Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Wenn es um BeamtStG, BBG und Landesrecht richtig abgrenzen in Beamtenrecht geht: ordnet Sachverhalt, Norm, Beweislast, Gegenargumente und nächsten Schritt; liefert eine Tatbestands- oder Anspruchsmatrix mit Gegenargumenten.
.claude/skills/klotzkette-beamtstg-bbg-beforderung-erprobung-begrenzte/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-02 | ✓→✗ | ▼ Worse | 31% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 71% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 45% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 49% | 0% |
Der Skill fragt zuerst:
Das BeamtStG ist kein vollständiges Beamtenhandbuch. Es setzt gemeinsame statusrechtliche Eckpunkte für Landes- und Kommunalbeamte. Details der Laufbahn, Besoldung, Versorgung und des Verfahrens stehen in Landesgesetzen. Für Bundesbeamte ist das BBG Ausgangspunkt; daneben stehen BBesG, BeamtVG, BDG und Laufbahnverordnungen.
Der Skill erzeugt eine Normenampel:
| Farbe | Bedeutung | | --- | --- | | Grün | Norm ist sehr wahrscheinlich einschlägig, muss aber in aktueller Fassung geprüft werden. | | Gelb | Norm kann einschlägig sein, hängt an Status/Bundesland/Sonderlaufbahn. | | Rot | Norm gehört zu anderem Dienstherrn oder falscher Materie. |
Wenn der Nutzer ein Aktenzeichen, eine Behördennorm oder eine Verwaltungsvorschrift nennt, prüft der Skill:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 20,173 | 30,956 | +53% | 1 | 1 | 0% | 2,701 | 4,623 | +71% | 0 | 0 | — |
case-02 | pass→fail | 22,736 | 28,528 | +25% | 1 | 1 | 0% | 3,260 | 4,284 | +31% | 0 | 0 | — |
case-03 | pass→pass | 25,963 | 23,076 | -11% | 1 | 1 | 0% | 2,873 | 4,162 | +45% | 0 | 0 | — |
case-04 | fail→pass | 21,002 | 19,802 | -6% | 1 | 1 | 0% | 2,940 | 3,748 | +27% | 0 | 0 | — |
case-05 | pass→pass | 19,000 | 25,866 | +36% | 1 | 1 | 0% | 2,685 | 4,011 | +49% | 0 | 0 | — |
case-06 | pass→pass | 22,733 | 28,036 | +23% | 1 | 1 | 0% | 3,297 | 4,973 | +51% | 0 | 0 | — |
case-07 | pass→pass | 10,083 | 21,816 | +116% | 1 | 1 | 0% | 1,326 | 3,306 | +149% | 0 | 0 | — |
case-08 | pass→pass | 25,820 | 29,191 | +13% | 1 | 1 | 0% | 3,219 | 4,999 | +55% | 0 | 0 | — |
case-09 | pass→pass | 22,993 | 22,609 | -2% | 1 | 1 | 0% | 3,084 | 4,216 | +37% | 0 | 0 | — |
case-10 | pass→pass | 14,522 | 19,776 | +36% | 1 | 1 | 0% | 2,090 | 3,761 | +80% | 0 | 0 | — |
case-11 | pass→pass | 8,519 | 15,556 | +83% | 1 | 1 | 0% | 1,203 | 3,076 | +156% | 0 | 0 | — |
case-12 | fail→fail | 18,189 | 23,921 | +32% | 1 | 1 | 0% | 2,962 | 4,233 | +43% | 0 | 0 | — |
case-13 | pass→pass | 30,011 | 31,876 | +6% | 1 | 1 | 0% | 4,268 | 5,554 | +30% | 0 | 0 | — |
case-14 | pass→pass | 29,909 | 28,721 | -4% | 1 | 1 | 0% | 4,303 | 5,700 | +32% | 0 | 0 | — |
case-15 | pass→pass | 5,363 | 14,077 | +162% | 1 | 1 | 0% | 995 | 2,877 | +189% | 0 | 0 | — |
case-16 | pass→pass | 24,006 | 19,330 | -19% | 1 | 1 | 0% | 3,489 | 4,633 | +33% | 0 | 0 | — |
case-17 | pass→pass | 20,516 | 22,641 | +10% | 1 | 1 | 0% | 3,500 | 4,189 | +20% | 0 | 0 | — |
case-18 | pass→pass | 14,588 | 15,176 | +4% | 1 | 1 | 0% | 2,012 | 3,379 | +68% | 0 | 0 | — |
case-19 | pass→pass | 10,302 | 19,279 | +87% | 1 | 1 | 0% | 1,724 | 3,575 | +107% | 0 | 0 | — |
case-20 | pass→pass | 20,705 | 26,054 | +26% | 1 | 1 | 0% | 3,512 | 4,533 | +29% | 0 | 0 | — |
case-21 | pass→pass | 17,979 | 18,284 | +2% | 1 | 1 | 0% | 2,545 | 3,515 | +38% | 0 | 0 | — |
case-22 | pass→pass | 19,352 | 19,337 | -0% | 1 | 1 | 0% | 2,811 | 3,604 | +28% | 0 | 0 | — |
case-23 | pass→pass | 14,468 | 21,924 | +52% | 1 | 1 | 0% | 2,570 | 4,288 | +67% | 0 | 0 | — |
case-24 | pass→pass | 21,552 | 26,490 | +23% | 1 | 1 | 0% | 3,074 | 4,575 | +49% | 0 | 0 | — |
case-25 | pass→pass | 19,229 | 21,005 | +9% | 1 | 1 | 0% | 2,611 | 4,561 | +75% | 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. 25 cases were attempted. The headline lift of 0 percentage points is the difference between those two pass rates over the 25 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.