#2066 asked for input_examples on the top ~20 tools by call volume. The
binding constraint turned out to be budget, not writing: after #2089
reclaimed 3 763 tokens, its own policy required ratcheting the tools/list
ceiling down with it, so the real headroom was 317 tokens. Ten examples
across five tools cost 199, leaving ~118. The ceiling is not raised.
Tools picked from 30 days of mcp.tool_called crossed with the
combinations the descriptions already warn about and callers still get
wrong: account_override without an explicit vat_treatment (books gross,
no moms line), representation without deltagare and syfte, confirmed on
a high-risk approval, a balanced voucher where the moms leg is its own
line, and get_kpi_report, where one caller sent `metric` 604 times over
seven days to a tool whose only parameter is period_id.
Examples are also surfaced in the unknown-parameter error. That costs
nothing in tools/list, because it only ships on the response to a call
that already failed, and it reaches the caller that most needs it: a key
list told DueCue's agent which parameter was wrong but not what a
correct call looks like, and the same rejected call repeated for a week.
Every example is validated against its own schema by the same
findUnknownArgKeys guard the server runs, plus required/type/enum/pattern
checks. An example our own boundary would reject is worse than none: it
teaches the exact mistake the guard then punishes. That test caught three
invented enum values in this change's own first draft.
Claude-Session: https://claude.ai/code/session_01L3P2hr19PhQuCoTSGoegcY
Co-authored-by: Jakob Wennberg <311770904+jakobwennberg-oss@users.noreply.github.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Two reports from the 08-24 feedback sweep (seq 261545):
- The bulk-book queue title (Samlingsverifikation: 1 transaktioner
2026-07-22) carried no amount, direction or counterparty; the CEO
approving from a phone could not tell what he authorised. Titles now
read: Samlingsverifikation -1 000,00 SEK 2026-05-12: NORDNET UTTAG
(+1 till). Same per-tx text the categorize titles already carry;
preview_data stays aggregate-only.
- gnubok_query_journal called with {query} instead of {text} silently
returned the whole journal. tools/call now rejects unknown top-level
parameters for every tool (all schemas declare additionalProperties:
false) with a VALIDATION_ERROR that lists the valid keys. company_id
stays tolerated everywhere. codedError is exported from
company-routing for the dispatcher.
The copy fixes this branch originally carried (scope-honest
list_pending_operations, BFL 5 kap 6 § on create_voucher, bank-movement
only on categorize/bulk_book) landed independently in #1844 and were
dropped on rebase; no catalog token change remains.
Claude-Session: https://claude.ai/code/session_01ScVhg6XsDtNXkiEQNV7LaZ
Co-authored-by: Jakob Wennberg <311770904+jakobwennberg-oss@users.noreply.github.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
