Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股股票聚类/相似股票发现。当用户说"聚类"、"clustering"、"相似股票"、"类似的股票"、"同类股票"、"走势相似"时触发。量化聚类分析股票相似性。支持formal和brief风格。
.claude/skills/aifinlab-a-share-stock-clustering/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-20 | ✗→✓ | ▲ Improved | -54% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -43% | 0% |
| case-03 | ✓→✗ | ▼ Worse | -81% | 0% |
| case-04 | ✓→✗ | ▼ Worse | -73% | 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]
各簇的共同特征（行业/风格/基本面）
| 维度 | formal | brief | |------|--------|-------| | 聚类结果 | 完整分簇列表 | 目标股所在簇 | | 簇特征 | 各簇详细画像 | 同类股票 | | 应用 | 配对交易候选 | Top 5相似股 | 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-01 | fail→fail | 25,112 | 25,721 | +2% | 1 | 1 | 0% | 3,448 | 1,165 | -66% | 0 | 0 | — |
case-02 | fail→fail | 21,216 | 6,915 | -67% | 1 | 1 | 0% | 3,392 | 813 | -76% | 0 | 0 | — |
case-03 | pass→fail | 25,929 | 7,929 | -69% | 1 | 1 | 0% | 4,796 | 889 | -81% | 0 | 0 | — |
case-04 | pass→fail | 23,374 | 10,986 | -53% | 1 | 1 | 0% | 3,850 | 1,049 | -73% | 0 | 0 | — |
case-05 | pass→fail | 22,080 | 54,336 | +146% | 1 | 1 | 0% | 3,916 | 9,400 | +140% | 0 | 0 | — |
case-06 | pass→pass | 21,403 | 20,482 | -4% | 1 | 1 | 0% | 3,336 | 3,989 | +20% | 0 | 0 | — |
case-07 | pass→pass | 25,775 | 18,449 | -28% | 1 | 1 | 0% | 3,444 | 3,405 | -1% | 0 | 0 | — |
case-08 | pass→pass | 19,742 | 20,796 | +5% | 1 | 1 | 0% | 3,003 | 3,765 | +25% | 0 | 0 | — |
case-09 | pass→pass | 20,674 | 22,672 | +10% | 1 | 1 | 0% | 3,251 | 3,673 | +13% | 0 | 0 | — |
case-10 | pass→pass | 22,506 | 19,451 | -14% | 1 | 1 | 0% | 2,869 | 3,095 | +8% | 0 | 0 | — |
case-11 | pass→pass | 17,335 | 13,430 | -23% | 1 | 1 | 0% | 2,666 | 2,697 | +1% | 0 | 0 | — |
case-12 | pass→pass | 20,784 | 15,640 | -25% | 1 | 1 | 0% | 2,777 | 3,051 | +10% | 0 | 0 | — |
case-13 | pass→fail | 30,413 | 8,795 | -71% | 1 | 1 | 0% | 4,610 | 962 | -79% | 0 | 0 | — |
case-14 | fail→pass | 17,868 | 3,893 | -78% | 1 | 1 | 0% | 3,258 | 1,261 | -61% | 0 | 0 | — |
case-15 | pass→pass | 21,281 | 11,216 | -47% | 1 | 1 | 0% | 3,359 | 2,206 | -34% | 0 | 0 | — |
case-16 | pass→pass | 20,396 | 15,595 | -24% | 1 | 1 | 0% | 3,033 | 3,040 | +0% | 0 | 0 | — |
case-17 | pass→pass | 19,682 | 15,681 | -20% | 1 | 1 | 0% | 2,970 | 3,054 | +3% | 0 | 0 | — |
case-18 | fail→fail | 15,478 | 7,709 | -50% | 1 | 1 | 0% | 2,598 | 1,547 | -40% | 0 | 0 | — |
case-19 | pass→pass | 15,895 | 10,014 | -37% | 1 | 1 | 0% | 2,317 | 2,086 | -10% | 0 | 0 | — |
case-20 | fail→pass | 14,339 | 5,169 | -64% | 1 | 1 | 0% | 2,594 | 1,185 | -54% | 0 | 0 | — |
case-21 | fail→pass | 18,985 | 6,942 | -63% | 1 | 1 | 0% | 2,359 | 1,348 | -43% | 0 | 0 | — |
case-22 | pass→pass | 14,604 | 5,611 | -62% | 1 | 1 | 0% | 2,293 | 1,248 | -46% | 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, and 16 counted toward the lift figure. The other 6 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 -5 percentage points is the difference between those two pass rates over the 16 comparable cases. 6 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.