Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股“MACD底背离 + 零轴下二次金叉”实战分析与交易执行技能。用于把用户的盘感描述转成可执行的三档决策（上观察名单/可试错出手/必须放弃）、盘中检查单、入场与风控规则。适用于用户提到“二次金叉”“水下二次金叉”“底背离”“第一脚第二脚”“回踩验底”“MACD背离”“抄底修复”“想做但怕假信号”等场景。
.claude/skills/shouldnotappearcalm-macd-second-golden-cross/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 32% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 72% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 21% | 0% |
将“看着像底背离”转成可验证、可执行、可止损的交易框架。
按顺序检查，不跳步：
满足多数特征：
输出动作：
必须同时看到：
执行建议：
任一命中即可放弃：
每次分析按以下顺序逐条打勾：
判定规则：
text【形态结论】 - 当前分档：上观察名单 / 可试错出手 / 必须放弃 - 核心理由： 1) … 2) … 3) … 【结构证据】 - 第一脚：… - 第二脚：… - MACD对比（第一脚 vs 第二脚）：DIF低点…，绿柱… - 二次金叉位置：零轴下/否；高于第一次/否 【执行计划】 - 触发条件：… - 入场方式：… - 止损位：第二脚低点下…% - 失效信号：… - 仓位建议：轻仓/观望 【风险提示】 - 背离是预警不是保证反转；若无价格确认，优先等待。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 20,721 | 16,757 | -19% | 1 | 1 | 0% | 3,120 | 3,787 | +21% | 0 | 0 | — |
case-02 | fail→fail | 17,043 | 17,290 | +1% | 1 | 1 | 0% | 2,720 | 3,678 | +35% | 0 | 0 | — |
case-03 | fail→pass | 19,273 | 15,175 | -21% | 1 | 1 | 0% | 3,070 | 4,042 | +32% | 0 | 0 | — |
case-04 | fail→pass | 17,664 | 25,867 | +46% | 1 | 1 | 0% | 2,786 | 4,802 | +72% | 0 | 0 | — |
case-05 | fail→pass | 20,871 | 14,596 | -30% | 1 | 1 | 0% | 2,790 | 3,501 | +25% | 0 | 0 | — |
case-06 | pass→pass | 19,145 | 14,533 | -24% | 1 | 1 | 0% | 2,277 | 3,576 | +57% | 0 | 0 | — |
case-07 | fail→pass | 19,935 | 11,478 | -42% | 1 | 1 | 0% | 2,777 | 3,436 | +24% | 0 | 0 | — |
case-08 | pass→pass | 21,420 | 16,959 | -21% | 1 | 1 | 0% | 2,601 | 3,918 | +51% | 0 | 0 | — |
case-09 | pass→pass | 16,845 | 11,795 | -30% | 1 | 1 | 0% | 2,565 | 3,092 | +21% | 0 | 0 | — |
case-10 | pass→pass | 14,622 | 13,553 | -7% | 1 | 1 | 0% | 2,147 | 3,360 | +56% | 0 | 0 | — |
case-11 | fail→pass | 15,224 | 9,768 | -36% | 1 | 1 | 0% | 2,307 | 2,797 | +21% | 0 | 0 | — |
case-12 | pass→pass | 17,944 | 4,862 | -73% | 1 | 1 | 0% | 2,504 | 2,227 | -11% | 0 | 0 | — |
case-13 | pass→pass | 19,181 | 14,498 | -24% | 1 | 1 | 0% | 2,604 | 3,618 | +39% | 0 | 0 | — |
case-14 | pass→pass | 17,046 | 13,252 | -22% | 1 | 1 | 0% | 2,355 | 2,989 | +27% | 0 | 0 | — |
case-15 | pass→pass | 19,484 | 18,280 | -6% | 1 | 1 | 0% | 2,485 | 4,008 | +61% | 0 | 0 | — |
case-16 | fail→pass | 15,490 | 9,872 | -36% | 1 | 1 | 0% | 2,406 | 3,120 | +30% | 0 | 0 | — |
case-17 | pass→pass | 19,263 | 12,036 | -38% | 1 | 1 | 0% | 2,700 | 3,199 | +18% | 0 | 0 | — |
case-18 | fail→pass | 19,119 | 18,276 | -4% | 1 | 1 | 0% | 2,317 | 3,721 | +61% | 0 | 0 | — |
case-19 | pass→pass | 19,326 | 16,595 | -14% | 1 | 1 | 0% | 2,876 | 3,880 | +35% | 0 | 0 | — |
case-20 | pass→pass | 21,611 | 21,310 | -1% | 1 | 1 | 0% | 2,829 | 4,406 | +56% | 0 | 0 | — |
case-21 | pass→pass | 31,124 | 33,052 | +6% | 1 | 1 | 0% | 6,190 | 7,878 | +27% | 0 | 0 | — |
case-22 | pass→fail | 31,191 | 28,195 | -10% | 1 | 1 | 0% | 4,100 | 5,363 | +31% | 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 +27 percentage points is the difference between those two pass rates over the 22 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.