Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户需要整理或理解理赔案件信息时使用此 skill。适用于理赔案件摘要生成、案件进度跟踪、关键信息提取等场景。
.claude/skills/aifinlab-claims-case-summary/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 7% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 4% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-21 | ✗→✓ | ▲ Improved | 60% | 0% |
你是一名经验丰富的保险理赔顾问，目标是帮助用户清晰整理理赔案件信息，跟踪案件进度，识别关键节点和风险点。
围绕用户的理赔案件，产出结构化、清晰、可追溯的案件摘要。优先帮助用户解决以下任务：
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 25,377 | 28,526 | +12% | 1 | 1 | 0% | 3,911 | 4,609 | +18% | 0 | 0 | — |
case-02 | fail→fail | 5,620 | 8,023 | +43% | 1 | 1 | 0% | 1,003 | 1,919 | +91% | 0 | 0 | — |
case-03 | fail→fail | 20,211 | 17,192 | -15% | 1 | 1 | 0% | 3,048 | 3,198 | +5% | 0 | 0 | — |
case-04 | pass→pass | 19,809 | 17,563 | -11% | 1 | 1 | 0% | 3,038 | 3,283 | +8% | 0 | 0 | — |
case-05 | pass→pass | 20,358 | 17,455 | -14% | 1 | 1 | 0% | 3,043 | 3,097 | +2% | 0 | 0 | — |
case-06 | fail→pass | 19,642 | 17,511 | -11% | 1 | 1 | 0% | 3,000 | 3,219 | +7% | 0 | 0 | — |
case-07 | fail→pass | 16,369 | 16,563 | +1% | 1 | 1 | 0% | 2,414 | 2,960 | +23% | 0 | 0 | — |
case-08 | pass→pass | 11,112 | 10,524 | -5% | 1 | 1 | 0% | 2,295 | 2,398 | +4% | 0 | 0 | — |
case-09 | pass→fail | 29,008 | 28,765 | -1% | 1 | 1 | 0% | 4,171 | 4,836 | +16% | 0 | 0 | — |
case-10 | fail→fail | 14,298 | 14,707 | +3% | 1 | 1 | 0% | 2,230 | 2,990 | +34% | 0 | 0 | — |
case-11 | fail→fail | 9,835 | 12,261 | +25% | 1 | 1 | 0% | 1,461 | 2,321 | +59% | 0 | 0 | — |
case-12 | pass→pass | 22,583 | 19,754 | -13% | 1 | 1 | 0% | 3,294 | 3,371 | +2% | 0 | 0 | — |
case-13 | fail→pass | 28,962 | 25,840 | -11% | 1 | 1 | 0% | 4,211 | 4,391 | +4% | 0 | 0 | — |
case-14 | fail→fail | 13,572 | 16,895 | +24% | 1 | 1 | 0% | 2,116 | 3,204 | +51% | 0 | 0 | — |
case-15 | fail→fail | 14,244 | 13,421 | -6% | 1 | 1 | 0% | 2,101 | 2,552 | +21% | 0 | 0 | — |
case-16 | fail→pass | 23,180 | 20,894 | -10% | 1 | 1 | 0% | 3,307 | 3,643 | +10% | 0 | 0 | — |
case-17 | fail→fail | 14,687 | 13,788 | -6% | 1 | 1 | 0% | 2,288 | 2,866 | +25% | 0 | 0 | — |
case-18 | pass→pass | 21,197 | 20,199 | -5% | 1 | 1 | 0% | 3,132 | 3,725 | +19% | 0 | 0 | — |
case-19 | pass→pass | 28,210 | 25,659 | -9% | 1 | 1 | 0% | 4,107 | 4,561 | +11% | 0 | 0 | — |
case-20 | fail→fail | 13,686 | 16,114 | +18% | 1 | 1 | 0% | 2,132 | 3,110 | +46% | 0 | 0 | — |
case-21 | fail→pass | 9,636 | 12,078 | +25% | 1 | 1 | 0% | 1,512 | 2,419 | +60% | 0 | 0 | — |
case-22 | pass→pass | 19,525 | 18,642 | -5% | 1 | 1 | 0% | 2,989 | 3,338 | +12% | 0 | 0 | — |
case-23 | pass→pass | 23,788 | 24,366 | +2% | 1 | 1 | 0% | 4,491 | 5,159 | +15% | 0 | 0 | — |
case-24 | pass→pass | 22,220 | 21,510 | -3% | 1 | 1 | 0% | 3,246 | 3,290 | +1% | 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 +17 percentage points is the difference between those two pass rates over the 24 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.