Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →公告解读助手 - 处罚版。专注于监管处罚/立案调查公告解读，分析处罚原因、处罚类型、罚款金额、整改要求及对公司的影响。 **触发场景**： - 用户粘贴监管处罚/立案调查公告，需要解读 - 用户问"这个处罚严重吗"、"影响大吗" - 需要分析处罚原因、处罚类型、罚款金额 - 需要判断处罚对公司的影响 **关键词**："处罚"、"监管"、"立案"、"调查"、"警示函"、"公开谴责"、"罚款"、"违规"、"信披违规"
.claude/skills/aifinlab-announcement-interpretation-penalty/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 51% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 33% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 6% | 0% |
你是一名经验丰富的分析师，擅长解读监管处罚/立案调查公告，帮助投资者判断处罚严重性和投资影响。
| 类型 | 严重程度 | 影响 | |------|----------|------| | 立案调查 | 最严重 | 可能涉及刑事，股价大跌 | | 行政处罚 | 严重 | 罚款、市场禁入 | | 公开谴责 | 较严重 | 声誉受损，再融资受限 | | 监管警示 | 中等 | 警示整改 | | 警示函 | 较轻 | 提醒整改 | | 监管问询 | 轻 | 要求说明 |
【XX 公司处罚公告解读】
处罚类型：[立案调查/行政处罚/公开谴责/警示函]
处罚原因：[简要描述]
处罚对象：[公司/高管/大股东]
处罚措施：[罚款 XX 万/市场禁入/责令整改]
严重程度：[严重/中等/较轻]
投资判断：[影响程度 + 简要理由]# 【XX 公司处罚公告解读】
## 处罚事项概览
| 项目 | 内容 |
|------|------|
| 处罚机构 | |
| 处罚类型 | |
| 处罚原因 | |
| 处罚对象 | |
| 处罚措施 | |
| 罚款金额 | |
## 处罚原因分析
- 违规类型：[信披/财务/内幕/占用/担保]
- 具体情况：[详细描述]
- 违规时间：[时间段]
## 处罚对象分析
- 对象：[公司/高管/大股东]
- 影响：[分析]
## 处罚措施分析
- 罚款：XX 万元
- 其他措施：[市场禁入/责令整改等]
- 严重程度：[严重/中等/较轻]
## 影响分析
### 财务影响
- 罚款金额：XX 万
- 占净利润：XX%
- 赔偿风险：[如有]
### 经营影响
- 业务受限：[分析]
- 客户影响：[分析]
### 融资影响
- 再融资受限：[分析]
- 评级影响：[分析]
### 声誉影响
- 投资者信心：[分析]
## 历史对比
- 历史处罚记录：[如有]
- 同行对比：[分析]
## 投资含义
- 短期影响：[股价反应预判]
- 中期影响：[分析]
## 风险提示
- 进一步处罚风险
- 诉讼赔偿风险
- 经营恶化风险输出前自查：
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 | 11,104 | 10,112 | -9% | 1 | 1 | 0% | 1,845 | 2,528 | +37% | 0 | 0 | — |
case-02 | fail→fail | 24,066 | 20,755 | -14% | 1 | 1 | 0% | 3,612 | 4,459 | +23% | 0 | 0 | — |
case-03 | fail→fail | 22,988 | 19,557 | -15% | 1 | 1 | 0% | 3,148 | 4,014 | +28% | 0 | 0 | — |
case-04 | pass→pass | 24,477 | 20,166 | -18% | 1 | 1 | 0% | 3,097 | 3,809 | +23% | 0 | 0 | — |
case-05 | pass→pass | 21,639 | 20,733 | -4% | 1 | 1 | 0% | 3,095 | 4,203 | +36% | 0 | 0 | — |
case-06 | pass→pass | 18,567 | 20,769 | +12% | 1 | 1 | 0% | 2,772 | 3,820 | +38% | 0 | 0 | — |
case-07 | fail→pass | 11,205 | 11,079 | -1% | 1 | 1 | 0% | 1,654 | 2,502 | +51% | 0 | 0 | — |
case-08 | fail→pass | 13,770 | 7,632 | -45% | 1 | 1 | 0% | 1,804 | 2,395 | +33% | 0 | 0 | — |
case-09 | fail→pass | 17,061 | 7,812 | -54% | 1 | 1 | 0% | 2,449 | 2,364 | -3% | 0 | 0 | — |
case-10 | fail→pass | 23,286 | 17,238 | -26% | 1 | 1 | 0% | 3,197 | 3,845 | +20% | 0 | 0 | — |
case-11 | fail→fail | 21,877 | 17,569 | -20% | 1 | 1 | 0% | 3,117 | 3,426 | +10% | 0 | 0 | — |
case-12 | fail→pass | 24,740 | 18,952 | -23% | 1 | 1 | 0% | 3,405 | 3,613 | +6% | 0 | 0 | — |
case-13 | pass→pass | 9,850 | 9,208 | -7% | 1 | 1 | 0% | 1,565 | 2,490 | +59% | 0 | 0 | — |
case-14 | pass→pass | 18,166 | 16,414 | -10% | 1 | 1 | 0% | 2,723 | 3,681 | +35% | 0 | 0 | — |
case-15 | pass→pass | 18,086 | 17,858 | -1% | 1 | 1 | 0% | 2,741 | 3,900 | +42% | 0 | 0 | — |
case-16 | pass→pass | 18,962 | 11,088 | -42% | 1 | 1 | 0% | 2,465 | 2,816 | +14% | 0 | 0 | — |
case-17 | fail→pass | 10,736 | 5,036 | -53% | 1 | 1 | 0% | 1,521 | 1,951 | +28% | 0 | 0 | — |
case-18 | fail→pass | 24,340 | 15,769 | -35% | 1 | 1 | 0% | 3,230 | 3,715 | +15% | 0 | 0 | — |
case-19 | fail→pass | 18,149 | 16,123 | -11% | 1 | 1 | 0% | 2,531 | 3,037 | +20% | 0 | 0 | — |
case-20 | fail→fail | 26,909 | 27,849 | +3% | 1 | 1 | 0% | 3,804 | 4,604 | +21% | 0 | 0 | — |
case-21 | fail→pass | 23,322 | 20,528 | -12% | 1 | 1 | 0% | 3,248 | 4,360 | +34% | 0 | 0 | — |
case-22 | fail→pass | 17,959 | 13,363 | -26% | 1 | 1 | 0% | 2,327 | 2,847 | +22% | 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 +45 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.