Stockselectionmodel.Bak

A股板块实时研报。全球新闻→A股映射、板块趋势研判、美股关联、利好利空、大事件提醒。零配置可用，输出兼容飞书/扣子/Markdown平台。

Security Scan

The skill's code and instructions match its stated purpose (A‑share sector reports) but there are small inconsistencies you should be aware of before installing (notably the scripts expect to run with Node yet no runtime/binary requirement is declared, and the skill writes local files).

Purpose & Capability

The name/description (A股板块实时研报) aligns with the included scripts: fetching news, mapping global news to A‑share candidates, sector analysis, market monitoring, and generating markdown reports. Data sources used (EastMoney, public news sites, optional Tavily) are consistent with the stated functionality. One mismatch: the skill runtime instructions and scripts assume executing 'node' commands, but the registry metadata does not declare 'node' or any required binary — this is an omission/incoherence to surface to the user.

Instruction Scope

SKILL.md instructs the agent to run local Node scripts that fetch public news and market data, synthesize summaries, and produce markdown. The instructions do not direct the agent to read unrelated system files or to exfiltrate secrets beyond the optional Tavily API key. The required output layout and message-splitting behaviour are explicit and bounded.

Install Mechanism

There is no install spec (instruction-only), and all code is present in the skill bundle. No remote downloads or install steps are performed by the skill itself, which reduces supply-chain risk. The scripts are plain JS (no obfuscated payloads or external shell downloads).

Credentials

The only declared environment variable is an optional TAVILY_API_KEY for an enhanced news source; that is proportionate to the feature. The code does not request other credentials. Scripts forward process.env to child processes but do not explicitly read or transmit other environment secrets.

Persistence & Privilege

The orchestrator script creates a .local directory inside the skill folder and writes JSON reports there (local persistence). The skill does not request elevated privileges, is not 'always:true', and does not modify other skills or system-wide agent config. If you care about disk writes, be aware the skill will store files under its .local directory.

What to consider before installing

What to check before installing: - Runtime: The scripts are Node.js scripts and SKILL.md runs them with `node ...`. The skill metadata did not declare 'node' as a required binary — ensure your agent environment provides a compatible Node runtime (Node 18+ recommended for built-in fetch/AbortSignal). If your environment lacks Node, the skill's instructions will fail. - Network access: The skill fetches data from public endpoints (EastMoney push APIs, news websites, and optionally Tavily). If you run this in a restricted/sandboxed environment, verify outbound HTTP(S) is allowed and acceptable. - Optional key: Tavily is optional; only provide TAVILY_API_KEY if you trust that service and want enhanced search results. - Local writes: The skill writes JSON files into a .local directory inside its repository — if you have concerns about disk storage or logs, review these files or run the skill in an isolated workspace. - Review code if unsure: The JS files are readable and use public APIs; no obfuscated or clearly malicious behavior was found, but if you plan to use the skill in production or with sensitive contexts, consider auditing the scripts (especially network endpoints and any future modifications). Overall: functionality is coherent with its description, but the missing declaration of the required runtime (node) and the local file writes are small but important operational/integration inconsistencies to verify before enabling the skill.

AI 股票内参 / 板块实时研报

实时拉取板块行情 + 新闻 + 美股关联 + 大事件，生成专业板块研报。

你可以这样问

板块研报（核心功能）

• "新能源板块怎么看？"
• "化工今天有什么催化？"
• "医药最近有哪些利空？"
• "AI 方向对应哪些美股龙头？"
• "券商板块这周有没有大事件？"

AI 内参

• "今天的 AI 新闻"
• "AI 晨报"

异动提醒

• "看一下 AI 概念股有没有异动"

Agent 工作流

当用户问板块/个股问题时，按以下顺序执行：

第 1 步：实时生成研报

node scripts/generate_brief.js --mode sector --sector <板块名> --output pretty

脚本会自动：

1. 拉取板块实时行情（东方财富免费API）
2. 抓取板块相关新闻（东方财富搜索）
3. 抓取美股关联标的 + 近期大事件
4. 自动提取驱动逻辑、利好利空
5. 生成完整 markdown 研报

第 2 步：发送给用户

拿到 markdown 后直接发送。如果内容超过平台单条消息限制（如飞书约4000字符），拆成上下两条发送。

注意：新闻摘要需要人工精炼为一句话概括，不要直接用脚本原始输出的长文本。

第 3 步：补充分析（可选）

如果用户需要更深入的分析，再结合 references/sector_prompt.md 做 LLM 研判。

研报固定布局（每次输出必须包含以下全部元素）

📊 {板块名} · 板块研报
{时间}

▎趋势判断
{趋势图标} {趋势文字}  ·  成分股 X 只  ·  涨停 X  跌停 X
{趋势说明}

▎核心指标（表格）
趋势 / 成分股 / 涨停 / 跌停 / 涨跌比

▎板块动态（表格）
资讯摘要（一句话概括） | 时间 | 来源 | 链接

▎驱动逻辑
• 从新闻自动提取

▎涨跌分布
🟥🟩 可视化条 + 数字

▎领涨个股（表格）
股票 | 代码 | 涨跌幅 | 状态

▎领跌个股（表格）
股票 | 代码 | 涨跌幅 | 状态

▎观点与预测
🔮 一句话预测

▎美股关联（表格）
美股标的 | 代码 | 关联说明

▎利好 / 利空
📈 利好因素（✅ 列表）
📉 利空因素（❌ 列表）

▎近期大事件（表格）
事件概括 | 时间 | 来源 | 链接

▎关注要点
1. 2. 3. 编号列表

▎风险提示
⚠️ 列表

— 免责声明 —

重要：每次输出必须包含以上全部板块，不可省略任何元素。

支持的板块

新能源 / 光伏 / 锂电池 / 储能 / 风电 / 充电桩 AI / 算力 / 大模型 / 半导体 / 光模块 / 机器人 / 自动驾驶 医药 / 创新药 / 中药 / 医美 / 减肥药 / 生物疫苗 化工 / 化学制品 / 煤化工 / 磷化工 / 化纤 白酒 / 消费 / 券商 / 银行 / 保险 / 军工 / 房地产 / 数字经济 / 华为

文件结构

ai-stock-insider/
├── SKILL.md
├── scripts/
│   ├── generate_brief.js      # 一键总控入口
│   ├── fetch_news.js          # 全球 AI 新闻抓取
│   ├── fetch_sector_news.js   # 板块新闻抓取（东方财富搜索）
│   ├── fetch_us_related.js    # 美股关联 + 大事件抓取
│   ├── map_stocks.js          # 全球新闻 → A 股映射
│   ├── market_monitor.js      # AI 概念股异动监控
│   └── sector_analyze.js      # A 股板块趋势分析
└── references/
    ├── analysis_prompt.md     # AI 内参 prompt
    ├── sector_prompt.md       # 板块研判 prompt
    ├── stock_mapping.md       # A 股概念映射表
    └── source_list.md         # 数据源说明

平台兼容

• 飞书：markdown 表格 + 链接均可渲染
• 扣子（Coze）：标准 markdown 兼容
• 其他 markdown 平台：标准语法，无特殊依赖

环境变量（可选）

变量	说明	必需
TAVILY_API_KEY	Tavily 搜索 API Key（增强新闻模式）	否

省 token 设计

1. 行情数据、新闻、美股关联全部由脚本拉取，不消耗 LLM token
2. 驱动逻辑、利好利空由脚本正则提取
3. LLM 只在用户需要深度分析时介入
4. 异动提醒不走 LLM

免责声明

本工具仅提供信息参考，不构成投资建议。股市有风险，投资需谨慎。