46c0b72ab0dc555d40cde0b9322224712c055ee9
6
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 MattssonandClaude Fable 5
|
f24b26a139 |
fix: similar-sweep currency remediation, security hardening and v1 API fixes (#1215)
* fix(security): gate replace_sie_import behind owner/admin membership The RPC was SECURITY DEFINER with EXECUTE granted to PUBLIC and anon, no company_members lookup, no auth.uid() reference and no unauthorized raise, while setting gnubok.allow_delete to disarm the BFL immutability and retention triggers. Any caller holding a company_id and an import id could hard delete another tenant's verifikationer. Confirmed live in production. Applies the same fail closed owner/admin guard that undo_sie_import already carries (migration 20260624120000), resolving the actor from COALESCE(p_user_id, auth.uid()) so it denies when the role is NULL, then revokes EXECUTE from PUBLIC and anon. search_path and the raised statement_timeout are restated, since CREATE OR REPLACE drops settings that are not repeated. userId is a required parameter on replaceSIEImport: the service client has a NULL auth.uid(), so a caller without an explicit actor now fails to compile rather than hitting the closed gate at runtime. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(security): validate arcim OAuth callback state server side The callback route is skipAuth and decoded the state parameter as plain base64url JSON, trusting consentId and provider from it. A one time code was minted at flow start and never read. An unauthenticated attacker who learned a consent id could run an OAuth flow on their own provider account and post the callback with a forged state, landing their tokens on another tenant's consent, so the victim's next migration imported the attacker's ledger. State is now an opaque randomBytes(32) pointer to a provider_otc row, consumed by a single atomic UPDATE guarded on used_at IS NULL and expires_at, so a replay loses the row lock race and updates nothing. provider is read from provider_consents rather than trusted from the client. provider_otc already existed for exactly this purpose and was never wired up. Also scopes getConsent to an owning company, closing a cross tenant status oracle where the preview and migrate paths echoed a consent's status before the scoped check ran. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(security): scope documents storage to company_id (phase A) The documents bucket policies matched on auth.uid(), and upload keys were documents/{userId}/..., so company membership was never consulted. Removing a member revoked nothing: their session still authenticated and they kept direct Storage read access to every receipt, supplier invoice and bank statement they had uploaded. The same bug was fixed for sie-files in 20260416120000; this bucket was left behind. Phase A is additive. Company scoped policies are added alongside the uploader scoped ones, uploads move to documents/{companyId}/{userId}/..., and reads accept either layout so nothing breaks mid migration. Phase C, which drops the old policies, is gated on the backfill reporting zero remaining legacy prefix objects. The policy compares the company segment as text rather than casting to uuid the way sie-files does: this bucket holds keys whose second segment is not a uuid (MCP audit packages), and Postgres does not guarantee the bucket prefix qual runs before the cast, so a planner reordering would raise 22P02 and fail the whole query instead of filtering the row out. deleteDocument now removes both candidate keys. Removing only the stored pointer would leave a readable orphan copy of a document the user asked to erase. The backfill script is included but has never been run. It defaults to dry run, refuses .env.local by name, and verifies each copy is readable and SHA-256 identical before repointing the row. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(security): enforce events:read scope and membership on /api/events This was the only one of the three validateApiKey call sites with no downstream guard: v1 and the MCP server both check scope and re-verify company membership, this route did neither. An events:read scope existed and was documented as gating the endpoint but was never called, so a legacy key falling back to DEFAULT_SCOPES read the full log. The bound company id went straight from the api_keys row into a service role query, so a key whose user had been removed from the company kept reading. Adds the scope check before any database access, re-verifies company_members with archived_at IS NULL, honours test mode by stamping X-Gnubok-Mode instead of ignoring it, applies minimisePayload so the pull surface can never return a wider payload than the push surface, and replaces the three flat error strings with the canonical envelope. Test key reads are served rather than blocked: TEST_KEY_WRITE_BLOCKED is gated on mutations in with-api-v1, so a read gets the same treatment as every other v1 read endpoint. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * perf(bookkeeping): sweep remaining journal_entries!inner embeds A previous refactor removed this pattern from lib/reports and introduced fetchEntryLines, but the class was never swept. Seventeen sites remained and had become the top application consumer of production database time: measured across the resulting query shapes, 32,694 calls and 25,848 seconds of execution, mean 790ms, with shapes averaging 2.6s and 3.0s and maxing at 7,962ms against the 8s statement_timeout, which surfaced to users as 500s on the booking path. PostgREST compiles an embed with filters on the embedded side into a correlated INNER JOIN LATERAL with a parameterized LIMIT, which stops Postgres reordering the join, so each query walked the whole journal_entry_lines table across all tenants. Driving from the entries side instead turns that into two indexed round trips. Converted sites keep their existing shape: the helper reattaches the parent entry under the same key the embed produced. Several conversions also remove a latent silent truncation where an unpaginated query was capped at PostgREST's 1000 row ceiling. Two deliberate exceptions. The free text ilike legs of the MCP display query stay on the embed, because each is capped at legLimit and that cap drives the truncation contract the tool reports, while the helper is unbounded. The accounts route moves to the existing get_account_usage_counts RPC instead, since its embed was a head count and the helper returns rows. commitEntry's write path is untouched: the change there is confined to the read query of the pre-commit dimension rule check. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(api): anchor v1 list cursors on created_at Page two returned page one, forever, while still advertising a fresh next_cursor. The three routes sorted by and encoded a Postgres date column, which serializes as YYYY-MM-DD, but decodeDefaultCursor validates the cursor timestamp as full ISO-8601 and returned null, so the keyset filter was never applied and has_more never went false. An integrator syncing verifikat looped on the newest rows indefinitely. The transactions route already solved this and its comment names the trap; the fix was never ported. All three now order and encode on created_at with an id tie break, matching the transactions keyset predicate exactly. ISO_TIMESTAMP is deliberately left alone: relaxing it would silently change sort semantics on the route that currently works. Default ordering therefore moves from business date to insert order. Every business date is still on the row, and the invoices list gains date_from and date_to filters so a date range is still reachable; the other two already had them. The tests use an in-memory PostgREST that actually evaluates the filters, because the repo's pass-through mock cannot catch this class of bug: the bug is that the filter is never sent. They walk to exhaustion with a hard iteration cap, so an unterminated walk fails instead of hanging. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(api): separate dry run from commit in the idempotency hash The request hash was built from url.pathname, which excludes the query string, so a dry run and its commit hashed identically. Following the flow documented in dry-run.ts, re-issuing the request with the same Idempotency-Key returned the cached preview with Idempotent-Replayed set and wrote nothing, while reporting 200. An agent or integrator saw success for a write that never happened. dry_run is folded into the hash only when true, not as an unconditional boolean. Including it as false would change the hash of every ordinary write, and with a 24h idempotency TTL any key in flight across the deploy would fail the request_hash comparison and 409 on a legitimate retry. Both hash call sites now go through one shared helper so they cannot drift into a permanent cache miss, and dry run responses are no longer stored at all. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * ci: install the Bedrock SDK out of tree in the compliance review The Swedish accounting compliance gate had failed ten consecutive runs and so was posting nothing. With --no-package-lock npm discarded the lockfile and re-resolved the whole tree from package.json, floating @hookform/resolvers to 5.4.3, whose valibot ^1 peer conflicts with the pinned valibot 0.39.0. Installing into the parent of the checkout resolves only that one package, so an unrelated peer conflict can never take the gate down again. Node still finds it because ESM bare specifiers walk up parent node_modules; NODE_PATH would not have worked, as it is CommonJS only. --legacy-peer-deps was rejected because it masks future genuine peer conflicts and still reifies the full tree. The same step's SDK version is aligned from 0.31.0 back to the 0.29.1 that package.json and check:guards enforce after the streaming outage. That drift went unnoticed because the pin guard only inspects package.json and the lockfile, never workflow files. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * build(docker): generate crontabs from vercel.json vercel.json defines 16 cron jobs; both Docker crontabs carried 9, and were byte identical to each other. Self hosted deployments therefore never sent recurring invoices, never dispatched webhooks and never cleaned up idempotency keys. tax-deadlines also ran once a year on 2 January instead of daily, and documents/verify weekly instead of daily. Extension crons are included rather than excluded. The Dockerfile copies the whole tree before building, so every extension cron route is compiled into the image regardless of the enabled preset, and each returns 200 when its extension is unconfigured, so curl -sf logs no failure. Two such entries were already present in the crontab for extensions absent from the preset, which settles the intent. documents/verify is treated as drift rather than a self hosted concession: the weekly cadence was present in the hosted crontab too, and the run is capped at 200 documents walking a nulls-first queue, so weekly drains the integrity queue seven times slower on a check that exists for BFL retention. webhooks/dispatch keeps its per minute cadence, adding 1,440 requests a day on self hosted. A gentler tick would silently stretch the first retry, since the retry ladder opens at 60 seconds. SCHEDULE_OVERRIDES is the one line place to change that. A parity test asserts the path sets match minus a documented exclusion list, and ratchets three cron routes that are currently scheduled nowhere so they are named rather than silently rotting. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * chore(observability): add a provider agnostic error sink There is no error tracking in this codebase: logs go to console and Vercel retention and nowhere else, nothing alerts on the 16 cron jobs, and seven code comments across lib, app, components and extensions asserted that Sentry captures errors when Sentry is not a dependency. The two most recent bug fixes on this repo were both discovered by customer email. This adds the sink, not a vendor. No dependency is taken: the interface has a no-op default and a registration point, so behaviour is unchanged until an adapter is registered. Releases are tagged from the build id already inlined by next.config.ts. Redaction moved out of lib/logger.ts into a leaf module that both the logger and the sink import, so there is one denylist and no path from application data to a third party can skip the personnummer regex, including direct sink calls that bypass the logger. That matters here because these logs carry personnummer and financial data. verifyCronSecret now reports its own 401s, which covers all 16 jobs without touching a route file and catches the case where CRON_SECRET is rotated without updating the scheduler and every job silently 401s forever. The threshold is one failure rather than the backup alert's three: suppressing the first occurrence is precisely how an outage stays invisible. The seven misleading comments are corrected to describe what the code actually does, including the two cases that still are not covered: the client side one, since the sink is server side, and a warn level call that is not forwarded. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix: remediate the 2026-07-26 similar-sweep findings across all surfaces Resolves the ~150-finding sweep (dev_docs/similar-sweep-2026-07-26.md) with one agent per finding; every behavioural fix carries a regression test proven to fail at HEAD. Full status, corrections to the sweep, refusals and open decisions in dev_docs/similar-sweep-2026-07-26-remediation-status.md. Structural roots closed: - resolveSekAmountOrNull(): honest SEK resolution refuses instead of booking 1:1; four duplicated toSek closures now refuse via INVOICE_FX_RATE_MISSING - ledger-line-amount.ts: journal_entry_lines.currency labels the document, not the amount; SQL pre-filter decoy proven and fixed - sparse-patch.ts: .partial() does not strip .default() in Zod 4.4.3; the exploitable salary payslip-line PATCH and KPI preferences sinks fixed - tests/schema: migration-replay phantom-column guard (13k+ refs, closed CHECK sets, onConflict targets); found 28 real defects, all fixed, all four baselines now empty - three new ratchet guards: sek-labelled-amount, cross-extension-import, ungated-extension-route Highlights: lawful VAT-rate set on all seven invoice surfaces (ML 6 kap), RC input VAT mismatch wired on web + both MCP callers, missing-underlag resource delegates to the shared RPC predicate, push-notifications consent polarity fail-closed, deadlines undo honours requested state, silent-failure and read-side-fabrication classes fixed across settings/KPI/inbox/Stripe/ Arcim/kassaflodesanalys, error-envelope stringification fixed at 10+ sites with isSwedishUserMessage extended. Also includes the parallel session's MCP invoice tools (update_invoice, recurring schedules, invoice deliveries) which share files with the sweep work and are verified green together. 13 new migrations are NOT applied anywhere; they apply via branch merge. 20260726120000 backfills 1247 supplier-invoice rows. pg tests for new DDL are written but unrun (no local Postgres). Verified: 11088 tests / 881 files green, tsc 0 non-test errors, lint 0 errors, check:guards passing, MCP payload 57475/57500. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(migrations): rename replace_sie_import migration off main's 20260726090000 version origin/main shipped 20260726090000_agent_quota_rpc_caller_guard.sql; keeping our replace_sie_import migration on the same version would abort the Supabase apply with a schema_migrations_pkey duplicate at merge time. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(review): remediate pre-publish deep-review findings across all slices A 13-agent review of the full branch diff surfaced 1 critical, 5 high and ~45 further findings; this commit resolves them in one pass: - replace_sie_import / undo_sie_import: p_user_id honored only for service_role callers; any other caller is pinned to auth.uid() (impersonation gate bypass), authz raise errcode 42501 mapped to a Swedish 403 in the route, new caller-guard migration for undo - bulk_book_transactions refuses homogeneous non-SEK batches instead of writing foreign magnitudes into SEK ledger columns - credit-note cap trigger: company-match on credited_invoice_id, no cross-tenant figures in exception text - link_voucher RPCs resolve NULL invoice currency as SEK end to end - personal-number ciphertext CHECK split into NOT VALID + VALIDATE - same-currency foreign settlements clear 1510 at booking rate and book realized diff to 3960/7960; rate-less foreign write paths refuse - receivables revaluation covers partially_paid and outstanding amounts - period lock guard paginates candidates past the PostgREST 1000 cap - documents: service-client storage removals after authz, dual-layout reads in integrity cron and archive export, backfill delete-source sweep actually deletes with hash verification and shared-key grouping - invoice matching normalizes NULL/lowercase currencies (regression), duplicate candidates stop claiming amount matches they never ran - match-invoice aborts on any booking failure (no paid-without-verifikat) - refresh-exchange-rate reverts on concurrent booking (TOCTOU window) - KPI preferences upsert arbiter aligned to the company-scoped constraint - personnummer_last4 stripped from all salary responses incl. MCP tools - worked-hours batch restores destroyed rows on conflict and error paths - MCP: shared duplicate-claim builder (no more 'null kr'), short-circuit on tag_journal_lines overflow, auto_send schedules stage as high risk - observability sink redacts emails/IBANs/API keys and keeps redacted stacks in prod; assorted small guards (safe-return-to /@, dry_run=True, cursor helper off-by-one, OAuth state TTL 10 min, arcim saveMappings call removed) Full dispositions, deferred items and hand-verified accounting numbers are documented in the PR body and DECISIONS.md. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * feat(personnummer): implement masking and encryption for personal numbers with tests * fix(review): address CI and compliance-bot findings for PR #1215 pg-real: the CI image's auth shim reads the legacy request.jwt.claim.role GUC, so both service-role simulations (runAsServiceRole and the invoice-delivery test's local helper) never satisfied auth.role() = 'service_role' and every legitimate p_user_id path failed closed; the shared helper now sets both GUC shapes plus SET LOCAL ROLE with a fail-loud sanity check, and the delivery test reuses it. The link-voucher migration had recreated both RPCs from pre-rewrite file text, reintroducing the NULL-unsafe membership pattern the null-safe-tenant-guards ratchet bans; both guards now use public.caller_is_company_member() with all currency changes preserved. Compliance bots: the customers export now emits the standard masked form instead of raw AES-256-GCM ciphertext in the Org-/personnummer column, and maskCustomerRow returns a non-round-trippable placeholder on decrypt failure instead of 500ing the list. MCP parity: gnubok_lock_period's staging pre-check now runs the exact countUnbookedInPeriod the commit path enforces (exported from period-service; local mirror deleted), and gnubok_agi_status resolves AGI state run-scoped so a correction run no longer renders as already filed. Declined with evidence: PR-Agent's opening-balances null-zeroing concern (all mergeable columns are NOT NULL with defaults per 20260713101000). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(review): address codex review findings on PR #1215 - restore 20260726140000 to its preview-recorded content and restate the NULL-safe tenant guard under 20260727130000: a recorded migration version never re-runs, so the in-place edit could not reach the preview branch - replace toFixed() with sv-SE two-decimal formatting in the ROT/RUT cap warning texts and update the pinned test expectations - drop the em dash in the fiscal-periods route comment - strip trailing whitespace in import-existing.test.ts Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * test(reports): raise timeout on real PDF render tests renderToBuffer does real @react-pdf layout work and exceeds the 5s default when the full suite saturates the CPU; tests pass in isolation. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
Jakob WennbergandClaude Fable 5
|
9d91ee0168 |
fix(vat): keep the momsrapport intact after a manual nollställning (#990)
* fix(vat): keep the momsrapport intact after a manual nollställning (#984) The momsrapport already excludes settlement verifikat tagged with source_type 'vat_settlement' (#983), but settlements booked any other way still zeroed every ruta the moment they were posted: manual momsomföringar booked before the tagged flow existed (the report in issue #984), SIE-imported settlements, and storno reversals of a settlement, which inverted the sign instead and silently doubled the rutor after an annullera. Exclude settlement entries by SHAPE as well: an entry with at least one line on a declaration account (ACCOUNT_RUTA) and at least one on a settlement net account (2650/1650) is bookkeeping about the declaration, not VAT-bearing activity, in both the web projection (fetchVatAccountTotals) and the MCP twin (computeVatReport). Opening-balance entries are exempt: carried-in 26xx balances are unsettled VAT that belongs in the next declaration. Shape-detected POSTED settlements now also gate the "Skapa verifikat" button through existing_entries, since the proposal re-clears the full period and booking it on top of a manual settlement would corrupt the 26xx balances. Stornos never gate, so annullera still re-enables booking. Fixes #984 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(mcp): stable id order for the paginated VAT-report line fetch (CodeRabbit) fetchAllRows pages with .range(); without a unique .order() rows can shift across page boundaries once a period exceeds 1000 lines, skipping or double-counting journal lines in the rutor. Same discipline as the web projection (fetch-all.ts). Pre-existing, but the query was already being touched for #984. Also documents the shape-rule triage from the compliance-bot review in DECISIONS.md: compound business-VAT-plus-2650 verifikat stay a known accepted residual (a direction guard would break the storno exclusion), and the opening-balance concern is false for app flows (SIE import and set_opening_balances both tag source_type 'opening_balance'). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com> |
||
|
Jakob WennbergandClaude Fable 5
|
2774e01258 |
feat(vat): book the momsrapport as an editable settlement verifikat (#980) (#983)
* feat(vat): book the momsrapport as an editable settlement verifikat (#980) Adds a "Bokfor momsrapporten" card under the VAT declaration that builds an editable verifikat proposal from the report and books it through the ordinary journal entry form: - lib/reports/vat-settlement.ts: proposal builder. Clears each 26xx account at exact ore, books the net on 2650 (att betala) or 1650 (att aterfa) at the filed whole-krona amount (buildFiledAmounts, oretal faller bort per SFL 22 kap 1 par), balances the gap on 3740. Surfaces existing vat_settlement entries in the period so the UI can warn before a double booking. - GET /api/reports/vat-declaration/settlement-proposal: same period params as the sibling report routes. - VatBookingCard (reports view): fetches the proposal, warns when the period already has a posted settlement or draft, and opens the JournalEntryForm (bare, prefilled, source_type vat_settlement) in a dialog so every line is editable before committing. Booking uses the existing engine path: balance validation, period locks, voucher series per source type. - vat_settlement entries are excluded from the declaration projection (calculateVatDeclaration via new shared fetchVatAccountTotals, and the MCP computeVatReport for parity): a pure-projection report would otherwise read zero, and a later Skatteverket submission would file zeros, the moment the settlement is booked. No migration needed: the vat_settlement source type shipped in 20260708100000. Closes #980 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(vat): block re-booking a settled period, fail loud on lookup errors (CodeRabbit) The proposal is not delta-aware (it re-clears the FULL period), so a second booking while a posted settlement exists would corrupt the 26xx balances: disable "Skapa verifikat" until that verifikat is annulled (storno restores the balances). And since the existing-settlement lookup now gates that button, a swallowed query error would silently re-enable it: throw instead. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com> |
||
|
Jakob WennbergandClaude Sonnet 5
|
ec27228a8e |
style: remove em/en dashes repo-wide, add CLAUDE.md rule against them (#890)
Em dashes (—) and en dashes (–) had spread across comments, docs, tests, and a few UI strings, reading as AI-generated boilerplate rather than house style. Replaced each with punctuation matching its context: colon for explanatory clauses, comma for asides, plain hyphen for numeric/legal ranges (e.g. "21-23§"), "to"/"till" for date ranges, parentheses for paired-dash asides. messages/en.json and messages/sv.json were fixed by hand together to keep sv/en in sync. Left untouched where the dash is the functional subject rather than decorative punctuation: date-range-parser.ts's separator regex, charset-repair.ts's CP1252 byte-mapping table (and its test), the SIE encoding mojibake docs, generic-csv.ts's minus-sign normalizer, the agent system-prompt files that already instruct against em dashes, and a golden iXBRL test fixture compared byte-for-byte. Also fixes two bugs surfaced along the way: an off-by-one in ApiKeysPanel's scope-label split (a leftover from an earlier partial pass), and a charset-repair test that had lost the literal en-dash it exists to verify. Regenerated the agent atom seed migration (skills:generate) since 27 SKILL.md files changed. Added a CLAUDE.md rule against em/en dashes, with an explicit carve-out for the functional-dash cases above. Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com> |
||
|
|
37ee125b9b |
fix(mcp): paginate trial-balance and VAT aggregations (1000-row truncation) (#806)
* fix(mcp): paginate trial-balance and VAT aggregations (1000-row truncation) The gnubok_get_trial_balance tool and computeVatReport each ran an unbounded journal_entry_lines aggregation. PostgREST caps an unpaginated .select() at 1000 rows, so any period with >1000 entry lines silently truncated: wrong per-account sums and a false "not balanced" trial balance, and an under-reported momsdeklaration for yearly or busy quarterly VAT periods. - get_trial_balance now delegates to the canonical generateTrialBalance (lib/reports), which paginates via fetchAllRows and rolls opening balances forward, also fixing a latent bug where the tool ignored IB. - computeVatReport now paginates its line fetch via fetchAllRows. The library fixed this class of bug in #79; these two MCP paths kept their own copies that were never updated. Signed-off-by: Jonas Hagberg <jonas@lindan.se> * fix(mcp): non-null assert periodId in generateTrialBalance call Fixes the core-only TS build error (string | undefined not assignable to string). periodId is guaranteed defined by the !period guard above; mirrors the existing periodId! call later in the file. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * test(mcp): support .range() in computeVatReport mock for paginated query computeVatReport now fetches journal_entry_lines via fetchAllRows (.range), but the hand-rolled mock terminated at .lte(). Move the terminal to .range() so the 8 VAT-aggregation tests exercise the paginated path. Test-only. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Signed-off-by: Jonas Hagberg <jonas@lindan.se> Co-authored-by: Jakob Wennberg <jakob.wennberg@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Jakob WennbergandClaude Opus 4.7
|
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>
|