Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股财务健康度诊断/爆雷预警/财务风险排查。当用户说"财务健康"、"财务质量"、"会不会爆雷"、"财务诊断"、"financial health"、"F-score"、"Z-score"、"XX的财务怎么样"、"财报有没有问题"、"应收账款"、"商誉减值"、"现金流质量"、"财务健康度"、"爆雷风险"、"财务风险"时触发。MUST USE when user asks about financial health diagnosis, bankruptcy/default risk, F-score/Z-score analysis, or whether a company might have accounting issues. 通过多维度财务指标（盈利/偿债/营运/现金流）综合诊断企业财务健康程度，计算 Piotroski F-Score 和改良 Z-Score，识别潜在爆雷风险。支持研报风格（formal）和快速诊断风格（brief）。
.claude/skills/aifinlab-a-share-financial-health/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 40% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 13% | 0% |
python "$SKILLS_ROOT/cn-stock-data/scripts/cn_stock_data.py" finance --code [CODE]Piotroski F-Score（0-9分）：对照 references/financial-health-guide.md 中的9项标准逐项打分
改良 Z-Score：使用A股适配公式计算
对照参考文档中的爆雷信号清单逐项排查：
每项标注：✅ 健康 / ⚠️ 关注 / 🔴 危险
formal 模式：
brief 模式（默认）：
| F-Score | Z-Score | 红旗数 | 等级 | |---------|---------|--------|------| | 7-9 | > 2.99 | 0-1 | 🟢 财务健康 | | 5-6 | 1.8-2.99 | 2-3 | 🟡 亚健康/需关注 | | 3-4 | < 1.8 | 4-5 | 🔴 高风险 | | 0-2 | < 1.2 | 6+ | ⛔ 极高风险/爆雷预警 |
综合三个维度取最差等级，任一维度触发高风险即整体标记高风险。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-25 | pass→pass | 13,360 | 2,768 | -79% | 1 | 1 | 0% | 1,756 | 1,635 | -7% | 0 | 0 | — |
case-01 | fail→fail | 34,736 | 13,355 | -62% | 1 | 1 | 0% | 4,569 | 2,329 | -49% | 0 | 0 | — |
case-02 | fail→fail | 10,300 | 8,170 | -21% | 1 | 1 | 0% | 1,557 | 1,629 | +5% | 0 | 0 | — |
case-03 | fail→fail | 34,053 | 29,778 | -13% | 1 | 1 | 0% | 4,545 | 1,973 | -57% | 0 | 0 | — |
case-04 | fail→fail | 20,589 | 29,404 | +43% | 1 | 1 | 0% | 3,105 | 5,414 | +74% | 0 | 0 | — |
case-05 | fail→fail | 29,342 | 10,434 | -64% | 1 | 1 | 0% | 5,820 | 1,628 | -72% | 0 | 0 | — |
case-06 | fail→fail | 24,127 | 19,952 | -17% | 1 | 1 | 0% | 3,137 | 4,195 | +34% | 0 | 0 | — |
case-07 | pass→pass | 15,700 | 6,946 | -56% | 1 | 1 | 0% | 2,361 | 2,358 | -0% | 0 | 0 | — |
case-08 | fail→pass | 17,770 | 9,557 | -46% | 1 | 1 | 0% | 2,227 | 2,707 | +22% | 0 | 0 | — |
case-09 | pass→pass | 11,061 | 6,615 | -40% | 1 | 1 | 0% | 1,606 | 2,230 | +39% | 0 | 0 | — |
case-10 | fail→pass | 14,228 | 5,153 | -64% | 1 | 1 | 0% | 1,813 | 2,041 | +13% | 0 | 0 | — |
case-11 | pass→pass | 15,558 | 9,648 | -38% | 1 | 1 | 0% | 2,100 | 2,443 | +16% | 0 | 0 | — |
case-12 | fail→pass | 10,637 | 5,708 | -46% | 1 | 1 | 0% | 1,408 | 1,965 | +40% | 0 | 0 | — |
case-13 | pass→pass | 8,926 | 4,073 | -54% | 1 | 1 | 0% | 1,456 | 1,729 | +19% | 0 | 0 | — |
case-14 | pass→pass | 7,421 | 5,009 | -33% | 1 | 1 | 0% | 1,065 | 2,086 | +96% | 0 | 0 | — |
case-15 | pass→pass | 47,889 | 13,270 | -72% | 1 | 1 | 0% | 3,057 | 3,440 | +13% | 0 | 0 | — |
case-16 | fail→pass | 16,766 | 3,311 | -80% | 1 | 1 | 0% | 2,103 | 1,665 | -21% | 0 | 0 | — |
case-17 | pass→pass | 13,464 | 4,573 | -66% | 1 | 1 | 0% | 1,991 | 1,993 | +0% | 0 | 0 | — |
case-18 | pass→pass | 8,489 | 10,930 | +29% | 1 | 1 | 0% | 1,427 | 2,788 | +95% | 0 | 0 | — |
case-19 | fail→pass | 15,282 | 9,021 | -41% | 1 | 1 | 0% | 2,211 | 2,503 | +13% | 0 | 0 | — |
case-20 | pass→pass | 13,455 | 13,400 | -0% | 1 | 1 | 0% | 1,917 | 3,064 | +60% | 0 | 0 | — |
case-21 | fail→fail | 9,325 | 2,860 | -69% | 1 | 1 | 0% | 1,300 | 1,585 | +22% | 0 | 0 | — |
case-22 | fail→pass | 16,296 | 9,235 | -43% | 1 | 1 | 0% | 2,199 | 2,732 | +24% | 0 | 0 | — |
case-23 | fail→pass | 9,053 | 2,500 | -72% | 1 | 1 | 0% | 1,262 | 1,574 | +25% | 0 | 0 | — |
case-24 | pass→pass | 13,494 | 6,981 | -48% | 1 | 1 | 0% | 1,713 | 2,365 | +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. 25 cases were attempted, and 20 counted toward the lift figure. The other 5 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +28 percentage points is the difference between those two pass rates over the 20 comparable cases. 2 cases got worse with the skill loaded, and they are 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.