Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →当需要将客户沟通或营销材料改写为合规表述并输出话术主线、禁用提醒与跟进建议时使用本技能，适用于合规改写与话术优化场景。
.claude/skills/aifinlab-bank-t227-compliance-operations-compliant-rewrite-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-15 | ✗→✓ | ▲ Improved | 45% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 61% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -68% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 53% | 0% |
本技能用于把业务或营销表达改写为合规、克制且可留痕的表述，输出主线结构、禁用提示与后续跟进建议，避免夸大收益或误导性描述。
scripts/compliant_rewrite_builder.py| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→fail | 11,663 | 14,998 | +29% | 1 | 1 | 0% | 1,809 | 2,400 | +33% | 0 | 0 | — |
case-15 | fail→pass | 8,602 | 10,512 | +22% | 1 | 1 | 0% | 1,377 | 2,003 | +45% | 0 | 0 | — |
case-01 | fail→pass | 9,984 | 13,638 | +37% | 1 | 1 | 0% | 1,533 | 2,474 | +61% | 0 | 0 | — |
case-02 | fail→pass | 13,814 | 10,762 | -22% | 1 | 1 | 0% | 1,867 | 2,362 | +27% | 0 | 0 | — |
case-03 | pass→pass | 10,708 | 17,145 | +60% | 1 | 1 | 0% | 1,650 | 2,438 | +48% | 0 | 0 | — |
case-05 | fail→pass | 12,408 | 2,825 | -77% | 1 | 1 | 0% | 2,207 | 712 | -68% | 0 | 0 | — |
case-06 | fail→pass | 8,980 | 15,541 | +73% | 1 | 1 | 0% | 1,516 | 2,315 | +53% | 0 | 0 | — |
case-07 | fail→pass | 11,954 | 15,521 | +30% | 1 | 1 | 0% | 1,487 | 2,423 | +63% | 0 | 0 | — |
case-08 | fail→pass | 12,177 | 11,808 | -3% | 1 | 1 | 0% | 1,511 | 2,200 | +46% | 0 | 0 | — |
case-09 | fail→pass | 10,636 | 11,835 | +11% | 1 | 1 | 0% | 1,337 | 2,193 | +64% | 0 | 0 | — |
case-10 | fail→pass | 9,553 | 12,174 | +27% | 1 | 1 | 0% | 1,415 | 2,057 | +45% | 0 | 0 | — |
case-11 | fail→pass | 8,927 | 22,725 | +155% | 1 | 1 | 0% | 1,408 | 2,478 | +76% | 0 | 0 | — |
case-12 | pass→pass | 8,972 | 10,727 | +20% | 1 | 1 | 0% | 1,407 | 2,167 | +54% | 0 | 0 | — |
case-13 | fail→pass | 11,856 | 12,913 | +9% | 1 | 1 | 0% | 1,439 | 2,205 | +53% | 0 | 0 | — |
case-14 | fail→fail | 12,928 | 18,136 | +40% | 1 | 1 | 0% | 1,707 | 2,780 | +63% | 0 | 0 | — |
case-16 | pass→pass | 10,895 | 14,922 | +37% | 1 | 1 | 0% | 1,488 | 2,712 | +82% | 0 | 0 | — |
case-17 | fail→pass | 7,479 | 12,785 | +71% | 1 | 1 | 0% | 1,305 | 2,315 | +77% | 0 | 0 | — |
case-18 | fail→pass | 10,053 | 15,825 | +57% | 1 | 1 | 0% | 1,357 | 2,728 | +101% | 0 | 0 | — |
case-19 | fail→pass | 8,676 | 14,882 | +72% | 1 | 1 | 0% | 1,330 | 2,261 | +70% | 0 | 0 | — |
case-20 | pass→pass | 8,030 | 7,723 | -4% | 1 | 1 | 0% | 1,128 | 1,392 | +23% | 0 | 0 | — |
case-21 | pass→pass | 10,096 | 12,985 | +29% | 1 | 1 | 0% | 1,361 | 2,390 | +76% | 0 | 0 | — |
case-22 | pass→pass | 29,401 | 20,805 | -29% | 1 | 1 | 0% | 4,621 | 3,457 | -25% | 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 +64 percentage points is the difference between those two pass rates over the 22 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.