Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户需要对重疾险理赔申请资料进行结构化受理检查时使用本 skill。它用于识别理赔申请书、身份证明、银行卡信息、诊断证明、出院记录、住院病历、病理报告、手术记录、影像报告、专项检查报告等与重大疾病责任认定相关的关键材料是否齐全、清晰、有效、相互一致，并输出补件清单、受理前检查结果和跟进建议。
.claude/skills/aifinlab-claims-material-check-critical-illness-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-15 | ✗→✓ | ▲ Improved | 135% | 0% |
| case-26 | ✗→✓ | ▲ Improved | 89% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 55% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 24% | 0% |
你是“理赔材料检查助手-重疾险版”。
你的任务是对重疾险理赔申请资料进行专业、结构化的检查与核验，重点识别理赔申请中与重大疾病责任认定相关的关键材料是否齐全、清晰、有效、相互一致，并判断这些材料是否基本满足受理与后续审核要求，帮助理赔受理人员、运营支持人员和审核协同人员快速发现缺失项、无效项、模糊项和需补充事项，降低重疾险理赔补件率并提升材料初审效率。
本技能只用于重疾险理赔材料的完整性检查、问题识别和补件建议，不替代正式理赔审核决定，不替代重疾定义解释、责任认定结论、赔付结论、拒赔结论或法律意见。凡涉及关键病理或诊断页缺失、疾病分期分级依据不足、手术记录不完整、后遗状态资料不足或材料归属不清的情形，必须明确标注“需进一步核实”或“建议补查”。
claims-material-check：通用理赔材料检查，适合跨险种基础核对claims-material-check-medical-insurance-assistant：医疗险报销型材料检查，重点在病历、发票、费用清单和结算资料claims-case-summary：侧重案件过程摘要，不替代本技能的受理前材料核验当用户表达以下意图时，使用本技能：
以下情况不按本技能直接处理，应提醒用户这是更细分任务：
围绕“判断重疾险理赔材料是否具备受理与责任审核基础”完成以下输出：
除非用户另有要求，严格按以下顺序输出：
遇到以下情况时，按保守口径输出：
输出结果应让理赔受理人员能够在较短时间内回答以下问题：
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-15 | fail→pass | 14,857 | 20,946 | +41% | 1 | 1 | 0% | 2,273 | 5,352 | +135% | 0 | 0 | — |
case-26 | fail→pass | 18,907 | 22,684 | +20% | 1 | 1 | 0% | 2,774 | 5,248 | +89% | 0 | 0 | — |
case-01 | fail→fail | 25,071 | 34,902 | +39% | 1 | 1 | 0% | 3,692 | 6,971 | +89% | 0 | 0 | — |
case-02 | fail→pass | 21,874 | 20,706 | -5% | 1 | 1 | 0% | 3,185 | 4,925 | +55% | 0 | 0 | — |
case-03 | fail→pass | 17,455 | 6,667 | -62% | 1 | 1 | 0% | 2,631 | 2,778 | +6% | 0 | 0 | — |
case-04 | fail→pass | 16,820 | 8,351 | -50% | 1 | 1 | 0% | 2,491 | 3,099 | +24% | 0 | 0 | — |
case-05 | fail→fail | 21,203 | 13,794 | -35% | 1 | 1 | 0% | 3,282 | 4,044 | +23% | 0 | 0 | — |
case-06 | fail→pass | 16,935 | 20,789 | +23% | 1 | 1 | 0% | 2,664 | 5,194 | +95% | 0 | 0 | — |
case-07 | fail→pass | 19,244 | 22,288 | +16% | 1 | 1 | 0% | 2,778 | 5,399 | +94% | 0 | 0 | — |
case-08 | fail→fail | 14,784 | 23,435 | +59% | 1 | 1 | 0% | 2,232 | 5,598 | +151% | 0 | 0 | — |
case-09 | pass→pass | 20,624 | 23,044 | +12% | 1 | 1 | 0% | 2,913 | 5,609 | +93% | 0 | 0 | — |
case-10 | fail→pass | 15,333 | 21,271 | +39% | 1 | 1 | 0% | 2,201 | 5,265 | +139% | 0 | 0 | — |
case-11 | fail→pass | 17,228 | 24,061 | +40% | 1 | 1 | 0% | 2,516 | 5,737 | +128% | 0 | 0 | — |
case-12 | pass→pass | 13,349 | 20,266 | +52% | 1 | 1 | 0% | 2,007 | 4,954 | +147% | 0 | 0 | — |
case-13 | fail→pass | 17,953 | 24,511 | +37% | 1 | 1 | 0% | 2,763 | 5,866 | +112% | 0 | 0 | — |
case-14 | fail→pass | 19,200 | 26,206 | +36% | 1 | 1 | 0% | 2,910 | 6,180 | +112% | 0 | 0 | — |
case-16 | fail→fail | 13,445 | 21,216 | +58% | 1 | 1 | 0% | 2,044 | 5,222 | +155% | 0 | 0 | — |
case-17 | pass→pass | 13,840 | 24,317 | +76% | 1 | 1 | 0% | 2,143 | 5,592 | +161% | 0 | 0 | — |
case-18 | pass→pass | 13,316 | 23,962 | +80% | 1 | 1 | 0% | 2,033 | 5,759 | +183% | 0 | 0 | — |
case-19 | pass→pass | 13,077 | 18,911 | +45% | 1 | 1 | 0% | 1,985 | 4,912 | +147% | 0 | 0 | — |
case-20 | pass→pass | 15,399 | 18,989 | +23% | 1 | 1 | 0% | 2,051 | 4,741 | +131% | 0 | 0 | — |
case-21 | fail→pass | 16,067 | 23,208 | +44% | 1 | 1 | 0% | 2,340 | 5,527 | +136% | 0 | 0 | — |
case-22 | fail→pass | 22,945 | 24,332 | +6% | 1 | 1 | 0% | 3,412 | 5,630 | +65% | 0 | 0 | — |
case-23 | pass→pass | 17,656 | 23,998 | +36% | 1 | 1 | 0% | 2,541 | 5,638 | +122% | 0 | 0 | — |
case-24 | fail→fail | 16,928 | 23,753 | +40% | 1 | 1 | 0% | 2,446 | 5,566 | +128% | 0 | 0 | — |
case-25 | fail→pass | 13,283 | 26,179 | +97% | 1 | 1 | 0% | 2,046 | 6,271 | +207% | 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. 26 cases were attempted. The headline lift of +54 percentage points is the difference between those two pass rates over the 26 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.