skills/agent-evals/assets/trigger-queries.elster-form-capture.json
|
[ { "id": "pos-01", "query": "trag mir die Steuererklärung 2024 in Mein ELSTER ein", "should_trigger": true, "split": "train" }, { "id": "pos-02", "query": "the Anlage V line numbers in my guide don't match the form this year", "should_trigger": true, "split": "train", "note": "the central teaching: Kennzahlen stable, sub-page and line numbers not" }, { "id": "pos-03", "query": "welche Kennzahl hat das Feld für Erhaltungsaufwand in der Anlage V", "should_trigger": true, "split": "validation", "note": "resolved by references/feldkarte-est.md" }, { "id": "pos-04", "query": "my second entry keeps overwriting the first one when I add rows with Playwright", "should_trigger": true, "split": "train", "note": "never says ELSTER; the moving AddMzbItem index and the id-prefix collision" }, { "id": "pos-05", "query": "continue the existing draft and get it to say es sind keine Fehler vorhanden", "should_trigger": true, "split": "validation" }, { "id": "pos-06", "query": "prüf bitte maschinell nach, ob wirklich alle Beträge im Formular gelandet sind", "should_trigger": true, "split": "train", "note": "the target-value comparison against the summary page text" }, { "id": "pos-07", "query": "ELSTER complains about missing Krankenversicherungsbeiträge although the employer transmitted a value", "should_trigger": true, "split": "validation", "note": "the transferred 0.00 that must be emptied rather than left at zero" }, { "id": "pos-08", "query": "which Teilseiten of the Anlage Vorsorgeaufwand still show no data", "should_trigger": true, "split": "train", "note": "the Daten vorhanden completeness test" }, { "id": "pos-09", "query": "automate filling a German tax portal form with the browser tools", "should_trigger": true, "split": "validation", "note": "names neither ELSTER nor an Anlage" }, { "id": "pos-10", "query": "der Betrag im Formular weicht um einen Euro von meiner Kalkulation ab", "should_trigger": true, "split": "train", "note": "ELSTER rounds in the taxpayer's favour, not commercially" }, { "id": "neg-01", "query": "kann ich das häusliche Arbeitszimmer 2024 überhaupt absetzen", "should_trigger": false, "split": "train", "note": "substantive deductibility — german-tax-research" }, { "id": "neg-02", "query": "schreib mir eine Einspruchsbegründung gegen den Schätzungsbescheid", "should_trigger": false, "split": "train", "note": "drafting a submission — german-tax-research" }, { "id": "neg-03", "query": "wann läuft die Einspruchsfrist für einen Bescheid vom 12. März ab", "should_trigger": false, "split": "validation", "note": "deadline computation under the notification fiction — german-tax-research" }, { "id": "neg-04", "query": "set up a persistent Edge profile so the login survives a restart", "should_trigger": false, "split": "train", "note": "authenticated-web-extraction; this Skill never handles credentials" }, { "id": "neg-05", "query": "scrape my LinkedIn contacts into a table", "should_trigger": false, "split": "validation", "note": "authenticated-web-extraction" }, { "id": "neg-06", "query": "extract the amounts from these scanned Handwerkerrechnungen", "should_trigger": false, "split": "train", "note": "pdf-to-markdown" }, { "id": "neg-07", "query": "read the Betriebskostenabrechnung spreadsheet and give me the totals", "should_trigger": false, "split": "validation", "note": "xlsx-to-markdown" }, { "id": "neg-08", "query": "bau mir aus den Belegen ein durchnummeriertes Anlagenkonvolut mit Deckblatt", "should_trigger": false, "split": "train", "note": "evidence-package-assembly" }, { "id": "neg-09", "query": "der Vermieter rechnet die Gartenpflege falsch um, was sagt das Mietrecht", "should_trigger": false, "split": "validation", "note": "german-legal-research" }, { "id": "neg-10", "query": "fill in this PDF form with the AcroForm field values", "should_trigger": false, "split": "train", "note": "near miss on 'fill a form'; pdf-to-markdown covers AcroForm filling" } ] |