Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →生成 IPO 尽职调查问题清单，覆盖业务、财务、法律等各领域
.claude/skills/aifinlab-ipo/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 50% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-06 | ✓→✗ | ▼ Worse | 27% | 0% |
本技能用于生成 IPO 尽职调查问题清单，覆盖业务、财务、法律、公司治理等各领域，帮助系统性地开展尽职调查工作。
根据项目情况选择需要尽调的领域
bashcd "C:\Users\51994\Desktop\skills\证券 2\14. 尽调问题清单助手-IPO 版" python scripts/generate_dd_questions.py --industry <行业> --output <输出路径>
问题清单将按领域分类，包含具体问题
根据问题清单开展尽职调查，收集资料并核实
markdown## IPO 尽职调查问题清单 ### 一、公司基本情况 #### 1. 公司设立及历史沿革 1.1 请说明公司设立背景及设立过程 1.2 请提供公司历次股权变更的工商档案 1.3 请说明是否存在重大资产重组情况 ... ### 二、业务与技术 #### 3. 主营业务 3.1 请说明公司主营业务及产品/服务 3.2 请说明公司业务模式及盈利模式 3.3 请说明公司核心竞争力 ... ### 三、财务会计 ... ### 四、公司治理 ... ### 五、法律合规 ... ### 六、募投项目 ...
scripts/generate_dd_questions.py - 问题清单生成脚本references/dd_checklist_template.md - 尽调清单模板| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 61,824 | 64,263 | +4% | 1 | 1 | 0% | 4,261 | 6,335 | +49% | 0 | 0 | — |
case-02 | fail→fail | 27,460 | 58,263 | +112% | 1 | 1 | 0% | 3,509 | 5,384 | +53% | 0 | 0 | — |
case-03 | fail→fail | 27,698 | 37,208 | +34% | 1 | 1 | 0% | 3,683 | 6,697 | +82% | 0 | 0 | — |
case-04 | fail→pass | 11,787 | 33,769 | +186% | 1 | 1 | 0% | 1,808 | 2,193 | +21% | 0 | 0 | — |
case-05 | pass→pass | 23,901 | 54,295 | +127% | 1 | 1 | 0% | 3,400 | 5,134 | +51% | 0 | 0 | — |
case-06 | pass→fail | 23,190 | 16,795 | -28% | 1 | 1 | 0% | 3,238 | 4,119 | +27% | 0 | 0 | — |
case-07 | fail→fail | 25,497 | 28,251 | +11% | 1 | 1 | 0% | 3,404 | 5,348 | +57% | 0 | 0 | — |
case-08 | pass→pass | 24,872 | 22,205 | -11% | 1 | 1 | 0% | 3,695 | 4,626 | +25% | 0 | 0 | — |
case-09 | pass→pass | 18,673 | 10,534 | -44% | 1 | 1 | 0% | 2,607 | 3,162 | +21% | 0 | 0 | — |
case-10 | fail→pass | 18,481 | 16,861 | -9% | 1 | 1 | 0% | 2,563 | 3,837 | +50% | 0 | 0 | — |
case-11 | fail→pass | 19,327 | 15,574 | -19% | 1 | 1 | 0% | 2,936 | 3,949 | +35% | 0 | 0 | — |
case-12 | pass→pass | 20,701 | 21,611 | +4% | 1 | 1 | 0% | 2,897 | 4,619 | +59% | 0 | 0 | — |
case-13 | fail→pass | 21,508 | 23,208 | +8% | 1 | 1 | 0% | 2,788 | 4,602 | +65% | 0 | 0 | — |
case-14 | pass→pass | 29,170 | 23,676 | -19% | 1 | 1 | 0% | 4,095 | 5,021 | +23% | 0 | 0 | — |
case-15 | pass→pass | 22,576 | 21,419 | -5% | 1 | 1 | 0% | 3,223 | 4,539 | +41% | 0 | 0 | — |
case-16 | pass→pass | 22,872 | 21,912 | -4% | 1 | 1 | 0% | 3,413 | 4,782 | +40% | 0 | 0 | — |
case-17 | pass→pass | 25,539 | 24,220 | -5% | 1 | 1 | 0% | 3,530 | 4,992 | +41% | 0 | 0 | — |
case-18 | pass→pass | 27,259 | 23,489 | -14% | 1 | 1 | 0% | 3,593 | 4,673 | +30% | 0 | 0 | — |
case-19 | pass→pass | 19,785 | 20,835 | +5% | 1 | 1 | 0% | 2,728 | 4,690 | +72% | 0 | 0 | — |
case-20 | pass→pass | 11,462 | 10,763 | -6% | 1 | 1 | 0% | 1,759 | 2,992 | +70% | 0 | 0 | — |
case-21 | pass→pass | 27,304 | 41,875 | +53% | 1 | 1 | 0% | 3,609 | 5,935 | +64% | 0 | 0 | — |
case-22 | pass→pass | 25,342 | 22,686 | -10% | 1 | 1 | 0% | 4,020 | 5,545 | +38% | 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 +14 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.