Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股产业链分析/上下游梳理。当用户说"产业链"、"上下游"、"供应链"、"industry chain"、"XX的上游是什么"、"XX的下游客户"、"产业链图谱"、"国产替代"、"卡脖子"、"哪些公司在这个产业链上"时触发。梳理特定产业的上中下游结构、核心环节、关键公司、竞争格局和投资机会，支持新能源/半导体/AI/医药等热门产业链。支持研报风格（formal）和快速梳理风格（brief）。
.claude/skills/aifinlab-a-share-industry-chain/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | 56% | 0% |
| case-04 | ✓→✗ | ▼ Worse | 92% | 0% |
| case-05 | ✓→✗ | ▼ Worse | -15% | 0% |
| case-06 | ✓→✗ | ▼ Worse | -36% | 0% |
| case-07 | ✓→✗ | ▼ Worse | -17% | 0% |
bashSCRIPTS="$SKILLS_ROOT/cn-stock-data/scripts" # 产业链内公司行情（市值、PE、PB 等） python "$SCRIPTS/cn_stock_data.py" quote --code [CODE1],[CODE2],... # 各公司财务指标（营收/利润/毛利率/ROE） python "$SCRIPTS/cn_stock_data.py" finance --code [CODE] # 各公司近 1 年日线（产业链环节走势对比） python "$SCRIPTS/cn_stock_data.py" kline --code [CODE] --freq daily --start [1年前日期]
补充数据源：
python# 行业板块分类（确定产业链相关行业） import akshare as ak df = ak.stock_board_industry_name_em() # 东方财富行业板块列表
Web 搜索（必须使用，产业链数据时效性强）：
references/industry-chain-guide.md 中的热门产业链速查通过 web 搜索 + 行业知识，绘制产业链结构：
上游（原材料/设备/核心零部件）
├── 环节A：XX材料 → 代表公司、市场规模、国产化率
├── 环节B：XX设备 → 代表公司、市场规模、国产化率
└── 环节C：XX零部件 → 代表公司、市场规模、国产化率
中游（制造/加工/集成）
├── 环节D：XX制造 → 代表公司、产能、市占率
└── 环节E：XX加工 → 代表公司、产能、市占率
下游（应用/终端/消费）
├── 环节F：XX应用 → 代表公司/客户、市场需求
└── 环节G：XX终端 → 代表品牌、渗透率每个环节标注：
通过 cn-stock-data 获取各环节上市公司数据，构建产业链公司地图：
| 环节 | 公司 | 代码 | 市值(亿) | 营收(亿) | 市占率(%) | 毛利率(%) | PE(TTM) | 上市/未上市 | |------|------|------|---------|---------|----------|----------|---------|-----------| | 上游-材料 | ... | ... | ... | ... | ... | ... | ... | A股 | | 中游-制造 | ... | ... | ... | ... | ... | ... | ... | A股 | | 下游-应用 | ... | ... | ... | ... | ... | ... | ... | A股 |
关键要求：
分析各环节的投资价值：
| 环节 | 利润率 | 壁垒高低 | 增速 | 国产替代空间 | 投资吸引力 | |------|--------|---------|------|------------|-----------| | 上游-材料 | XX% | 高/中/低 | XX% | 大/中/小 | 星级评分 | | 中游-制造 | XX% | 高/中/低 | XX% | 大/中/小 | 星级评分 | | 下游-应用 | XX% | 高/中/低 | XX% | 大/中/小 | 星级评分 |
重点分析：
根据风格要求输出（默认 brief）。参考 references/industry-chain-guide.md 中的分析框架。
| 维度 | formal（产业链深度报告） | brief（快速梳理） | |------|----------------------|-----------------| | 篇幅 | 5-8 页 | 1-2 页 | | 产业链结构 | 完整（每个环节详细展开） | 精简（核心环节+关键公司） | | 公司地图 | 全面（15-20 家，含未上市） | Top 公司（8-10 家，A股为主） | | 数据深度 | 含财务对比+估值分析 | 市值+毛利率+PE 核心指标 | | 国产替代 | 单独章节详细分析 | 1-2 段概述 | | 投资建议 | 分环节给出投资逻辑 | 总结最值得关注的 2-3 个方向 | | 免责声明 | 需要 | 不需要 |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 31,381 | 58,717 | +87% | 1 | 1 | 0% | 4,819 | 3,476 | -28% | 0 | 0 | — |
case-02 | fail→fail | 25,372 | 15,870 | -37% | 1 | 1 | 0% | 4,268 | 2,525 | -41% | 0 | 0 | — |
case-03 | fail→fail | 34,618 | 20,303 | -41% | 1 | 1 | 0% | 5,305 | 2,813 | -47% | 0 | 0 | — |
case-04 | pass→fail | 26,477 | 60,477 | +128% | 1 | 1 | 0% | 5,325 | 10,243 | +92% | 0 | 0 | — |
case-05 | pass→fail | 19,046 | 9,054 | -52% | 1 | 1 | 0% | 3,450 | 2,916 | -15% | 0 | 0 | — |
case-06 | pass→fail | 20,936 | 8,945 | -57% | 1 | 1 | 0% | 3,333 | 2,127 | -36% | 0 | 0 | — |
case-07 | pass→fail | 19,386 | 30,172 | +56% | 1 | 1 | 0% | 3,209 | 2,672 | -17% | 0 | 0 | — |
case-08 | fail→fail | 53,287 | 17,315 | -68% | 1 | 1 | 0% | 8,259 | 2,303 | -72% | 0 | 0 | — |
case-09 | fail→fail | 30,534 | 21,751 | -29% | 1 | 1 | 0% | 4,751 | 3,041 | -36% | 0 | 0 | — |
case-10 | fail→fail | 30,226 | 11,358 | -62% | 1 | 1 | 0% | 4,712 | 2,390 | -49% | 0 | 0 | — |
case-11 | pass→fail | 29,083 | 13,541 | -53% | 1 | 1 | 0% | 3,800 | 2,340 | -38% | 0 | 0 | — |
case-12 | fail→fail | 14,264 | 14,136 | -1% | 1 | 1 | 0% | 1,878 | 2,364 | +26% | 0 | 0 | — |
case-13 | pass→fail | 29,089 | 21,995 | -24% | 1 | 1 | 0% | 4,604 | 2,942 | -36% | 0 | 0 | — |
case-14 | fail→fail | 28,849 | 20,299 | -30% | 1 | 1 | 0% | 4,220 | 3,005 | -29% | 0 | 0 | — |
case-15 | fail→fail | 26,276 | 10,805 | -59% | 1 | 1 | 0% | 3,925 | 2,407 | -39% | 0 | 0 | — |
case-16 | pass→fail | 36,254 | 17,632 | -51% | 1 | 1 | 0% | 5,496 | 2,429 | -56% | 0 | 0 | — |
case-17 | fail→fail | 27,675 | 53,072 | +92% | 1 | 1 | 0% | 3,973 | 8,875 | +123% | 0 | 0 | — |
case-18 | fail→pass | 33,387 | 42,480 | +27% | 1 | 1 | 0% | 4,615 | 7,190 | +56% | 0 | 0 | — |
case-19 | fail→fail | 27,344 | 10,750 | -61% | 1 | 1 | 0% | 4,206 | 1,973 | -53% | 0 | 0 | — |
case-20 | fail→fail | 19,151 | 9,968 | -48% | 1 | 1 | 0% | 3,088 | 1,924 | -38% | 0 | 0 | — |
case-21 | pass→fail | 29,090 | 17,765 | -39% | 1 | 1 | 0% | 4,526 | 2,656 | -41% | 0 | 0 | — |
case-22 | fail→fail | 28,996 | 27,174 | -6% | 1 | 1 | 0% | 4,447 | 3,613 | -19% | 0 | 0 | — |
case-23 | pass→pass | 22,939 | 36,728 | +60% | 1 | 1 | 0% | 3,506 | 6,378 | +82% | 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 3 counted toward the lift figure. The other 20 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 -30 percentage points is the difference between those two pass rates over the 3 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.