MattssonandClaude Fable 5
4d6dd68df4
* fix(mcp): surface BFL 5 kap 6 § underlag requirements in voucher and transaction tooling Addresses a user report that the MCP surface treats document anchoring as optional convenience while the law treats it as mandatory: - create_voucher: description and inbox_item_id reframed as the compliant path for received handlingar; staging without a document now carries an advisory compliance_warning in the preview and a WARNING in the message (never blocks: IB/migration vouchers legitimately lack a kvitto). - High-risk approval guidance now tells the agent to surface any preview compliance_warning alongside the 5 kap 5 § irreversibility. - Approval-flow copy is scope-aware: staging responses and list_pending_operations explain that keys without pending_operations:approve (SoD) approve at /pending instead of hunting for a tool their catalog hides. - categorize/create/bulk_book transactions: state that a transaction models a cash-account movement and point cashless events (privat utlagg) at create_voucher instead of silently fabricating a bank line. - link_document_to_voucher: signposts the pre-posting inbox path. - /pending VoucherPreview shows an attn line when document_attached=false. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(mcp): entity-neutral utlagg hint, coherent no-document envelope, real attn styling Addresses the skeptic refutations on da811a58f: - categorize_transaction: drop the 6540/2893 example; 2893 is AB-only (EF books egen insattning 2013/2018 per category-mapping.ts) and a static description cannot know the entity type. - create_voucher: when inbox_item_id is supplied but the item has no stored document (document_id NULL, ON DELETE SET NULL), the will-text no longer claims an OCR attach and the compliance warning gets an inbox variant instead of a dead-end 'restage with inbox_item_id'. Pinned by a new test. - /pending VoucherPreview: use the AttnLine component (the bare 'attn' class does not exist) and conditional wording so IB/internal vouchers are not falsely flagged under BFL 5 kap 6 \u00a7. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(mcp): exempt IB vouchers from the underlag warning; UI mirrors the staged warning Swedish accounting review round 2: the warning keyed off document_attached alone, so migrated IB entries (and the /pending card for any old no-document row) were falsely flagged under BFL 5 kap 6 \u00a7. Staging now skips the complianceNote when is_opening_balance=true, and VoucherPreview gates on the staged compliance_warning instead of overloading document_attached, so the server owns the policy in one place. Pinned by a new IB test. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(mcp): trim two descriptions to fit the tools/list payload ceiling after merging main PR #1846 spent the 59,950 headroom; the merged state crossed by 7 tokens. Dropped the redundant filter enumeration from list_pending_operations (the schema documents the filters) and the dims-bags aside from bulk_book_transactions (the lines schema documents the bags). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
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. Add ?tool_namespace=accounted for the Accounted tool names. Requests without it retain the legacy Gnubok namespace. OAuth and stdio bridges live alongside the API surface: see app/api/mcp-oauth/, packages/accounted-mcp/, and the compatibility package in packages/gnubok-mcp/.
Tool authoring contract
Enforced by tests in __tests__/: these are not style preferences, they're guard rails.
additionalProperties: falseon everyinputSchema. Guarded bystrict-schemas.test.ts. Forces clear rejections on hallucinated fields instead of silent ignores.- Descriptions ≤ 280 chars. Guarded by
output-schema.test.ts. NoArgs:/Returns:/Examples:prose: those belong in JSON Schema. Use agent-native hints ("Use to…", "Call X first", "HIGH risk"). - Staged-operation envelope for write tools:
outputSchema: STAGED_OPERATION_SCHEMA(server.ts). Fields:staged, risk_level, actor, message, preview, period_status?, next?. Thestaged: trueboolean is the explicit completion signal; agents must not infer completion from prose. Do NOT introduce a parallel{ success, shouldContinue, output }envelope. period_statusthreading: any tool that ties to a fiscal-period-bound date (categorize, mark paid, create voucher, correct/reverse entry, approve supplier invoice) passesdateForPeriodChecktostagePendingOperation. Response then includesperiod_status: { period_id, status: open|locked|closed, lock_date }so widgets and agents disable writes without round-trips.- Scope mapping: every new tool needs an entry in
lib/auth/api-keys.tsTOOL_SCOPE_MAP. Missing entries default to deny. - Tests for new write tools: add staging-gate coverage to
__tests__/voucher-tools.test.ts(or a sibling) plus executor coverage tolib/pending-operations/__tests__/voucher-executors.test.tsif the tool stages a newoperation_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. If the test fires, the right answer is rarely "raise the ceiling". Instead, trim descriptions or set specialized wide tools to catalogVisibility: 'search'. Those tools remain discoverable with full schemas through gnubok_search_tools and callable through tools/call without bloating the default catalog.
Where things live
server.ts: the tools array + JSON-RPC dispatchertool-result.ts:withNext(),toToolError()response helpersresources/: read-onlyAccounted://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 promptsskills/: domain-knowledge skill bodies served viagnubok_load_skill__tests__/: strictness guards + per-tool coverage