Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Wenn es um Pdfs: Compliance-Dokumentation und Aktenvermerk in Barrierefreiheit Web Checker geht: ordnet Akteninhalt, Belege, Lücken und Nachforderungen; liefert ein direkt nutzbares Arbeitsprodukt mit Prüfpunkten, Risiken und nächstem Schritt.
.claude/skills/klotzkette-pdfs-compliance-dokumentation-und-akte/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 30% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 74% | 0% |
Vor einer rechtlichen Schlussfolgerung diese Anker am aktuellen Normtext prüfen; Spezial- und Landesrecht nur hinzunehmen, wenn es den konkreten Auftrag traegt:
§ 241 Abs. 2 BGB — Rücksichtnahme-, Schutz- und Organisationspflichten.§ 242 BGB — Treu und Glauben als Korrektiv enger Klausel- und Anspruchsarbeit.§ 280 Abs. 1 BGB — Pflichtverletzung, Vertretenmuessen, Schaden.§ 286 Abs. 1 BGB — Verzug und Fristlogik.§ 195 BGB — regelmäßige Verjährung.§ 199 Abs. 1 BGB — Beginn der regelmäßigen Verjährung.§ 253 Abs. 2 ZPO — Bestimmtheit von Antrag und Klagegrund.§ 138 Abs. 1 ZPO — Wahrheitspflicht und vollstaendiger Tatsachenvortrag.Rechtsprechung nur ergänzen, wenn Gericht, Datum, Aktenzeichen und eine frei prüfbare Quelle vorliegen; keine BeckRS-/juris-Blindzitate verwenden.
Wenn Unterlagen vorhanden sind, arbeite zuerst aus den Unterlagen. Stelle nur Rückfragen, die die nächste Weiche verändern:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 38,149 | 38,626 | +1% | 1 | 1 | 0% | 4,849 | 5,841 | +20% | 0 | 0 | — |
case-02 | fail→pass | 26,667 | 33,310 | +25% | 1 | 1 | 0% | 4,459 | 5,375 | +21% | 0 | 0 | — |
case-03 | fail→pass | 36,218 | 38,880 | +7% | 1 | 1 | 0% | 5,090 | 6,620 | +30% | 0 | 0 | — |
case-04 | fail→fail | 31,998 | 34,267 | +7% | 1 | 1 | 0% | 3,675 | 6,098 | +66% | 0 | 0 | — |
case-05 | fail→fail | 27,155 | 33,913 | +25% | 1 | 1 | 0% | 3,903 | 6,118 | +57% | 0 | 0 | — |
case-06 | fail→fail | 22,850 | 40,379 | +77% | 1 | 1 | 0% | 3,819 | 6,089 | +59% | 0 | 0 | — |
case-07 | fail→pass | 23,895 | 20,945 | -12% | 1 | 1 | 0% | 3,280 | 4,551 | +39% | 0 | 0 | — |
case-08 | fail→fail | 25,678 | 30,376 | +18% | 1 | 1 | 0% | 3,348 | 5,309 | +59% | 0 | 0 | — |
case-09 | fail→fail | 23,042 | 28,718 | +25% | 1 | 1 | 0% | 3,768 | 5,981 | +59% | 0 | 0 | — |
case-10 | pass→pass | 24,647 | 30,343 | +23% | 1 | 1 | 0% | 3,360 | 5,829 | +73% | 0 | 0 | — |
case-11 | fail→fail | 24,337 | 26,079 | +7% | 1 | 1 | 0% | 3,504 | 4,635 | +32% | 0 | 0 | — |
case-12 | fail→fail | 25,501 | 27,851 | +9% | 1 | 1 | 0% | 3,302 | 4,696 | +42% | 0 | 0 | — |
case-13 | fail→fail | 21,039 | 27,348 | +30% | 1 | 1 | 0% | 2,720 | 4,653 | +71% | 0 | 0 | — |
case-14 | fail→fail | 20,452 | 30,329 | +48% | 1 | 1 | 0% | 2,577 | 4,279 | +66% | 0 | 0 | — |
case-15 | fail→pass | 23,275 | 32,613 | +40% | 1 | 1 | 0% | 2,811 | 4,895 | +74% | 0 | 0 | — |
case-16 | fail→fail | 31,189 | 46,834 | +50% | 1 | 1 | 0% | 3,353 | 4,415 | +32% | 0 | 0 | — |
case-17 | fail→pass | 27,046 | 36,771 | +36% | 1 | 1 | 0% | 3,623 | 5,486 | +51% | 0 | 0 | — |
case-18 | pass→pass | 23,441 | 28,157 | +20% | 1 | 1 | 0% | 2,989 | 4,308 | +44% | 0 | 0 | — |
case-19 | fail→pass | 27,860 | 27,679 | -1% | 1 | 1 | 0% | 3,221 | 4,575 | +42% | 0 | 0 | — |
case-20 | fail→fail | 22,001 | 22,306 | +1% | 1 | 1 | 0% | 2,824 | 4,599 | +63% | 0 | 0 | — |
case-21 | fail→fail | 20,478 | 33,191 | +62% | 1 | 1 | 0% | 3,465 | 5,131 | +48% | 0 | 0 | — |
case-22 | pass→pass | 20,717 | 30,113 | +45% | 1 | 1 | 0% | 3,279 | 5,067 | +55% | 0 | 0 | — |
case-23 | pass→pass | 23,824 | 27,394 | +15% | 1 | 1 | 0% | 2,942 | 4,742 | +61% | 0 | 0 | — |
case-24 | fail→pass | 29,953 | 32,718 | +9% | 1 | 1 | 0% | 4,024 | 5,626 | +40% | 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. 24 cases were attempted. The headline lift of +33 percentage points is the difference between those two pass rates over the 24 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.