Files
accounted/extensions/general/mcp-server
Jakob WennbergandClaude Opus 4.8 f7cd1b86e7 fix(import): preserve customized SIE #KONTO account names (#669)
* feat(import): add syncMappedAccounts helper for account create + rename

Single home for the create-missing-accounts logic that exists in three
near-identical copies (executeSIEImport, the SIE execute route, and the
arcim-migration extension), plus a new rename pass that carries customized
SIE #KONTO names into accounts that already exist (e.g. K1-seeded defaults).

The file's name applies only to identity mappings (source === target);
remapped targets keep their BAS/current name. With updateAccountNames=false
the behavior matches the legacy code exactly. Not wired up yet.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(import): preserve SIE #KONTO account names; add updateAccountNames option

Customer report: account names customized in Fortnox did not follow into
Accounted via SIE import. The import always used BAS default names for
accounts in the BAS reference and never touched accounts that already
existed (the K1-seeded chart), so the file's names were silently dropped.

executeSIEImport now routes account creation through syncMappedAccounts,
which prefers the file's #KONTO name for identity-mapped accounts and
renames existing accounts whose name differs (surfaced as a warning).
New option updateAccountNames (default true) restores the old behavior
when disabled. The duplicated pre-create blocks in the execute route and
the arcim-migration extension are removed — executeSIEImport owns account
sync on every path now, including the Fortnox re-sync (idempotent renames).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(mcp): expose update_account_names on gnubok_import_sie

Optional boolean on the tool schema, staged into the pending operation and
threaded through commitImportSie to executeSIEImport. Defaults to true at
both stage and commit time — the commit-side default also covers operations
staged before the param existed (Boolean(undefined) would have silently
flipped it off).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(api): v1 SIE import generated no account mappings

The route passed [] as mappings to executeSIEImport, which the
mapping-coverage guard (added in #613) rejects for any real file — and
before that guard, every voucher was silently skipped as unmapped. The
route has never produced a working import for files with vouchers.

Generate mappings server-side from the file's #KONTO records plus stored
per-company overrides (same as the dashboard execute route), reject
unmappable files with a clean 400 before the operation row is created,
and expose options.updateAccountNames (default true).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(import): "Använd kontonamn från filen" toggle in import review step

New switch (default on) controlling whether the SIE file's #KONTO names
are carried into the chart of accounts. Helper text shows how many
identity-mapped accounts carry names that differ from the BAS defaults.
The page already serializes the whole options object to the execute
route, so no further wiring is needed.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(import): address PR #669 review — parallel renames, rename audit trail

- Rename pass now runs UPDATEs concurrently in bounded batches of 25
  (greptile P2): a re-sync with many custom names no longer serializes
  N round trips, and a pathological full-chart rename cannot stampede
  the API. Per-rename failures stay non-fatal via Promise.allSettled.
- Persist the per-account rename detail (number, from, to) into
  sie_imports.migration_documentation as accountRenames — the
  behandlingshistorik record per BFNAR 2013:2 (swedish-compliance
  review); the result warnings only carry the count.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-04 20:35:49 +02:00
..
2026-05-04 11:12:29 +02:00

Accounted MCP server

JSON-RPC 2.0 server exposing the Accounted bookkeeping engine to MCP clients (Claude Desktop, Claude Code, etc.). Endpoint: /api/extensions/ext/mcp-server/mcp. OAuth and stdio bridge live alongside the API surface — see app/api/mcp-oauth/ and packages/gnubok-mcp/.

Tool authoring contract

Enforced by tests in __tests__/ — these are not style preferences, they're guard rails.

  1. additionalProperties: false on every inputSchema. Guarded by strict-schemas.test.ts. Forces clear rejections on hallucinated fields instead of silent ignores.
  2. Descriptions ≤ 280 chars. Guarded by output-schema.test.ts. No Args: / Returns: / Examples: prose — those belong in JSON Schema. Use agent-native hints ("Use to…", "Call X first", "HIGH risk").
  3. Staged-operation envelope for write tools — outputSchema: STAGED_OPERATION_SCHEMA (server.ts). Fields: staged, risk_level, actor, message, preview, period_status?, next?. The staged: true boolean is the explicit completion signal; agents must not infer completion from prose. Do NOT introduce a parallel { success, shouldContinue, output } envelope.
  4. period_status threading — any tool that ties to a fiscal-period-bound date (categorize, mark paid, create voucher, correct/reverse entry, approve supplier invoice) passes dateForPeriodCheck to stagePendingOperation. Response then includes period_status: { period_id, status: open|locked|closed, lock_date } so widgets and agents disable writes without round-trips.
  5. Scope mapping — every new tool needs an entry in lib/auth/api-keys.ts TOOL_SCOPE_MAP. Missing entries default to deny.
  6. Tests for new write tools — add staging-gate coverage to __tests__/voucher-tools.test.ts (or a sibling) plus executor coverage to lib/pending-operations/__tests__/voucher-executors.test.ts if the tool stages a new operation_type.

Determinism / cache stability

Tool definitions (name, description, inputSchema, outputSchema, annotations) are declared as static object literals at module load — no timestamps, no UUIDs, no Date/Math.random in the definition layer. This makes the tools/list JSON payload byte-stable across requests, which lets agent-side prompt caches stay warm. Do not introduce per-request non-determinism into the definitions block. Anything time-bound or random belongs inside execute().

For internal Anthropic API usage (today only extensions/general/invoice-inbox/lib/extract-invoice-fields.ts): annotate stable prefixes with cache_control: { type: 'ephemeral' } and log usage.cache_read_input_tokens for hit-ratio observability. The 1h TTL from the agent-native API plan (item 10) requires the direct Anthropic API; Accounted's Bedrock path defaults to a shorter TTL.

Payload-size watchdog

payload-size.bench.test.ts enforces a tools/list JSON payload ceiling (currently 25,000 tokens). If the test fires, the right answer is rarely "raise the ceiling" — instead, trim descriptions or leverage gnubok_search_tools (already deployed; tool definitions can defer to it for discovery rather than enumerating in tools/list).

Where things live

  • server.ts — the tools array + JSON-RPC dispatcher
  • tool-result.ts — withNext(), toToolError() response helpers
  • resources/ — read-only Accounted:// URIs (active company, period, recent activity, capabilities, attention items, voucher gaps, chart of accounts, VAT treatments)
  • widgets/ — inline HTML widgets (receipt-matcher, vat-review)
  • prompts/ — slash-command-style prompts
  • skills/ — domain-knowledge skill bodies served via gnubok_load_skill
  • __tests__/ — strictness guards + per-tool coverage