Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股价值陷阱/低估值陷阱识别。当用户说"价值陷阱"、"value trap"、"低估值陷阱"、"便宜有道理"、"为什么一直不涨"时触发。量化识别低估值股票是否为价值陷阱。支持formal和brief风格。
.claude/skills/aifinlab-a-share-value-trap/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -1% | 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]
PE/PB低于行业均值的股票
多维度打分判断是真便宜还是价值陷阱
| 维度 | formal | brief | |------|--------|-------| | 低估原因 | 多维度分析 | 主要原因 | | 陷阱信号 | 各信号检测结果 | 陷阱概率 | | 建议 | 深度分析+建议 | 买/避 | 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→pass | 25,047 | 35,309 | +41% | 1 | 1 | 0% | 3,363 | 2,821 | -16% | 0 | 0 | — |
case-02 | fail→fail | 18,605 | 23,757 | +28% | 1 | 1 | 0% | 2,520 | 3,597 | +43% | 0 | 0 | — |
case-03 | fail→fail | 26,033 | 29,014 | +11% | 1 | 1 | 0% | 3,419 | 911 | -73% | 0 | 0 | — |
case-04 | fail→pass | 11,697 | 5,716 | -51% | 1 | 1 | 0% | 2,147 | 1,350 | -37% | 0 | 0 | — |
case-05 | fail→pass | 5,367 | 2,869 | -47% | 1 | 1 | 0% | 956 | 1,009 | +6% | 0 | 0 | — |
case-06 | pass→pass | 31,105 | 8,804 | -72% | 1 | 1 | 0% | 2,046 | 2,085 | +2% | 0 | 0 | — |
case-07 | fail→pass | 19,406 | 20,831 | +7% | 1 | 1 | 0% | 2,797 | 3,297 | +18% | 0 | 0 | — |
case-08 | pass→pass | 20,328 | 23,842 | +17% | 1 | 1 | 0% | 2,868 | 3,209 | +12% | 0 | 0 | — |
case-09 | pass→pass | 17,333 | 20,048 | +16% | 1 | 1 | 0% | 2,486 | 3,453 | +39% | 0 | 0 | — |
case-10 | fail→pass | 25,087 | 21,112 | -16% | 1 | 1 | 0% | 3,649 | 3,613 | -1% | 0 | 0 | — |
case-11 | pass→pass | 17,205 | 13,919 | -19% | 1 | 1 | 0% | 2,558 | 2,780 | +9% | 0 | 0 | — |
case-12 | pass→pass | 20,130 | 17,641 | -12% | 1 | 1 | 0% | 2,774 | 2,884 | +4% | 0 | 0 | — |
case-13 | pass→pass | 21,036 | 22,399 | +6% | 1 | 1 | 0% | 2,853 | 3,499 | +23% | 0 | 0 | — |
case-14 | pass→pass | 19,544 | 22,838 | +17% | 1 | 1 | 0% | 2,887 | 3,512 | +22% | 0 | 0 | — |
case-15 | fail→pass | 17,995 | 5,714 | -68% | 1 | 1 | 0% | 2,273 | 1,256 | -45% | 0 | 0 | — |
case-16 | fail→pass | 26,718 | 12,699 | -52% | 1 | 1 | 0% | 2,709 | 2,206 | -19% | 0 | 0 | — |
case-17 | pass→pass | 17,601 | 11,643 | -34% | 1 | 1 | 0% | 2,543 | 2,319 | -9% | 0 | 0 | — |
case-18 | pass→pass | 10,989 | 4,893 | -55% | 1 | 1 | 0% | 1,647 | 1,096 | -33% | 0 | 0 | — |
case-19 | pass→pass | 19,070 | 16,564 | -13% | 1 | 1 | 0% | 2,741 | 2,955 | +8% | 0 | 0 | — |
case-20 | pass→pass | 28,530 | 32,754 | +15% | 1 | 1 | 0% | 4,398 | 4,757 | +8% | 0 | 0 | — |
case-21 | pass→pass | 23,354 | 34,185 | +46% | 1 | 1 | 0% | 4,061 | 5,207 | +28% | 0 | 0 | — |
case-22 | pass→fail | 29,098 | 10,728 | -63% | 1 | 1 | 0% | 4,628 | 1,038 | -78% | 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 19 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 +27 percentage points is the difference between those two pass rates over the 19 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.