Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股事件驱动分析/公告解读/政策影响评估。当用户说"事件分析"、"公告解读"、"政策影响"、"并购"、"增发"、"回购"、"股权激励"、"XX出了什么公告"、"这个政策对XX有什么影响"、"解禁"、"定增"、"事件驱动"、"event analysis"、"公告分析"、"重组"、"资产注入"、"战略合作"、"利好还是利空"时触发。MUST USE when user asks about event-driven analysis, corporate announcement interpretation, policy impact assessment, M&A analysis, or any event's impact on stock price. 分析公司公告、政策变化、并购重组等事件对股价和基本面的潜在影响，评估事件的正面/负面程度和持续性。通过 cn-stock-data 获取行情和财务数据，结合 web 搜索获取事件详情。支持机构事件点评风格（formal）和个人事件笔记风格（brief）。
.claude/skills/aifinlab-a-share-event/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-13 | ✗→✓ | ▲ Improved | 100% | 0% |
| case-04 | ✓→✗ | ▼ Worse | -67% | 0% |
| case-19 | ✓→✗ | ▼ Worse | 46% | 0% |
| case-21 | ✓→✗ | ▼ Worse | -62% | 0% |
| case-22 | ✓→✗ | ▼ Worse | 39% | 0% |
bashSCRIPTS="$SKILLS_ROOT/cn-stock-data/scripts" # 实时行情（事件发生后的市场反应） python "$SCRIPTS/cn_stock_data.py" quote --code [CODE] # K线（事件前后股价走势对比） python "$SCRIPTS/cn_stock_data.py" kline --code [CODE] --freq daily --start [事件前30日] # 财务指标（评估事件对基本面的影响） python "$SCRIPTS/cn_stock_data.py" finance --code [CODE] # 资金流向（机构对事件的反应） python "$SCRIPTS/cn_stock_data.py" fund_flow --code [CODE] --days 30
补充 web 搜索：公告原文、政策全文、分析师点评、历史类似事件案例。
Step 1: 事件识别与分类
| 事件类型 | 子类 | 典型影响方向 | |---------|------|------------| | 公司公告 | 业绩预告/快报 | 取决于超/低预期 | | | 分红方案 | 通常正面 | | | 增发/配股 | 短期偏负面（摊薄） | | | 回购 | 正面（公司认为低估） | | | 大股东增持 | 正面信号 | | | 大股东减持/解禁 | 负面信号 | | | 股权激励 | 通常正面（绑定管理层） | | | 并购重组 | 取决于标的质量和价格 | | | 关联交易 | 需评估是否损害小股东利益 | | 政策事件 | 行业政策（利好） | 板块普涨 | | | 行业政策（利空） | 板块普跌 | | | 货币/财政政策 | 系统性影响 | | | 监管政策 | 取决于具体内容 | | 市场事件 | 指数调整（纳入/剔除） | 纳入正面，剔除负面 | | | 北向资金大幅增减 | 信号意义 | | | 融券/做空 | 负面压力 |
Step 2: 事件详情获取
Step 3: 影响评估
定量评估：
定性评估：
综合评级： | 维度 | 评分 | |------|------| | 影响方向 | 正面 / 中性 / 负面 | | 影响程度 | 重大 / 中等 / 轻微 | | 持续时间 | 一次性 / 短期 / 中长期 | | 确定性 | 高 / 中 / 低 |
Step 4: 历史类比（可选） Web 搜索类似事件的历史案例：
Step 5: 输出
| 维度 | formal（机构事件点评） | brief（个人事件笔记） | |------|---------------------|---------------------| | 篇幅 | 2-4 页 | 半页 | | 事件描述 | 完整（含公告/政策原文引用） | 一段话概述 | | 定量影响 | 详细计算（EPS/净资产影响） | 核心数字 | | 定性评估 | 完整评级表 | 一句话判断 | | 历史类比 | 2-3 个案例 | 无或 1 个 | | 结论 | 客观分析正反面 | 可加操作建议 | | 免责声明 | 需要 | 不需要 |
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 | 34,353 | 23,923 | -30% | 1 | 1 | 0% | 4,841 | 2,127 | -56% | 0 | 0 | — |
case-02 | fail→fail | 17,272 | 11,163 | -35% | 1 | 1 | 0% | 2,269 | 2,065 | -9% | 0 | 0 | — |
case-03 | fail→fail | 36,089 | 59,806 | +66% | 1 | 1 | 0% | 5,655 | 1,890 | -67% | 0 | 0 | — |
case-04 | pass→fail | 35,334 | 11,156 | -68% | 1 | 1 | 0% | 5,335 | 1,774 | -67% | 0 | 0 | — |
case-05 | pass→pass | 25,984 | 40,960 | +58% | 1 | 1 | 0% | 5,177 | 9,523 | +84% | 0 | 0 | — |
case-06 | pass→pass | 39,024 | 30,743 | -21% | 1 | 1 | 0% | 5,752 | 6,998 | +22% | 0 | 0 | — |
case-07 | fail→fail | 16,427 | 9,950 | -39% | 1 | 1 | 0% | 2,398 | 1,923 | -20% | 0 | 0 | — |
case-08 | fail→fail | 31,580 | 12,542 | -60% | 1 | 1 | 0% | 5,204 | 1,925 | -63% | 0 | 0 | — |
case-09 | fail→fail | 11,455 | 17,698 | +55% | 1 | 1 | 0% | 1,710 | 1,665 | -3% | 0 | 0 | — |
case-10 | fail→fail | 35,844 | 10,577 | -70% | 1 | 1 | 0% | 4,927 | 1,864 | -62% | 0 | 0 | — |
case-11 | fail→fail | 11,188 | 36,277 | +224% | 1 | 1 | 0% | 1,768 | 1,692 | -4% | 0 | 0 | — |
case-12 | fail→fail | 34,515 | 56,622 | +64% | 1 | 1 | 0% | 5,408 | 2,062 | -62% | 0 | 0 | — |
case-13 | fail→pass | 22,244 | 37,101 | +67% | 1 | 1 | 0% | 3,212 | 6,439 | +100% | 0 | 0 | — |
case-14 | fail→fail | 38,225 | 40,672 | +6% | 1 | 1 | 0% | 5,927 | 2,031 | -66% | 0 | 0 | — |
case-23 | fail→fail | 32,589 | 9,087 | -72% | 1 | 1 | 0% | 5,520 | 1,912 | -65% | 0 | 0 | — |
case-15 | fail→fail | 12,294 | 17,854 | +45% | 1 | 1 | 0% | 1,941 | 3,152 | +62% | 0 | 0 | — |
case-16 | fail→fail | 24,144 | 9,445 | -61% | 1 | 1 | 0% | 3,420 | 1,903 | -44% | 0 | 0 | — |
case-17 | fail→fail | 30,805 | 42,908 | +39% | 1 | 1 | 0% | 5,280 | 1,724 | -67% | 0 | 0 | — |
case-18 | fail→fail | 28,967 | 9,725 | -66% | 1 | 1 | 0% | 4,161 | 1,895 | -54% | 0 | 0 | — |
case-19 | pass→fail | 13,591 | 25,379 | +87% | 1 | 1 | 0% | 1,799 | 2,630 | +46% | 0 | 0 | — |
case-20 | fail→fail | 42,227 | 9,162 | -78% | 1 | 1 | 0% | 6,248 | 1,813 | -71% | 0 | 0 | — |
case-21 | pass→fail | 29,476 | 48,979 | +66% | 1 | 1 | 0% | 5,028 | 1,909 | -62% | 0 | 0 | — |
case-22 | pass→fail | 8,536 | 23,806 | +179% | 1 | 1 | 0% | 1,310 | 1,816 | +39% | 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. 23 cases were attempted, and 4 counted toward the lift figure. The other 19 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 -13 percentage points is the difference between those two pass rates over the 4 comparable cases. 17 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.