Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Wenn es um Avalrahmenlinie und Kautionsaval in Bank-Rechtsabteilung geht: ordnet Sachverhalt, Norm, Beweislast, Gegenargumente und nächsten Schritt; liefert ein direkt nutzbares Arbeitsprodukt mit Prüfpunkten, Risiken und nächstem Schritt.
.claude/skills/klotzkette-avalrahmenlinie-kautionsaval-praxis/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 77% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 64% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 171% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 82% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 78% | 0% |
Avalrahmen und Kautionsaval in der Bankpraxis prüfen: Kreditgeschäft, Avalprovision, Limit, Sicherheiten, Abrufrisiko, Text der Garantie/Bürgschaft, § 1 KWG, §§ 765 ff. BGB, §§ 349 und 350 HGB und Regress sauber dokumentieren. Arbeite entlang dieser konkreten Prüfungslinie und trenne Rolle, Frist, Zuständigkeit, Beweislast und gewünschten Output.
| Punkt | Eintrag | | --- | --- | | Kunde / Gruppe | Schuldner, Mutter, Projektgesellschaft, Konsortialbezug | | Begünstigter | Name, Land, Rolle, Streitnähe | | Betrag / Währung | Nennbetrag, Teilabruf, Kumulation mit anderen Linien | | Laufzeit | fest, unbefristet, auto-renewal, Kündigungsfenster | | Texttyp | Bankformular, Begünstigtenformular, FIDIC/öffentliche Hand/Sondertext | | Sicherheiten | Cash Cover, Grundschuld, Globalzession, Patronat, Rang | | Abrufrisiko | niedrig, mittel, hoch mit Grund |
Prüfe den Avaltext in dieser Reihenfolge:
Avale schmieren Liquidität, weil der Kunde keine Barkaution bindet. Genau deshalb muss die Bank prüfen:
Liefere eine Avalfreigabe-Notiz:
kreditsicherheiten-bestellung-verwertung.garantieabruf-missbrauch-und-zahlungsstopp.restrukturierung-kreditengagement, forbearance-npe-risikoklassifizierung, insolvenz-anfechtung-bank.trade-finance-sanctions-lc-guarantee.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 35,796 | 56,554 | +58% | 1 | 1 | 0% | 4,911 | 7,787 | +59% | 0 | 0 | — |
case-02 | fail→fail | 41,806 | 43,239 | +3% | 1 | 1 | 0% | 5,826 | 8,240 | +41% | 0 | 0 | — |
case-03 | fail→pass | 34,753 | 61,088 | +76% | 1 | 1 | 0% | 4,842 | 8,577 | +77% | 0 | 0 | — |
case-04 | pass→pass | 27,285 | 36,236 | +33% | 1 | 1 | 0% | 3,400 | 6,805 | +100% | 0 | 0 | — |
case-05 | pass→pass | 18,660 | 26,071 | +40% | 1 | 1 | 0% | 2,060 | 4,889 | +137% | 0 | 0 | — |
case-06 | fail→pass | 21,272 | 31,128 | +46% | 1 | 1 | 0% | 3,398 | 5,568 | +64% | 0 | 0 | — |
case-07 | pass→pass | 25,568 | 28,334 | +11% | 1 | 1 | 0% | 3,261 | 5,628 | +73% | 0 | 0 | — |
case-08 | pass→pass | 23,476 | 34,586 | +47% | 1 | 1 | 0% | 2,880 | 6,580 | +128% | 0 | 0 | — |
case-09 | pass→pass | 20,744 | 30,371 | +46% | 1 | 1 | 0% | 2,619 | 5,474 | +109% | 0 | 0 | — |
case-10 | pass→pass | 19,411 | 32,187 | +66% | 1 | 1 | 0% | 2,415 | 5,817 | +141% | 0 | 0 | — |
case-11 | pass→pass | 19,600 | 26,880 | +37% | 1 | 1 | 0% | 2,867 | 5,529 | +93% | 0 | 0 | — |
case-12 | pass→pass | 18,203 | 26,022 | +43% | 1 | 1 | 0% | 2,589 | 5,221 | +102% | 0 | 0 | — |
case-13 | pass→pass | 18,844 | 19,457 | +3% | 1 | 1 | 0% | 2,155 | 5,192 | +141% | 0 | 0 | — |
case-14 | fail→pass | 10,552 | 17,992 | +71% | 1 | 1 | 0% | 1,868 | 5,070 | +171% | 0 | 0 | — |
case-15 | fail→fail | 18,464 | 28,408 | +54% | 1 | 1 | 0% | 2,460 | 6,235 | +153% | 0 | 0 | — |
case-16 | fail→pass | 20,381 | 22,298 | +9% | 1 | 1 | 0% | 2,935 | 5,340 | +82% | 0 | 0 | — |
case-17 | fail→pass | 20,032 | 25,101 | +25% | 1 | 1 | 0% | 3,086 | 5,507 | +78% | 0 | 0 | — |
case-18 | pass→pass | 18,858 | 21,825 | +16% | 1 | 1 | 0% | 3,293 | 5,704 | +73% | 0 | 0 | — |
case-19 | fail→fail | 19,705 | 24,226 | +23% | 1 | 1 | 0% | 3,357 | 5,728 | +71% | 0 | 0 | — |
case-20 | pass→pass | 19,610 | 24,198 | +23% | 1 | 1 | 0% | 3,573 | 5,977 | +67% | 0 | 0 | — |
case-21 | pass→pass | 18,274 | 31,321 | +71% | 1 | 1 | 0% | 3,431 | 7,395 | +116% | 0 | 0 | — |
case-22 | pass→pass | 20,833 | 29,355 | +41% | 1 | 1 | 0% | 3,462 | 7,118 | +106% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +23 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.