mapped to Swedish messages via dedicated SkatteverketSkattekontoError. - skattekonto-sync.ts: parallel saldo + transaktioner fetch, UPSERT on (company_id, dedup_key) so kommande rows graduate to tidigare in place. Dedup key uses transaktionsidentitet when available, else sha256 of (date|amount|text). Caches saldo snapshot in extension_data. Emits skattekonto.synced / balance.changed (sign flip) / transaction.upcoming (first appearance) / connection.expired. - skattekonto-booking.ts: keyword→counter-account rules with AB/EF differentiation (2510 vs 2012 for preliminärskatt; 2731/2710/2650 for arbetsgivaravgifter/avdragen skatt/moms; 8423/8313 for kostnads-/intäktsränta). Creates a draft journal entry against BAS 1630, leaves it for the user to review and commit. Throws NO_COUNTER_ACCOUNT instead of guessing when no rule matches. - Daily cron at 0 4 * * * (Swedish 06:00). Double-gated by CRON_SECRET and NEXT_PUBLIC_SKATTEVERKET_ENABLED. Per-company cooldown of 1 hour, time budget 50s, distinct `expired` status for token-exhaustion separate from generic errors. Database: - skattekonto_transactions: company-scoped with RLS, unique (company_id, dedup_key), indexed on (company_id, date DESC) and (company_id, status). journal_entry_id FK with ON DELETE SET NULL so a row can be re-bokförd after entry deletion. Frontend: - /skattekonto/page.tsx: dashboard with saldo card, transactions list (booked + upcoming), per-row "Bokför" action. - /settings/skatteverket: connection panel showing scope/expiry. - Extension toggle in SettingsSidebar (gated by ENABLED_EXTENSION_IDS). Tests: 9 booking-rule cases (counter-account guessing, AB/EF divergence, no-match throw) + 7 mapper cases (dedup key stability, sign convention, kommande→tidigare graduation). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix: address PR review findings Build: - Fix Next.js build failure: Zod refuses .partial() on a refined schema. Replace AbsenceRangeQuerySchema.partial().extend(...) in the absence DELETE handler with a fresh z.object that defines its own optional fields. Greptile findings (PR #388): - skattekonto_transactions UPDATE policy was missing WITH CHECK; without it a user could mutate company_id to one they don't belong to. Edit the original migration for fresh applies + add a follow-up migration that drops/recreates the policy with both clauses (already applied to prod via Supabase MCP). - FK499 TotalSjuklonekostnad now reads sjuklonRate from run.calculation_params (snapshot taken at calc time) instead of a hardcoded 0.80, so an operator override (e.g. CBA-specific rate) is honored. Falls back to 0.80 for older runs without the snapshot. - Rename NEXT_PUBLIC_SKATTEVERKET_ENABLED → SKATTEVERKET_ENABLED so the flag is server-side only. NEXT_PUBLIC_* vars are inlined into the client bundle at build time, which would create split-brain (server 503 vs client still rendering enabled flow) on a flag flip without redeploy. UI panels detect 503 by response code, not by reading the env directly, so no client-visible change is needed. - Add pg-real RLS smoke tests for both new tables (salary_absence_days and skattekonto_transactions): tenant SELECT isolation, UPDATE WITH CHECK enforcement, unique-constraint enforcement, cross-tenant dedup key allowed. Swedish compliance review: - Document the högriskskydd cap interpretation in derive-absence-line-items.ts. We count *sjuklöneperioder* in the rolling 12-month window, matching the law's plain reading ("från och med den 11:e sjukperioden ... görs inget karensavdrag"). An alternative reading counts only periods that actually had karens deducted; that requires persisting per-period karens-deduction state, which gnubok doesn't yet do. The period-count reading can over- suppress, never under-suppress, so it's the safer default. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(test): inline skattekonto fixtures so core-only CI runs without dev_docs dev_docs/ is gitignored, so the skattekonto-mappers test failed in CI when it tried to readFileSync from dev_docs/skattekonto(2.1.0)/examples/. Inline the saldoResponse + transaktionerResponse fixtures verbatim from the spec; the test still verifies our mappers + dedup-key logic against the same shape. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
bb855d2ddc |
Add/ai native supp (#385)
* feat(branding): implement dynamic branding in service worker and reports * feat(auth): enhance API key scopes and add bookkeeping write scope - Updated transaction write scope description to include additional tools. - Enhanced reports read scope description to reflect new functionality. - Introduced bookkeeping write scope with relevant description. - Updated SCOPE_GROUPS to include bookkeeping domain. - Modified TOOL_SCOPE_MAP to include new bookkeeping operations. - Updated validateApiKey function to return api_key_id and api_key_name for better actor attribution. feat(tests): add unit tests for MCP resource registry - Created tests for data resources to ensure all required fields are present. - Added tests for resource query parsing and retrieval. feat(resources): implement MCP resources for company and accounting data - Added capabilities resource to expose API key capabilities based on granted scopes. - Implemented chart of accounts resource to retrieve active BAS chart. - Created company current resource to fetch active company details. - Developed active fiscal period resource to check posting eligibility. - Implemented recent activity resource to fetch latest journal entries, invoices, and transactions. - Added VAT treatments resource to provide available VAT rates per customer type. feat(pending-operations): introduce risk tiers for operations - Added risk level classification for pending operations to determine auto-commit eligibility. - Implemented functions to classify operation risk levels and identify high-risk operations. feat(migrations): add actor model and risk tier to pending operations - Updated pending_operations table to include actor type and risk level columns. - Enhanced audit_log to mirror actor information for compliance. - Modified validate_and_increment_api_key function to return actor details. - Expanded operation types in pending_operations to include new high-risk operations. * feat: add auto-commit functionality for low-risk pending operations - Implemented shouldAutoCommit function to determine eligibility for auto-commit based on operation type, actor type, and company settings. - Created commitPendingOperation function to handle execution of pending operations with consistent status updates. - Added tests for shouldAutoCommit to cover various scenarios including high-risk operations, user actors, company opt-in status, and monetary thresholds. - Introduced new columns in company_settings for agent_auto_commit_enabled and agent_auto_commit_max_amount to allow companies to opt-in for auto-commit functionality. - Added SQL migration to update the database schema for new auto-commit settings. * feat(idempotency): implement idempotency key handling for safe retries and cleanup * feat: expand API key scopes and pending operations for bookkeeping - Added 'suppliers:write' scope to API key scopes for supplier invoice management. - Updated SCOPE_GROUPS to include the new 'suppliers:write' scope. - Introduced new pending operation types for bookkeeping: close_period, lock_period, run_year_end, set_opening_balances, run_currency_revaluation, explain_voucher_gap, uncategorize_transaction, approve_supplier_invoice, credit_supplier_invoice, and convert_invoice. - Implemented corresponding commit functions for the new operations in the pending operations module. - Enhanced PendingOperation type to include actor model and risk level attributes. - Added tests for new functionality, ensuring proper behavior and constraints in the database. * feat: implement unlockPeriod functionality and related tests * feat: add agent auto-commit settings and related functionality * feat: add attention resource with comprehensive summary of outstanding tasks * feat: enhance pending operations with 'committing' status and immutability checks, improve idempotency handling, and add original voucher reference for credit notes |
||
|
|
c86dbdc60d |
feat(branding): add hiddenNavHrefs to hide sidebar items via env var (#382)
Extends the branding service with a hiddenNavHrefs: string[] field so forks can hide sidebar entries (e.g. /salary, /customers) without patching DashboardNav.tsx. Configurable via NEXT_PUBLIC_BRANDING_HIDDEN_NAV as a comma-separated list. Default is [] — vanilla gnubok unchanged. Routes remain reachable; this is a nav-visibility switch only. Settings sidebar is intentionally out of scope. Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
064fb7f7a9 |
Add/white label (#381)
* feat(branding): add BrandingService with default-preserving env layer Introduce lib/branding/service.ts mirroring lib/email/service.ts. Defaults match current gnubok values exactly, so production behaviour is unchanged unless an env var (NEXT_PUBLIC_BRANDING_*, BRANDING_*) or extension override (via registerBrandingService) is set. Resolution order: defaults < env vars < extension override. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(branding): route root layout, manifest, and PWA assets through branding service - app/layout.tsx now reads title, description, themeColor, and apple-touch-icon from getBranding() instead of hardcoded values. - public/manifest.json replaced by dynamic app/manifest.ts so PWA name, short_name, description, theme_color, background_color, and icon paths are resolved at request time. The manifest now serves at /manifest.webmanifest (Next.js convention for the metadata file route). The previous /manifest.json URL is no longer populated; nothing in core references it after this commit. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(branding): route email service and templates through branding service - resend-service.ts: From line uses getBranding().appName instead of hardcoded "Gnubok" in both the with-fromName and bare cases. - invite-templates.ts: subject, HTML header, body, plain text, and the team-invite variants all read from branding (sentence case in prose, uppercased for the styled <p> header). - consent-notification-templates.ts: signature fallback (companyName || branding) for both HTML and plain text variants. Defaults preserve the exact current strings ("Gnubok", "GNUBOK", "gnubok" in their respective contexts) so no email content changes for production. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(branding): route OAuth consent page through branding service The MCP OAuth consent page rendered for Claude Desktop / Claude.ai connector flows now reads the app name from getBranding() for both the HTML <title> and the body copy. Default still produces "gnubok" in lowercase prose, matching current behaviour. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(branding): route auth, dashboard, and onboarding text through branding service Replace user-visible "gnubok" / "Gnubok" references with calls to getBranding(). Touches: - Auth pages (login, register, mfa/enroll): logo src/alt, MFA TOTP friendlyName. - Onboarding (companies/new, invite, sandbox, WelcomeOnboarding, Step2CompanyDetails, NewUserChecklist, BankIdCompanyPicker, ArcimMigrationWorkspace): logo, headings, error/help text. - Dashboard fallback (companyName="gnubok") and settings (backup copy, ApiKeysPanel MCP connector name + login note, CompanyDangerZone, retention-notice). - API routes (support contact subject prefix, enable-banking consent email companyName fallback, AI inbox receipt-request appUrl, pain001 messageId prefix). - MCP server "open the gnubok web app" review message. - Salary/reports filings (AGI Programnamn, KU10 Programnamn, payslip footer, full-archive system metadata, SRU #PROGRAM line). Internal identifiers (cookie names gnubok-company-id / gnubok-invite-token, API key prefix gnubok_sk_, invite token prefix gnubok_inv_, MCP tool names, npm package gnubok-mcp, GNUBOK_API_KEY env name) are deliberately left unchanged — they're stable contracts that whitelabels must not break. Defaults match current behaviour exactly. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(branding): support legal page field-level swaps for entity and contact Privacy and DPA pages now interpolate appName, legalEntity, and privacyEmail from the branding service instead of hardcoding "Gnubok", "Arcim", and "privacy@gnubok.se". Page metadata uses generateMetadata() so titles also reflect the brand. lib/support.ts now falls back to getBranding().supportEmail when SUPPORT_RECIPIENT_EMAIL is unset, so a single BRANDING_SUPPORT_EMAIL env var configures both the support form recipient and the displayed support address. Whitelabels with a different legal jurisdiction or entirely different DPA text should override the page route from an extension. Phase 1 intentionally only supports field-level swaps. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(branding): add WHITELABEL.md and example branding extension WHITELABEL.md: fork checklist, env var reference, the "do not change" list (cookies, API key prefixes, invite token prefixes, MCP tool names, gnubok-mcp npm package, GNUBOK_API_KEY env name), out-of-scope items, the upstream sync workflow YAML to copy into a fork, conflict avoidance guidance, and a verification checklist. extensions/general/_example-branding/: copy-paste starter extension with index.ts (commented placeholder values for registerBrandingService), manifest.json, and README.md. Disabled by default (not added to extensions.config.json); whitelabels cp the folder, edit, and enable. sectors.test.ts: bumped expected extension count 12 -> 13 to account for the new starter extension on disk. The generated registry is unchanged because the example is disabled. The sync workflow YAML is documented inline in WHITELABEL.md rather than checked in as a workflow file. It's only meaningful in a fork -- gnubok itself has nothing to sync from. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(branding): address PR review — lazy support email + escape brand in HTML/XML Three issues from code review: P1 — lib/support.ts: SUPPORT_RECIPIENT_EMAIL was a module-level const, evaluated at import time before extensions register branding overrides via ensureInitialized(). Convert to getSupportRecipientEmail() lazy accessor; update the only caller in app/api/support/contact/route.ts. Extension-supplied supportEmail values now route correctly. P2 — app/api/mcp-oauth/authorize/route.ts: appName was interpolated into the consent page HTML without escapeHtml(), inconsistent with the existing escaping of companyName. Wrap appName.toLowerCase() in escapeHtml() at use sites in <title> and the body paragraph. P2 — lib/salary/agi/xml-generator.ts and lib/salary/ku/ku10-generator.ts: appName placed inside <gem:Programnamn> / <Programnamn> XML elements without escapeXml(), the helper already used for other admin-controlled fields in the same files. Wrap accordingly to prevent malformed XML if a brand name contains XML reserved characters. All admin-controlled inputs only — no user-exploitable path. Defense in depth, not a known incident. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(branding): security follow-up — lazy metadata, SRU/email header sanitization Self-audit after the PR review surfaced four more concerns. Fixes them with the same defense-in-depth posture as the prior review fixes. 1. app/layout.tsx — same eager-evaluation class as P1 support.ts. The module-level `const branding = getBranding()` froze branding before extensions registered, so extension-based overrides for title, description, themeColor, and apple-touch-icon silently never applied. - Convert to generateMetadata() / generateViewport() (lazy, run per request, see extension-registered overrides). - Inline getBranding() inside RootLayout for the apple-touch-icon href so it picks up overrides too. - Add ensureInitialized() at module level so extensions are loaded before the first metadata call. Mirrors the API route pattern. 2. app/manifest.ts — same class. The dynamic manifest function reads getBranding() per request, but if the manifest is requested before any other module has triggered ensureInitialized(), extensions are still unloaded. Add ensureInitialized() at module level. 3. lib/reports/ink2/sru-generator.ts — appName interpolated into the SRU `#PROGRAM` directive without sanitization. SRU's reserved char is `#` (directive marker) and CRLF injects new directives. Wrap in the existing sanitizeString() helper to match the pattern used for other admin-controlled fields in this file (#NAMN, #ADRESS, etc.). 4. extensions/general/email/lib/resend-service.ts — appName and the user-controlled fromName both flow into the From header. Resend's API does its own validation, but defense in depth: strip CRLF and angle brackets via a small sanitizeHeaderPart() helper before building the header string. fromName was a pre-existing surface; appName is new with this whitelabel work. All four are admin-controlled inputs (env vars or extension code), not user-exploitable. No known incidents — defense in depth, and correctness for extension-based whitelabels. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
cd64c0e3fb |
feat(skatteverket): production-ready momsdeklaration submission (#380)
* feat(skatteverket): production-ready momsdeklaration submission
Brings the Skatteverket extension up to a state where it can ship moms
declaration submission to Vercel production. Verified end-to-end against
SKV's Komplett testtjänst — all 8 momsdeklaration operations tested
(kontrollera, spara/hämta/radera utkast, lås/lås upp, hämta inlämnade,
hämta beslutade) plus signing-link return.
Bundles three coherent changes:
1. Skatteverket extension (the main work)
- extensions.config.json: enable `skatteverket`, drop `invoice-inbox`
and `ai-agent` (those were enabled in config but lacked AWS env vars
in prod, so they loaded but failed at runtime)
- lib/reports/vat-declaration.ts: extend ACCOUNT_RUTA to populate
Ruta 06 (uttag 3401–3403), Ruta 20–24 (reverse-charge bases from
4xxx cost accounts), Ruta 50 (import 4545–4547), and Ruta 42
(3404/3994/3980); delete the supplier-type heuristic that made
Ruta 20 and Ruta 23 always 0
- extensions/general/skatteverket/lib/token-store.ts: work around
three real prod schema-drift issues — wrong column on read/delete
(was `company_id`, schema only has `user_id`), missing
UNIQUE(user_id) constraint that makes UPSERT fail (switched to
DELETE+INSERT), missing RLS policies (switched to service-role
client). Refresh path now reuses existing row's company_id when
none is passed.
- extensions/general/skatteverket/index.ts: 9 sites switched from
ctx.companyId to ctx.userId for the token-store key; pass
companyId from the OAuth callback
- extensions/general/skatteverket/types.ts + components/reports/
SkatteverketPanel.tsx: align field names with v1.0.24 RAML
(signeringsLank/kontrollResultat/resultat/kod/status/beskrivning).
Without this, the signing link never displayed.
- SkatteverketPanel: add Lås upp + Radera utkast + Hämta utkast +
Hämta beslut buttons so the full lifecycle is reachable from the UI
- lib/reports/__tests__/vat-declaration.test.ts: rewritten to match
the refactored calculator; new fixtures for cost-account-based
reverse charge (Ruta 20/21/22/23/24), Ruta 50 import, Ruta 06
uttag, Ruta 42 expansion; SKV §4.1.1.4 cross-field contract checks
- supabase/migrations/20260428120000_skatteverket_tokens_user_id_unique.sql:
idempotently adds the missing UNIQUE(user_id) constraint
- scripts/*: dev-only helpers used during the prod-of-test
verification (create test company, seed VAT data, inspect token
state, etc.)
2. Journal-entries cancelled-status filter
- app/api/bookkeeping/journal-entries/route.ts: when no status filter
is supplied, exclude `cancelled` entries by default
- supabase/migrations/20260428153500_journal_entries_with_related_exclude_statuses.sql
3. Swedish e-invoicing skill (reference docs only — no runtime code)
- .claude/skills/swedish-e-invoicing/
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(skatteverket): address PR review findings
- panel: handleFetchDraft read `result.data?.last` (typo) — switched to
`result.data?.locked` to match the field defined in
SkatteverketUtkastResponse and the v1.0.24 RAML. The "(låst)" suffix on
the success message would silently never appear before this fix.
- api-client: getValidToken had no concurrency guard, so two parallel
SKV requests from the same user could both call /token with the same
refresh_token. SKV rotates the refresh_token on first use, so the
second call would 401 with REFRESH_EXHAUSTED-adjacent failures. With
the new 6-button UI on SkatteverketPanel, rapid clicks made this a
realistic trigger. Added an in-process Promise map keyed on userId
that coalesces concurrent refresh attempts; cross-process races are
mitigated by re-reading tokens inside the critical section before
calling refreshAccessToken (if another process refreshed already, we
use the newer token instead of burning the old refresh_token).
- migration 20260428120000: dedup query used `created_at < max(...)`,
which failed to remove duplicates inserted in the same second. The
subsequent ALTER TABLE … ADD CONSTRAINT would then abort. Switched
to ctid (Postgres physical row identifier) to break timestamp ties.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(skatteverket): throw on token-store SELECT error before destructive DELETE
The company_id pre-read in storeTokens used destructuring that discarded
the error field. If the service-role SELECT failed for any reason (network
blip, overloaded DB, transient permissions issue), `existing` became null,
`resolvedCompanyId` stayed undefined, and execution fell through to the
DELETE. The old row got deleted successfully, then the INSERT omitted
company_id and failed with the NOT NULL constraint violation — leaving
the user with no token row at all and forcing a fresh BankID handshake.
Now we capture the SELECT error and throw before the DELETE runs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
||
|
|
9a7f067fdd |
Revert "fix(tic): call /enrichment before /collect to avoid session-consume bug (#367)" (#368)
This reverts commit
|
||
|
|
34955b0418 |
fix(tic): call /enrichment before /collect to avoid session-consume bug (#367)
* fix(tic): call /enrichment before /collect to avoid session-consume bug TIC support confirmed (2026-04-24) that reading session status via /poll or /collect after the BankID session reaches `complete` marks the session as "consumed" server-side, after which /api/v1/enrichment refuses the sessionId with `error: "Session not completed"` (returned as HTTP 200, not 400). Today's flow inside POST /api/extensions/ext/tic/bankid/complete called collectBankIdResult and then requestEnrichment — exactly the order that triggers the bug. Login still succeeded (collect itself returns OK), but enrichment failed 100% of the time, leaving extension_data.bankid_enrichment empty and /select-company's CompanyRoles picker blank. Reorder so /enrichment runs first, then /collect. Per TIC, the consume flag only blocks subsequent /enrichment calls — collect after enrichment is harmless. Refactor the existing fetchAndStoreEnrichment helper into fetchEnrichmentSafely (pure fetch, returns EnrichmentData | null) and storeEnrichment (pure DB upsert), so the data can be captured before the collect/user-creation logic and persisted afterward. Enrichment stays non-blocking: any /enrichment failure still logs a warning and lets login proceed, exactly as today. Adds a regression test asserting requestEnrichment is invoked before collectBankIdResult (via mock.invocationCallOrder) plus tests for both failure modes — enrichment throwing and TIC returning the production "Session not completed" body shape. Ref: TIC session 90f7da27-84b2-4f09-afc4-47239eef57c1. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(tic): drop redundant void _omit — _ prefix already silences lint Per PR review: @typescript-eslint/no-unused-vars ignores `_`-prefixed identifiers by default, so the void expression is dead weight. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
fd1db89603 |
Fix/invoice numbers (#365)
* feat: make invoice_number nullable and assign on send - Updated the invoices table to allow invoice_number to be nullable. - Modified the logic to assign invoice numbers only when the invoice status transitions to 'sent'. - Refactored related code to handle nullable invoice numbers, including UI components and API routes. - Added tests to ensure correct behavior when handling invoices with null invoice numbers. - Introduced a utility function to display invoice numbers, defaulting to '(Utkast)' for drafts. * fix: update fiscal period handling to return names of open periods in error messages * fix: enhance period creation logic to account for company-wide bookkeeping lock-through * fix: remove unnecessary customer_type field from customer insertion query * fix: scope invoice number count query to specific companies to avoid test interference * feat: Implement atomic invoice number generation and ensure compliance with invoice numbering rules - Introduced `ensureInvoiceNumber` function to assign invoice numbers atomically, handling concurrency and ensuring compliance with document types. - Updated invoice-related components to utilize the new `invoiceNumberDisplay` utility for consistent invoice number formatting. - Added checks to ensure that invoices in non-draft statuses have valid invoice numbers, preventing violations of legal requirements. - Created tests for the new invoice number generation logic, ensuring correct behavior under various scenarios, including concurrent requests. - Added a draft banner to PDF templates for invoices without assigned numbers, clarifying their status to users. - Updated database migrations to support the new atomic invoice number generation logic and enforce constraints on invoice statuses. |
||
|
|
a0485ca1c1 |
fix: BankID signup takeover + supplier invoice dedup + event_log RLS visibility (#358)
* fix(tic): reject BankID signup when email is already registered (CWE-287) Signup previously linked BankID to any pre-existing profile matching the submitted email. Because BankID proves identity but not email ownership, an attacker who knew a victim's email could bind their own BankID to the victim's account and then log in via BankID (which skips TOTP MFA). Now signup returns 409 account_exists; the register page shows a Swedish error toast and redirects to /login so the user can authenticate with their password first and link BankID from settings (via the authenticated /bankid/link route that already exists). Covered by new bankid-complete.test.ts with an explicit regression test asserting no side-effects occur on the account_exists path. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(bookkeeping): prevent duplicate registration entries for supplier invoices The invoice-inbox convert flow creates the registration journal entry inline and then emits supplier_invoice.confirmed. The core handler was also creating one, producing a second voucher on 2440/2641/expense and overwriting registration_journal_entry_id on the row. Two guards in the core handler: - Payload guard: skip if supplierInvoice.registration_journal_entry_id is already set (fast path for callers that include it in the payload). - DB re-fetch guard: re-read the row and skip if it has been linked since the payload was built (handles stale-payload callers). The inbox extension now stamps registration_journal_entry_id onto the in-memory invoice before emitting so the fast path trips. Also fixes a latent bug where the handler filtered company_settings by userId instead of companyId, which would have selected the wrong row (or none) on multi-company accounts. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(events): write company_id on event_log rows and refuse unscoped events event_log has a company-scoped SELECT RLS policy, so rows written with NULL company_id are invisible to every user — effectively silently dropped from the automation feed. The handler now reads companyId from the payload (all persisted event types mandate it in TS) and includes it in both single and batch inserts. If a caller ever bypasses the type system and emits without companyId, the handler logs an error and skips persistence rather than writing a poisoned row. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix: address PR review nits (event-log meta strip, redirect timer) - event-log-handler: rename stripUserId → stripMetaFields and also drop companyId from the stored data JSONB so it isn't duplicated alongside its dedicated column. - register page: drop the 1500ms setTimeout before router.push('/login') on the account_exists branch — the timer had no cleanup and fired a stale-closure push if the component unmounted first. The toast survives the route change via the root layout's Toaster. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
1af977950b |
Ai/full autonomous flow (#359)
* Refactor bookkeeping error handling and introduce new error classes - Introduced new error classes for better error categorization: - JournalEntryNotBalancedError - FiscalPeriodNotFoundError - EntryDateOutsideFiscalPeriodError - JournalEntryNotFoundError - CannotReverseNonPostedError - CannotCorrectNonPostedError - EntryAlreadyReversedError - CurrencyRevaluationAlreadyExistsError - InvalidMappingResultError - BookkeepingDatabaseError - Updated existing functions in engine.ts and transaction-entries.ts to throw specific errors instead of generic ones. - Enhanced error response handling in get-error-message.ts to provide localized messages for new error types. - Added unit tests for new error classes and error handling functions to ensure correctness and coverage. * feat(ai): implement AI proposal application and persistence - Add apply.ts to handle the application of AI proposals, including match and booking steps. - Introduce persist.ts for inserting and managing AI requests and proposals, ensuring unique constraints. - Create re-validate.ts for validating proposals before acceptance, checking for stale conditions. - Define database migrations for ai_requests and ai_proposals tables, including constraints and indexes. - Enhance journal_entries with AI provenance tracking, linking entries to AI proposals. - Update categorization_templates to distinguish AI-corrected templates. - Add company settings for toggling AI flow and managing backfill processes. - Extend processing_history to include AI-related events for better tracking. * feat: add uncategorized transactions API and UI for transaction selection - Implemented a new API endpoint for fetching uncategorized transactions with pagination and filtering options. - Created ChangeTransactionDialog component for selecting alternative transactions based on AI proposals. - Developed ReceiptDetailDialog to display detailed information about receipts, including upload functionality. - Added TransactionDetailDialog for viewing transaction details with links to the transaction list. - Introduced receipt quality assessment logic to evaluate extracted receipt data. - Implemented feature flagging for the AI bookkeeping agent to control availability in different environments. * feat: add manual receipt extraction dialog and integrate AWS Textract for expense analysis - Added ManualExtractDialog component for user input when AI fails to extract receipt data. - Implemented ReceiptsList component to manage and display uploaded receipts, including upload and rescan functionalities. - Introduced Textract integration for analyzing expenses, extracting fields like total, vendor, and date. - Updated package.json to include @aws-sdk/client-textract dependency. * fix(ai): handle livsmedel VAT transition (12% → 6%) in booking prompt and re-validate guard Add date-aware guidance to BOOKING_SYSTEM_PROMPT for the temporary livsmedel VAT cut (Prop. 2025/26:55, 2026-04-01 to 2027-12-31), with restaurang/servering carve-out at 12%. Add a re-validate safety net that rejects clearly-stale rate labels for grocery-chain merchants relative to the entry date. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
02f94ef631 |
Fix/critical issues (#351)
* fix: add 15s timeout to accounting provider HTTP clients Node's built-in fetch has no default timeout, so a stalled provider could hold a serverless worker open for many minutes — worse with withRetry (6x on Fortnox, 3x on others) and getPaginated stacking across pages. Wrap each fetch() in the Fortnox, Visma, Bokio, Briox, and Björn Lundén clients with signal: AbortSignal.timeout(15_000), and treat TimeoutError/AbortError as retryable so a single stalled attempt retries cleanly instead of hanging the request. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix: add timeouts to OAuth token endpoints Wrap every OAuth2 token exchange, refresh, and revoke POST in an AbortController via a new fetchWithTimeout helper. Without this, a hung provider endpoint holds the request thread indefinitely — worst case being Skatteverket, where refreshAccessToken sits on the hot path of every bookkeeping action and exchangeCodeForTokens races the 5-minute BankID auth-code TTL. On timeout, the Skatteverket OAuth callback now redirects to /reports?tab=vat-declaration with a Swedish retry message instead of leaving the user stranded on the callback URL. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix: close RLS escalation on membership and settings tables Any authenticated user who was a member (including viewer) could issue a direct PostgREST PATCH against company_members and promote themselves to owner, bypassing the app-layer requireWritePermission guard entirely. Reproduced on prod, then verified the fix on staging. Tighten INSERT/UPDATE/DELETE policies on company_members, team_members, api_keys, company_invitations, team_invitations, companies, teams, and company_settings to require the caller to hold role IN ('owner','admin') in the target company/team. Role check is wrapped in SECURITY DEFINER helpers (user_is_company_admin, user_is_team_admin, user_role_in_company) to avoid RLS recursion when a policy on company_members references company_members in its subquery. Add a BEFORE UPDATE trigger on company_members that rejects any role change unless the caller already holds role='owner', so admins cannot mint further owners even though they can otherwise write. Legitimate write paths are unaffected: company creation goes through the create_company_with_owner SECURITY DEFINER RPC, invite acceptance uses the service role, and team->company membership syncs via SECURITY DEFINER triggers. All bypass RLS. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(migrations): resolve duplicate schema_migrations version 20260421160000 Two migration files shared timestamp 20260421160000 on main (booking_template_usage.sql and opening_balances_rpc.sql), causing supabase_migrations.schema_migrations PK collisions on any fresh CI run: duplicate key value violates unique constraint "schema_migrations_pkey" Key (version)=(20260421160000) already exists. Bump opening_balances_rpc.sql to 20260421160500. booking_template_usage keeps 20260421160000 because its table already exists on prod; the renamed file has an idempotent CREATE OR REPLACE FUNCTION body and has not yet been deployed to prod, so moving its version is free. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(migrations): make booking_template_usage migration idempotent The table already exists on prod (applied out-of-band) but prod's schema_migrations does not track version 20260421160000, so the next PR-driven deploy would re-run this migration and fail on `CREATE TABLE public.booking_template_usage` with a duplicate-relation error. Add IF NOT EXISTS to CREATE TABLE and CREATE INDEX, and DROP POLICY IF EXISTS before each CREATE POLICY. No functional change on fresh databases; prod just silently no-ops the table/index creates and re-declares policies without dropping-then-missing them. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix: implement isTimeoutError utility and enforce role restrictions on company_members insert * fix: implement fallback for user_id in commit_journal_entry function when auth.uid() is NULL --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
|
7cbd791c5b |
fix: request only enabled TIC enrichment types (unblocks the picker) (#349)
* chore: translate TIC enrichment errors into actionable hints Adds self-diagnosing log output for known TIC failure shapes: - 'Session not completed' → tenant doesn't have enrichment enabled (the BankID consent-to-enrich dialog only fires when SPAR/CompanyRoles are configured on the tenant; without it, TIC rejects enrichment requests with this misleadingly-worded error) - 'not enabled' variants → explicit tenant disable - 'too old' → >30min between auth and enrichment request No behaviour change — purely diagnostic. Saves a future developer from re-discovering that "Session not completed" means "contact support@tic.io" rather than "retry the session." Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix: stop requesting disabled 'SPAR' enrichment type Root cause of the 'Session not completed' failures: TIC renamed 'SPAR' to 'Address' and our tenant has Address *disabled* (only CompanyRoles is on — verified via GET /api/v1/enrichment/types). Requesting an unknown or disabled type causes TIC to reject the whole enrichment with a misleading 'Session not completed' error rather than ignoring the unknown type or returning PartiallyCompleted. - Enrichment request now asks only for ['CompanyRoles'], which is what the /select-company picker actually needs. This is enough to stop the 'Session not completed' errors and make the picker work in prod. - If Address is enabled on TIC later, add it to the array (and wire up the address pre-fill paths in WelcomeOnboarding + createCompanyFromTicRole — both already look for a `.spar` field that TIC may have renamed). - Updated the diagnostic hint for 'Session not completed' to lead with the type-mismatch cause and point at the /enrichment/types curl for verification, since that's the first thing to check. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore: clean up stale SPAR references in tic extension (PR review) Greptile P2 nits — all cosmetic, no runtime impact: - fetchAndStoreEnrichment JSDoc updated from "SPAR + CompanyRoles" to just CompanyRoles, with a note about where to add Address back if/when it's enabled on the tenant. - hasSpar log field removed from the 'enrichment data shape' snapshot. enrichmentData.spar was permanently undefined after the SPAR → nothing swap; logging a constantly-false field would have misled a future developer chasing missing address data. - Inline comment on the signup-path call to fetchAndStoreEnrichment now matches the login-path comment (CompanyRoles, not SPAR + CompanyRoles). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |