Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股板块价差/行业估值差分析。当用户说"板块价差"、"sector spread"、"行业估值差"、"板块分化"、"分化有多大"时触发。量化分析板块间价差和估值差异。支持formal和brief风格。
.claude/skills/aifinlab-a-share-sector-spread/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 41% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 1% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -35% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -3% | 0% |
bashSCRIPTS="$SKILLS_ROOT/cn-stock-data/scripts" python "$SCRIPTS/cn_stock_data.py" kline --code [CODE] --freq daily --start [日期] python "$SCRIPTS/cn_stock_data.py" quote --code [CODE] python "$SCRIPTS/cn_stock_data.py" finance --code [CODE]
过度分化→可能反转
| 维度 | formal | brief | |------|--------|-------| | 价差 | 完整行业价差矩阵 | 最大价差 | | 分化度 | 离散度时序 | 当前分化水平 | | 回归信号 | 历史分位分析 | 是否过度分化 | 默认风格：brief。
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-20 | fail→fail | 9,588 | 3,290 | -66% | 1 | 1 | 0% | 1,205 | 836 | -31% | 0 | 0 | — |
case-01 | fail→pass | 11,157 | 18,881 | +69% | 1 | 1 | 0% | 1,565 | 2,206 | +41% | 0 | 0 | — |
case-02 | fail→pass | 12,293 | 11,816 | -4% | 1 | 1 | 0% | 1,993 | 2,016 | +1% | 0 | 0 | — |
case-03 | pass→pass | 12,277 | 11,105 | -10% | 1 | 1 | 0% | 1,717 | 2,125 | +24% | 0 | 0 | — |
case-04 | fail→fail | 15,487 | 5,052 | -67% | 1 | 1 | 0% | 2,418 | 1,352 | -44% | 0 | 0 | — |
case-05 | fail→pass | 33,354 | 18,119 | -46% | 1 | 1 | 0% | 2,928 | 3,260 | +11% | 0 | 0 | — |
case-06 | fail→pass | 18,440 | 9,859 | -47% | 1 | 1 | 0% | 2,782 | 1,801 | -35% | 0 | 0 | — |
case-07 | pass→pass | 17,372 | 11,488 | -34% | 1 | 1 | 0% | 2,679 | 2,326 | -13% | 0 | 0 | — |
case-08 | pass→pass | 28,255 | 19,202 | -32% | 1 | 1 | 0% | 2,894 | 2,832 | -2% | 0 | 0 | — |
case-09 | pass→pass | 14,567 | 14,683 | +1% | 1 | 1 | 0% | 2,293 | 2,523 | +10% | 0 | 0 | — |
case-10 | fail→pass | 11,898 | 6,762 | -43% | 1 | 1 | 0% | 1,619 | 1,578 | -3% | 0 | 0 | — |
case-11 | fail→pass | 14,316 | 8,349 | -42% | 1 | 1 | 0% | 2,311 | 1,944 | -16% | 0 | 0 | — |
case-12 | pass→pass | 18,609 | 17,962 | -3% | 1 | 1 | 0% | 2,765 | 3,170 | +15% | 0 | 0 | — |
case-13 | fail→pass | 13,663 | 3,943 | -71% | 1 | 1 | 0% | 2,301 | 1,001 | -56% | 0 | 0 | — |
case-14 | pass→pass | 10,616 | 2,539 | -76% | 1 | 1 | 0% | 1,832 | 841 | -54% | 0 | 0 | — |
case-15 | pass→pass | 13,551 | 3,068 | -77% | 1 | 1 | 0% | 1,645 | 828 | -50% | 0 | 0 | — |
case-16 | fail→fail | 6,937 | 4,036 | -42% | 1 | 1 | 0% | 1,023 | 927 | -9% | 0 | 0 | — |
case-17 | pass→fail | 9,935 | 2,102 | -79% | 1 | 1 | 0% | 1,359 | 770 | -43% | 0 | 0 | — |
case-18 | fail→pass | 12,562 | 13,778 | +10% | 1 | 1 | 0% | 2,027 | 1,627 | -20% | 0 | 0 | — |
case-19 | pass→fail | 16,645 | 17,194 | +3% | 1 | 1 | 0% | 2,730 | 3,000 | +10% | 0 | 0 | — |
case-21 | fail→pass | 18,521 | 16,291 | -12% | 1 | 1 | 0% | 2,629 | 2,661 | +1% | 0 | 0 | — |
case-22 | pass→fail | 35,549 | 6,999 | -80% | 1 | 1 | 0% | 7,326 | 814 | -89% | 0 | 0 | — |
case-23 | pass→fail | 29,844 | 12,565 | -58% | 1 | 1 | 0% | 4,579 | 1,257 | -73% | 0 | 0 | — |
case-24 | pass→fail | 16,550 | 6,422 | -61% | 1 | 1 | 0% | 3,339 | 834 | -75% | 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. 24 cases were attempted, and 21 counted toward the lift figure. The other 3 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 +17 percentage points is the difference between those two pass rates over the 21 comparable cases. 5 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.