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，对内控或质检缺陷进行跟踪、分级、整改进度汇总时使用本技能。
.claude/skills/aifinlab-bank-t233-compliance-operations-defect-tracking-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 57% | 0% |
本技能用于合规运营缺陷跟踪场景，帮助把缺陷清单、整改进度与风险等级形成结构化输出，便于闭环管理。
scripts/defect_tracking_manager.py：对缺陷台账进行分级、超期识别与整改进度汇总。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | pass→pass | 6,877 | 8,974 | +30% | 1 | 1 | 0% | 899 | 1,521 | +69% | 0 | 0 | — |
case-03 | fail→pass | 14,602 | 15,063 | +3% | 1 | 1 | 0% | 2,147 | 1,790 | -17% | 0 | 0 | — |
case-01 | fail→fail | 17,611 | 22,356 | +27% | 1 | 1 | 0% | 2,588 | 2,874 | +11% | 0 | 0 | — |
case-04 | pass→pass | 21,299 | 16,097 | -24% | 1 | 1 | 0% | 2,555 | 2,531 | -1% | 0 | 0 | — |
case-05 | pass→pass | 17,267 | 14,330 | -17% | 1 | 1 | 0% | 2,582 | 2,894 | +12% | 0 | 0 | — |
case-06 | pass→pass | 19,961 | 17,559 | -12% | 1 | 1 | 0% | 2,656 | 3,143 | +18% | 0 | 0 | — |
case-07 | pass→pass | 22,492 | 20,892 | -7% | 1 | 1 | 0% | 3,333 | 2,996 | -10% | 0 | 0 | — |
case-08 | pass→pass | 10,661 | 10,450 | -2% | 1 | 1 | 0% | 1,581 | 2,246 | +42% | 0 | 0 | — |
case-09 | fail→pass | 14,756 | 11,969 | -19% | 1 | 1 | 0% | 1,971 | 2,126 | +8% | 0 | 0 | — |
case-10 | fail→pass | 17,139 | 14,088 | -18% | 1 | 1 | 0% | 2,213 | 2,481 | +12% | 0 | 0 | — |
case-11 | fail→pass | 19,713 | 3,124 | -84% | 1 | 1 | 0% | 2,738 | 819 | -70% | 0 | 0 | — |
case-12 | pass→pass | 15,193 | 12,452 | -18% | 1 | 1 | 0% | 2,122 | 2,256 | +6% | 0 | 0 | — |
case-13 | pass→pass | 12,160 | 15,175 | +25% | 1 | 1 | 0% | 1,952 | 2,242 | +15% | 0 | 0 | — |
case-14 | fail→pass | 9,237 | 11,839 | +28% | 1 | 1 | 0% | 1,238 | 1,947 | +57% | 0 | 0 | — |
case-15 | fail→fail | 25,631 | 10,966 | -57% | 1 | 1 | 0% | 1,698 | 2,052 | +21% | 0 | 0 | — |
case-16 | pass→pass | 15,810 | 7,370 | -53% | 1 | 1 | 0% | 2,224 | 1,548 | -30% | 0 | 0 | — |
case-17 | fail→pass | 16,032 | 15,006 | -6% | 1 | 1 | 0% | 2,331 | 2,253 | -3% | 0 | 0 | — |
case-18 | pass→pass | 21,096 | 18,020 | -15% | 1 | 1 | 0% | 3,062 | 3,042 | -1% | 0 | 0 | — |
case-19 | pass→pass | 17,808 | 13,520 | -24% | 1 | 1 | 0% | 2,554 | 2,106 | -18% | 0 | 0 | — |
case-20 | pass→pass | 31,744 | 16,940 | -47% | 1 | 1 | 0% | 2,763 | 2,949 | +7% | 0 | 0 | — |
case-21 | pass→pass | 15,858 | 16,591 | +5% | 1 | 1 | 0% | 2,055 | 3,011 | +47% | 0 | 0 | — |
case-22 | pass→pass | 21,204 | 15,204 | -28% | 1 | 1 | 0% | 3,291 | 2,852 | -13% | 0 | 0 | — |
case-23 | pass→pass | 15,033 | 13,903 | -8% | 1 | 1 | 0% | 2,076 | 2,382 | +15% | 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. 23 cases were attempted. The headline lift of +26 percentage points is the difference between those two pass rates over the 23 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.