Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Wenn es um Internationale Harmonisierung Konzern-BAV in BAV Strategie Konzern — Treuenfels Yamamoto Rechtsanwälte geht: prüft Frist, Form, Zuständigkeit, Rechtsweg und Sofortmaßnahmen; liefert eine Fristen- und Risikoampel mit Sofortschritten.
.claude/skills/klotzkette-internationale-harmonisierung-japan-corporate/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 136% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 173% | 0% |
| case-20 | ✓→✗ | ▼ Worse | 123% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 141% | 0% |
| case-05 | ✓→✓ | = Same ✓ | 179% | 0% |
Erster Schritt bei jedem internationalen Mandat: lückenlose Bestandsaufnahme aller Pensionspläne konzernweit.
Inventory-Dimensionen:
| Dimension | Inhalt | |-----------|--------| | Plantyp | DB/DC/Hybrid; staatlich, branchentariflich, betrieblich | | Durchführungsweg | Trust/Versicherung/Direktzusage/staatliches System | | Begünstigte | Aktive Anwärter/Rentner/Ausgeschiedene (differenziert) | | DBO (IAS 19) | Je Plan in Lokalwährung und EUR | | Funding Level | Planvermögen / DBO in % | | Governance | Board/Trustee-Struktur; Sponsor-Rechte | | Rechtliche Basis | Tarifvertrag/Gesetz/Einzelvertrag | | Lokale Aufsicht | Zuständige Behörde und Regelwerk | | Risiken | Longevity/Investment/Political/Regulatory |
bAV-Projektteam erstellt für jeden Konzern-Mandanten eine länderspezifische Analyse folgender Kernländer (Hauptmärkte DE/UK/US/FR/SG/JP):
pensionsmodelle-fuenf-durchfuehrungswege)drei-stufen-theorie-eingriffsanalyse)expatriate-pensionsplanung-und-totalization)japan-bav-und-corporate-pension-iorp)fachliche Leitung entwickelt für Konzern-Mandanten eine Global Benefits Policy als konzerninterne Weisung:
Bestandteile GBP:
internationale-buyout-datenflows-und-datenschutz)| Entscheidung | Local (HR/Legal) | Group (Pension Committee) | |-------------|-----------------|--------------------------| | Plandesign-Änderungen < EUR 500k Kosten p.a. | ✓ (mit Notification) | — | | Plandesign-Änderungen ≥ EUR 500k Kosten p.a. | — | ✓ (Approval required) | | Neue Versorgungseinrichtung | — | ✓ | | Liquidation / Restrukturierung Plan | — | ✓ | | Trustee-Ernennung | Local Board | Group notified | | Kapitalanlage-Entscheidungen | Local Trustee | IPS genehmigt durch Group | | M&A-Pensionsrisiken ab EUR 10 Mio. | — | ✓ (Pre-deal Approval) |
GLOBAL PENSION INVENTORY REPORT
[Konzern Muster AG], Berichtsjahr [JJJJ]
Erstellt durch: bAV-Projektteam
Verantwortlich: fachliche Leitung
EXECUTIVE SUMMARY
Gesamtzahl Pensionspläne weltweit: [X]
Länder mit DB-Plänen: [X]
Gesamte DBO (IAS 19, EUR): EUR [Betrag] Mio.
Gesamtes Planvermögen (IAS 19, EUR): EUR [Betrag] Mio.
Net Pension Liability (Bilanz, EUR): EUR [Betrag] Mio.
Gewichteter Funding Level: [X]%
Anzahl aktiver Anwärter weltweit: [X]
Anzahl Rentner: [X]
Anzahl unverfallbar Ausgeschiedener: [X]
LÄNDERÜBERSICHT:
Land | Pläne | DBO EUR Mio. | FundLevel | Risikostufe | Status
DE | [X] | [X] | [X]% | Mittel-Hoch | Aktiv
UK | [X] | [X] | [X]% | Mittel | De-Risking
US | [X] | [X] | [X]% | Gering | Frozen
FR | [X] | [X] | [X]% | Gering | Aktiv
SG | [X] | CPF-only | n.a. | Sehr Gering | Pflicht
JP | [X] | [X] | [X]% | Mittel | DC-Reformpensionsmodelle-fuenf-durchfuehrungswege — Deutschlandspezifikjapan-bav-und-corporate-pension-iorp — Japan-Vertiefungcountry-by-country-benefits-matrix-konzern — Detailmatrix je Landexpatriate-pensionsplanung-und-totalization — Expat-Sonderfragen> Quellenregel: Entscheidungen nur nach Prüfung einer amtlichen oder frei zugänglichen Quelle mit Gericht, Entscheidungsform, Datum, Aktenzeichen und tragender Aussage ausgeben.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 39,656 | 46,723 | +18% | 1 | 1 | 0% | 5,075 | 9,402 | +85% | 0 | 0 | — |
case-02 | fail→fail | 29,906 | 34,018 | +14% | 1 | 1 | 0% | 3,940 | 7,617 | +93% | 0 | 0 | — |
case-03 | fail→fail | 63,437 | 56,177 | -11% | 1 | 1 | 0% | 8,200 | 10,384 | +27% | 0 | 0 | — |
case-04 | pass→pass | 26,002 | 29,825 | +15% | 1 | 1 | 0% | 2,710 | 6,541 | +141% | 0 | 0 | — |
case-05 | pass→pass | 16,634 | 30,706 | +85% | 1 | 1 | 0% | 2,716 | 7,588 | +179% | 0 | 0 | — |
case-06 | pass→pass | 20,631 | 25,199 | +22% | 1 | 1 | 0% | 2,747 | 6,469 | +135% | 0 | 0 | — |
case-07 | fail→fail | 26,102 | 27,467 | +5% | 1 | 1 | 0% | 3,780 | 7,394 | +96% | 0 | 0 | — |
case-08 | fail→pass | 23,209 | 27,445 | +18% | 1 | 1 | 0% | 3,012 | 7,099 | +136% | 0 | 0 | — |
case-09 | pass→pass | 24,169 | 19,012 | -21% | 1 | 1 | 0% | 3,316 | 6,840 | +106% | 0 | 0 | — |
case-10 | fail→fail | 12,522 | 21,376 | +71% | 1 | 1 | 0% | 2,023 | 6,386 | +216% | 0 | 0 | — |
case-11 | pass→pass | 19,995 | 35,676 | +78% | 1 | 1 | 0% | 2,610 | 8,475 | +225% | 0 | 0 | — |
case-12 | pass→pass | 23,762 | 28,131 | +18% | 1 | 1 | 0% | 3,234 | 7,295 | +126% | 0 | 0 | — |
case-13 | fail→fail | 22,074 | 23,168 | +5% | 1 | 1 | 0% | 3,281 | 6,428 | +96% | 0 | 0 | — |
case-14 | fail→pass | 18,776 | 38,998 | +108% | 1 | 1 | 0% | 2,439 | 6,653 | +173% | 0 | 0 | — |
case-15 | pass→pass | 21,342 | 26,522 | +24% | 1 | 1 | 0% | 3,382 | 7,523 | +122% | 0 | 0 | — |
case-16 | fail→fail | 18,534 | 34,820 | +88% | 1 | 1 | 0% | 2,801 | 6,520 | +133% | 0 | 0 | — |
case-17 | pass→pass | 18,357 | 24,154 | +32% | 1 | 1 | 0% | 2,885 | 6,732 | +133% | 0 | 0 | — |
case-18 | pass→pass | 17,213 | 22,872 | +33% | 1 | 1 | 0% | 2,694 | 6,944 | +158% | 0 | 0 | — |
case-19 | pass→pass | 10,548 | 17,848 | +69% | 1 | 1 | 0% | 1,916 | 6,754 | +253% | 0 | 0 | — |
case-20 | pass→fail | 16,964 | 21,986 | +30% | 1 | 1 | 0% | 3,122 | 6,973 | +123% | 0 | 0 | — |
case-21 | pass→pass | 17,931 | 28,758 | +60% | 1 | 1 | 0% | 3,093 | 7,685 | +148% | 0 | 0 | — |
case-22 | fail→fail | 28,422 | 36,979 | +30% | 1 | 1 | 0% | 4,819 | 9,360 | +94% | 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 +5 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are 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.