Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →用于合同条款结构化识别与完整性检查，自动识别关键条款类型并提示缺失项。适用于信托业务合同起草、条款体检、法务预审场景。
.claude/skills/aifinlab-contract-clause-assistant/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-11 | ✗→✓ | ▲ Improved | -29% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -43% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 29% | 0% |
本技能用于对合同条款进行基础解析与完整性检查，重点覆盖：
支持 JSON 数组或 JSONL，每条记录建议包含：
contract_nameclause_noclause_titleclause_text使用 scripts/contract_clause_analysis.py 生成报告：
bashpython scripts/contract_clause_analysis.py \ --input clauses.json \ --output clause_report.md
可选参数：
--required：自定义关键条款清单 JSON| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-11 | fail→pass | 12,643 | 6,626 | -48% | 1 | 1 | 0% | 1,833 | 1,299 | -29% | 0 | 0 | — |
case-12 | pass→pass | 8,211 | 3,026 | -63% | 1 | 1 | 0% | 1,329 | 884 | -33% | 0 | 0 | — |
case-04 | fail→fail | 27,669 | 36,112 | +31% | 1 | 1 | 0% | 4,358 | 6,302 | +45% | 0 | 0 | — |
case-01 | fail→fail | 22,996 | 41,282 | +80% | 1 | 1 | 0% | 3,956 | 4,728 | +20% | 0 | 0 | — |
case-02 | fail→fail | 17,393 | 22,413 | +29% | 1 | 1 | 0% | 2,698 | 4,066 | +51% | 0 | 0 | — |
case-03 | fail→fail | 13,025 | 21,001 | +61% | 1 | 1 | 0% | 1,946 | 3,664 | +88% | 0 | 0 | — |
case-05 | fail→fail | 21,285 | 22,848 | +7% | 1 | 1 | 0% | 3,129 | 3,771 | +21% | 0 | 0 | — |
case-06 | fail→fail | 29,347 | 35,462 | +21% | 1 | 1 | 0% | 5,494 | 7,150 | +30% | 0 | 0 | — |
case-07 | fail→pass | 10,881 | 2,293 | -79% | 1 | 1 | 0% | 1,844 | 693 | -62% | 0 | 0 | — |
case-08 | fail→pass | 8,788 | 2,785 | -68% | 1 | 1 | 0% | 1,394 | 799 | -43% | 0 | 0 | — |
case-09 | fail→pass | 16,523 | 6,164 | -63% | 1 | 1 | 0% | 2,798 | 1,341 | -52% | 0 | 0 | — |
case-10 | pass→pass | 16,360 | 8,844 | -46% | 1 | 1 | 0% | 2,517 | 1,800 | -28% | 0 | 0 | — |
case-13 | pass→pass | 8,423 | 6,077 | -28% | 1 | 1 | 0% | 1,361 | 1,353 | -1% | 0 | 0 | — |
case-14 | fail→pass | 8,396 | 7,917 | -6% | 1 | 1 | 0% | 1,271 | 1,645 | +29% | 0 | 0 | — |
case-15 | pass→pass | 7,215 | 3,939 | -45% | 1 | 1 | 0% | 1,189 | 1,037 | -13% | 0 | 0 | — |
case-16 | fail→pass | 11,698 | 3,827 | -67% | 1 | 1 | 0% | 1,730 | 905 | -48% | 0 | 0 | — |
case-17 | fail→pass | 14,999 | 9,313 | -38% | 1 | 1 | 0% | 2,213 | 1,835 | -17% | 0 | 0 | — |
case-18 | pass→pass | 14,971 | 4,679 | -69% | 1 | 1 | 0% | 2,311 | 1,132 | -51% | 0 | 0 | — |
case-19 | fail→pass | 23,262 | 6,846 | -71% | 1 | 1 | 0% | 4,334 | 1,632 | -62% | 0 | 0 | — |
case-20 | fail→pass | 13,410 | 2,385 | -82% | 1 | 1 | 0% | 2,072 | 707 | -66% | 0 | 0 | — |
case-21 | fail→fail | 10,255 | 9,769 | -5% | 1 | 1 | 0% | 1,719 | 1,924 | +12% | 0 | 0 | — |
case-22 | fail→pass | 14,940 | 5,077 | -66% | 1 | 1 | 0% | 2,207 | 1,066 | -52% | 0 | 0 | — |
case-23 | pass→pass | 19,067 | 18,377 | -4% | 1 | 1 | 0% | 2,757 | 3,064 | +11% | 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.