Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当用户需要在银行合规运营场景下，对客服录音或沟通记录进行结构化摘要、风险要点提炼与后续动作整理时使用本技能。
.claude/skills/aifinlab-bank-t231-compliance-operations-call-summary-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-22 | ✓→✗ | ▼ Worse | 15% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 6% | 0% |
本技能用于银行合规运营中对录音或沟通记录进行合规可用的摘要输出，强调保留关键事实、风险提示和待办事项，便于质检留痕与后续跟踪。
scripts/call_summary_builder.py：对录音转写文本进行分段提取、关键词识别与摘要结构化输出。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-22 | pass→fail | 10,261 | 9,558 | -7% | 1 | 1 | 0% | 1,643 | 1,893 | +15% | 0 | 0 | — |
case-01 | pass→pass | 15,767 | 12,355 | -22% | 1 | 1 | 0% | 2,041 | 2,158 | +6% | 0 | 0 | — |
case-02 | fail→pass | 13,316 | 15,087 | +13% | 1 | 1 | 0% | 1,626 | 1,924 | +18% | 0 | 0 | — |
case-03 | pass→pass | 6,340 | 7,129 | +12% | 1 | 1 | 0% | 778 | 1,476 | +90% | 0 | 0 | — |
case-04 | pass→pass | 18,447 | 12,363 | -33% | 1 | 1 | 0% | 2,394 | 2,041 | -15% | 0 | 0 | — |
case-05 | fail→fail | 24,499 | 15,584 | -36% | 1 | 1 | 0% | 3,045 | 2,443 | -20% | 0 | 0 | — |
case-06 | fail→pass | 23,748 | 12,812 | -46% | 1 | 1 | 0% | 2,974 | 2,138 | -28% | 0 | 0 | — |
case-07 | pass→pass | 14,977 | 14,021 | -6% | 1 | 1 | 0% | 2,137 | 2,348 | +10% | 0 | 0 | — |
case-08 | pass→pass | 9,983 | 10,537 | +6% | 1 | 1 | 0% | 1,299 | 2,200 | +69% | 0 | 0 | — |
case-09 | pass→pass | 7,719 | 5,341 | -31% | 1 | 1 | 0% | 987 | 1,392 | +41% | 0 | 0 | — |
case-10 | pass→pass | 13,882 | 11,122 | -20% | 1 | 1 | 0% | 1,740 | 2,190 | +26% | 0 | 0 | — |
case-11 | fail→pass | 12,799 | 11,598 | -9% | 1 | 1 | 0% | 1,614 | 2,039 | +26% | 0 | 0 | — |
case-12 | pass→pass | 4,989 | 6,541 | +31% | 1 | 1 | 0% | 655 | 1,390 | +112% | 0 | 0 | — |
case-13 | pass→pass | 17,956 | 18,862 | +5% | 1 | 1 | 0% | 2,431 | 2,898 | +19% | 0 | 0 | — |
case-14 | pass→pass | 9,000 | 8,828 | -2% | 1 | 1 | 0% | 1,140 | 1,816 | +59% | 0 | 0 | — |
case-15 | pass→pass | 14,096 | 13,174 | -7% | 1 | 1 | 0% | 2,080 | 2,166 | +4% | 0 | 0 | — |
case-16 | pass→pass | 13,700 | 13,005 | -5% | 1 | 1 | 0% | 2,077 | 2,188 | +5% | 0 | 0 | — |
case-17 | pass→pass | 13,897 | 10,837 | -22% | 1 | 1 | 0% | 1,779 | 2,269 | +28% | 0 | 0 | — |
case-18 | pass→pass | 7,850 | 10,835 | +38% | 1 | 1 | 0% | 1,314 | 2,179 | +66% | 0 | 0 | — |
case-19 | pass→pass | 18,016 | 12,198 | -32% | 1 | 1 | 0% | 1,646 | 2,263 | +37% | 0 | 0 | — |
case-20 | pass→pass | 9,204 | 10,596 | +15% | 1 | 1 | 0% | 1,319 | 1,935 | +47% | 0 | 0 | — |
case-21 | pass→pass | 9,466 | 9,502 | +0% | 1 | 1 | 0% | 1,292 | 1,901 | +47% | 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 +9 percentage points is the difference between those two pass rates over the 22 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.