Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →用于信托领域合同与法务中的合同版本对比助手-补充协议版场景。支持结构化输入处理、规则分析与Markdown结果输出。
.claude/skills/aifinlab-contract-version-compare-supplemental-agreement/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -43% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -41% | 0% |
本技能面向“合同与法务 / 版本对比”业务单元，输出结构化分析报告与风险提示，支持尽调、法务、风控及存续管理流程中的基础判断。
id, name, text, status, timestamp 及相关业务字段bashpython scripts/t817_analysis.py --input input.json --output report.md --title "合同版本对比助手-补充协议版（T817）"
python# 调用 skill result = run_skill({ "param1": "value1", "param2": "value2" })
bashpython scripts/run_skill.py --input data.json
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 11,839 | 10,003 | -16% | 1 | 1 | 0% | 2,417 | 1,145 | -53% | 0 | 0 | — |
case-02 | fail→pass | 12,837 | 5,048 | -61% | 1 | 1 | 0% | 2,287 | 1,311 | -43% | 0 | 0 | — |
case-03 | fail→pass | 12,045 | 2,755 | -77% | 1 | 1 | 0% | 1,863 | 892 | -52% | 0 | 0 | — |
case-04 | pass→pass | 12,657 | 2,651 | -79% | 1 | 1 | 0% | 2,011 | 845 | -58% | 0 | 0 | — |
case-05 | pass→pass | 8,555 | 3,301 | -61% | 1 | 1 | 0% | 1,521 | 1,075 | -29% | 0 | 0 | — |
case-06 | pass→pass | 12,272 | 8,274 | -33% | 1 | 1 | 0% | 1,774 | 1,782 | +0% | 0 | 0 | — |
case-07 | pass→pass | 14,570 | 5,400 | -63% | 1 | 1 | 0% | 2,353 | 1,301 | -45% | 0 | 0 | — |
case-08 | pass→pass | 14,210 | 6,533 | -54% | 1 | 1 | 0% | 2,187 | 1,517 | -31% | 0 | 0 | — |
case-09 | pass→pass | 11,133 | 7,479 | -33% | 1 | 1 | 0% | 1,787 | 1,659 | -7% | 0 | 0 | — |
case-10 | pass→pass | 13,188 | 14,035 | +6% | 1 | 1 | 0% | 2,097 | 2,512 | +20% | 0 | 0 | — |
case-11 | pass→pass | 13,421 | 5,602 | -58% | 1 | 1 | 0% | 2,145 | 1,228 | -43% | 0 | 0 | — |
case-12 | fail→pass | 10,601 | 2,406 | -77% | 1 | 1 | 0% | 1,502 | 824 | -45% | 0 | 0 | — |
case-13 | fail→pass | 11,141 | 2,774 | -75% | 1 | 1 | 0% | 1,523 | 891 | -41% | 0 | 0 | — |
case-14 | fail→pass | 10,711 | 3,673 | -66% | 1 | 1 | 0% | 1,654 | 1,075 | -35% | 0 | 0 | — |
case-15 | fail→pass | 14,441 | 3,122 | -78% | 1 | 1 | 0% | 2,325 | 976 | -58% | 0 | 0 | — |
case-16 | pass→pass | 16,447 | 10,052 | -39% | 1 | 1 | 0% | 2,448 | 2,077 | -15% | 0 | 0 | — |
case-17 | pass→pass | 9,328 | 2,044 | -78% | 1 | 1 | 0% | 1,634 | 742 | -55% | 0 | 0 | — |
case-18 | fail→pass | 9,347 | 2,643 | -72% | 1 | 1 | 0% | 1,458 | 830 | -43% | 0 | 0 | — |
case-19 | fail→pass | 26,809 | 2,843 | -89% | 1 | 1 | 0% | 5,050 | 745 | -85% | 0 | 0 | — |
case-20 | pass→pass | 20,798 | 2,201 | -89% | 1 | 1 | 0% | 3,903 | 772 | -80% | 0 | 0 | — |
case-21 | pass→pass | 7,603 | 8,797 | +16% | 1 | 1 | 0% | 1,303 | 1,889 | +45% | 0 | 0 | — |
case-22 | pass→pass | 11,287 | 12,013 | +6% | 1 | 1 | 0% | 1,777 | 2,197 | +24% | 0 | 0 | — |
case-23 | fail→pass | 15,249 | 13,626 | -11% | 1 | 1 | 0% | 2,373 | 2,667 | +12% | 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 +43 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.