Jakob Wennberg
a894af11d0
* feat(mcp): tool titles + Connect-with-Claude docs for Connectors Directory readiness Add a top-level Tool.title (MCP spec 2025-06-18) to all 92 MCP tools and flow it through the tools/list serializer for the Claude Connectors Directory. Trim the ~10 longest tool descriptions toward 180-200 chars (semantics + agent hints preserved) to partly offset the added bytes, and raise the payload-size bench ceiling 32K → 36K with headroom reserved for upcoming Skatteverket tools. Ship a "Connect with Claude" docs page (OAuth 2.1 connector + npx gnubok-mcp stdio bridge, sample sandbox prompts, 10-minute reviewer test, support pointer), wire it into the docs nav, and add a README to the packages/gnubok-mcp npm bridge. Bump one categorized sandbox business expense above 4 000 kr so the VAT close-check demo surfaces a high-value receipt-less expense. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(mcp): restore six spaces lost in the description-trimming pass Review finding (Greptile P2): the trim deleted single spaces in six tool descriptions ("Stages foruser", "länkatill", "kundfordran(1510)", …), producing malformed text in tools/list. All six were pure space deletions with no token gain — restored verbatim. Swept all 92 descriptions for residual merge artifacts; the two remaining matches are pre-existing function-call notation from main. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(compliance): correct BFL citation in seed + mark 4000 kr as app heuristic Swedish compliance review findings: - The seed comment cited BFL 5 kap 6§ as requiring verifikat on high-value affärshändelser; 6§ governs verifikationens innehåll. Rephrased: BFL 5 kap 6-7§ require every affärshändelse to be documented — the 4 000 kr cut-off is the tool's own heuristic, not a statutory threshold. - The connect-claude docs read as though 4 000 kr were a legal floor (conflatable with the förenklad faktura ceiling, ML 17 kap 26-28§). Clarified that BFL requires underlag regardless of amount. Dismissed as by-design: the gnubok_sk_test_... README placeholders (standard practice; the prefix schema is already public in llms.txt). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
3.8 KiB
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.
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 (currently 36,000 tokens — bumped from 32,000 when top-level Tool.title landed on all tools for Claude Connectors Directory readiness). 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 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