Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Wenn es um Pensionsmodelle: Risikoampel, Gegenargumente und Verteidigungslinien in BAV Strategie Konzern — Treuenfels Yamamoto Rechtsanwälte geht: zerlegt Ergebnis, Frist, Zuständigkeit, Beweislast und Gegenposition; liefert eine Fristen- und Risikoampel mit Sofortschritten.
.claude/skills/klotzkette-pensionsmodelle-risikoampel-und-gegenargumente/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 75% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 80% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 62% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 88% | 0% |
Wenn Unterlagen vorhanden sind, arbeite zuerst aus den Unterlagen. Stelle nur Rückfragen, die die nächste Weiche verändern:
Bei Veränderung von Versorgungszusagen sind drei Eingriffsklassen zu unterscheiden:
| Stufe | Gegenstand | Eingriffsvoraussetzung | |---|---|---| | 1 | bereits erdiente und dynamische Steigerungen (z. B. gehaltsabhängig) | nur bei zwingenden Gründen / Existenzbedrohung | | 2 | künftige dienstzeitabhängige Zuwächse | triftige Gründe (z. B. wirtschaftlich begründete Notlage) | | 3 | künftige dienstzeitunabhängige Zuwächse | sachlich-proportionale Gründe |
| Konstellation | Ampel | Grund | |---|---|---| | Schließung Versorgungswerk für Neueintritte (Cut-off) | Grün-Gelb | Stufe 3, regelmäßig zulässig bei sachlichen Gründen | | Festschreibung künftiger Zuwächse für Bestand | Gelb | Stufe 2, triftige Gründe nötig | | Kürzung bereits erdienter Anwartschaften | Rot | Stufe 1, nur bei Existenznotlage | | DB-zu-DC-Umstellung bei Neueintritten | Grün | reine Plan-Designänderung | | DB-zu-DC-Umstellung für Bestand mit Bestandsschutz | Gelb-Rot | Stufenprüfung, häufig Mitbestimmung |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 40,708 | 46,789 | +15% | 1 | 1 | 0% | 5,706 | 7,567 | +33% | 0 | 0 | — |
case-02 | fail→fail | 36,949 | 39,442 | +7% | 1 | 1 | 0% | 4,208 | 7,117 | +69% | 0 | 0 | — |
case-03 | fail→fail | 30,282 | 35,489 | +17% | 1 | 1 | 0% | 4,761 | 6,703 | +41% | 0 | 0 | — |
case-04 | pass→pass | 28,997 | 35,238 | +22% | 1 | 1 | 0% | 4,037 | 6,684 | +66% | 0 | 0 | — |
case-05 | pass→pass | 25,027 | 32,455 | +30% | 1 | 1 | 0% | 3,426 | 6,253 | +83% | 0 | 0 | — |
case-06 | pass→pass | 11,330 | 13,646 | +20% | 1 | 1 | 0% | 1,934 | 3,895 | +101% | 0 | 0 | — |
case-07 | fail→fail | 23,244 | 28,850 | +24% | 1 | 1 | 0% | 3,071 | 5,522 | +80% | 0 | 0 | — |
case-08 | fail→pass | 19,778 | 19,316 | -2% | 1 | 1 | 0% | 2,695 | 4,719 | +75% | 0 | 0 | — |
case-09 | fail→pass | 26,941 | 28,305 | +5% | 1 | 1 | 0% | 3,755 | 5,779 | +54% | 0 | 0 | — |
case-10 | fail→pass | 23,233 | 28,342 | +22% | 1 | 1 | 0% | 3,173 | 5,705 | +80% | 0 | 0 | — |
case-11 | fail→pass | 22,130 | 26,118 | +18% | 1 | 1 | 0% | 3,417 | 5,531 | +62% | 0 | 0 | — |
case-12 | pass→pass | 24,703 | 22,096 | -11% | 1 | 1 | 0% | 3,544 | 4,673 | +32% | 0 | 0 | — |
case-13 | pass→pass | 23,687 | 22,470 | -5% | 1 | 1 | 0% | 3,726 | 5,011 | +34% | 0 | 0 | — |
case-14 | pass→fail | 20,768 | 21,802 | +5% | 1 | 1 | 0% | 3,101 | 5,073 | +64% | 0 | 0 | — |
case-15 | pass→pass | 13,124 | 20,426 | +56% | 1 | 1 | 0% | 2,299 | 4,550 | +98% | 0 | 0 | — |
case-16 | pass→pass | 11,731 | 17,404 | +48% | 1 | 1 | 0% | 1,981 | 4,570 | +131% | 0 | 0 | — |
case-17 | pass→pass | 7,631 | 8,066 | +6% | 1 | 1 | 0% | 1,152 | 3,047 | +164% | 0 | 0 | — |
case-18 | pass→pass | 12,314 | 16,546 | +34% | 1 | 1 | 0% | 1,543 | 4,013 | +160% | 0 | 0 | — |
case-19 | fail→pass | 10,906 | 9,988 | -8% | 1 | 1 | 0% | 1,786 | 3,363 | +88% | 0 | 0 | — |
case-20 | pass→pass | 17,009 | 17,884 | +5% | 1 | 1 | 0% | 2,169 | 4,445 | +105% | 0 | 0 | — |
case-21 | pass→pass | 17,370 | 20,152 | +16% | 1 | 1 | 0% | 2,732 | 4,856 | +78% | 0 | 0 | — |
case-22 | pass→pass | 12,542 | 23,232 | +85% | 1 | 1 | 0% | 2,035 | 4,284 | +111% | 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 +18 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.