stock-monitor

股票监控分析技能 - 自定义股票池监控、实时行情、技术指标分析、涨跌趋势预测、信号提醒

Security Scan

The skill's code and instructions mostly match a stock‑monitoring purpose, but it calls multiple external search/data services and instructs pushing reports (including local holdings) to external channels while not clearly declaring required credentials — this creates potential for unintended data exposure and some coherence gaps.

Purpose & Capability

Name/description match the included Python scripts and docs: collecting market data, computing indicators and generating reports. Network access to market APIs (Tencent, Sina, Eastmoney, ifzq) is expected for this functionality.

Instruction Scope

Runtime instructions and templates explicitly read local config files (~/.openclaw/stock-*.json) and generate/push reports containing personal holdings. They also instruct using external search/agent services (Tavily, Exa/Agent‑Reach, Jina via r.jina.ai and web_fetch). Those external calls may transmit stock codes, holdings or queries outside the user's environment — the docs do not clearly limit or warn about what personal data may be sent.

Install Mechanism

No install spec; skill is instruction+bundled scripts (no remote downloads). Provided code is local and readable; no install-time arbitrary network downloads or archive extraction were observed.

Credentials

Registry metadata lists no required env vars, and SKILL.md marks a Tavily API key as optional — that is plausible. However the skill expects to push reports to Feishu/other delivery channels (message tool / delivery.channel: feishu) but does not declare any Feishu/webhook credentials. This is a gap: either OpenClaw provides messaging creds at runtime or the skill omitted required credentials. The use of third‑party agent/search services is legitimate for news gathering but increases the blast radius if private data (positions/trades) are included in queries.

Persistence & Privilege

always:false and no special platform privileges requested. The skill reads/writes its own config files under ~/.openclaw and references cron jobs; it does not request to modify other skills or system settings in the provided materials.

What to consider before installing

This skill appears to implement the advertised stock‑monitoring features using public market APIs and local JSON files. Before installing, consider: 1) Reports are designed to include personal holdings and may be pushed to external chat groups — confirm which Feishu/group will receive them and who can read them. 2) The skill will use external search/agent services (Tavily/Exa/Jina/web_fetch) for news; queries or snippets might be sent to those services — avoid including sensitive identifiers in stock configuration if you are concerned about leakage. 3) There is a minor metadata mismatch: SKILL.md mentions an optional Tavily API key and external delivery, but the registry entry lists no required credentials — ask the author whether any messaging/webhook tokens are required and where they should be stored. 4) If you plan to run scheduled tasks, review the cron/job config and restrict execution to an isolated agent session or test environment first. If you want lowest risk, run the scripts locally (offline with cached data) and inspect outgoing network calls; remove or disable automatic pushing to chat if you do not want holdings exposed.

Metadata

openclaw:
  requires:
    files:
      - ~/.openclaw/stock-pool.json        # 股票监控池配置（公开数据）
      - ~/.openclaw/stock-positions.json   # 持仓记录（个人数据，仅本地使用）
      - ~/.openclaw/stock-trades.json      # 交易记录（个人数据，仅本地使用）
      - ~/.openclaw/stock-alerts.json      # 预警配置
      - ~/.openclaw/cron/jobs.json         # 定时任务配置
    apis:
      - 腾讯行情API (qt.gtimg.cn)         # 实时行情（主要数据源）
      - 新浪财经API (sina.com.cn)          # 实时行情（备用数据源）
      - 东方财富API (eastmoney.com)        # K线数据、资讯（主备双通道）
      - Agent Reach (Jina/Exa)            # 资讯获取备选（当Tavily额度满时）
    credentials:
      - Tavily API Key (可选，额度满时使用Agent-Reach替代)
    network: true
    python: true

Stock Monitor - 股票监控分析技能

---

核心特色

• 多维度技术分析：ADX趋势、BOLL轨道、MACD/KDJ/RSI/OBV/DMI/WR指标
• 资讯情感分析：东方财富+新浪财经+雪球
• 持仓智能管理：自动追踪成本、盈亏、FIFO同步
• 买卖点位建议：基于支撑/阻力位+BOLL+均线的精确入场/出场/止损/目标价位
• 中国股市习惯：🔴红色=上涨/盈利 | 🟢绿色=下跌/亏损
• 数据源自动降级：行情/K线失败自动切换备用源，告别数据中断

---

目录结构

stock-monitor/
├── SKILL.md                    本文件
├── references/
│   ├── commands.md             CLI命令速查
│   ├── config.md               配置文件说明
│   ├── index.md                指标说明
│   ├── scheduled-tasks.md      定时任务配置
│   └── troubleshooting.md      故障排查
├── reports/
│   ├── prompts.md              Cron任务Prompt模板
│   └── templates.md            报告模板（完整模板）
└── scripts/
    ├── stock_monitor.py        主监控脚本
    └── stock_capital.py        资金流向脚本

---

数据源降级机制（v1.1+）

为应对数据源不稳定，脚本内置多级自动降级：

数据类型	主数据源	备用①	备用②	最终保底
A股实时行情	腾讯 qt.gtimg.cn	新浪 hq.sinajs.cn	—	返回空
港股实时行情	腾讯 qt.gtimg.cn	新浪 hq.sinajs.cn	—	返回空
A股K线	东方财富 push2his	新浪财经	—	返回空
港股K线	腾讯 web.ifzq.gtimg.cn	东方财富 push2his	实时行情构造单日K线	返回空
大盘指数	腾讯 qt.gtimg.cn	新浪 hq.sinajs.cn	—	返回空

降级标识：当触发备用源时，脚本会打印 [降级] 或 [保底] 日志，便于排查。

---

CLI 命令

工作目录：scripts | 格式：python stock_monitor.py <command> [args]

行情与指数

命令	说明
quote <code>	查询单只股票行情
index	查询大盘指数

技术分析

命令	说明
analyze <code>	分析股票技术指标
monitor-a	监控所有A股
monitor-hk	监控所有港股

综合报告

命令	说明
report	生成完整综合报告
report-a	生成A股综合报告
report-hk	生成港股综合报告

持仓管理

命令	说明
position add <code> <qty> <cost>	添加/更新持仓
position remove <code>	清除持仓
position list	查看持仓

交易记录

命令	说明
trade buy <code> <qty> <price>	记录买入
trade sell <code> <qty> <price>	记录卖出
trades [code]	查看交易记录

---

重要规则

数据优先级（强制）

1. 脚本行情数据（index/monitor-a/monitor-hk）— 权威数据
2. 搜索资讯数据（web_fetch）— 仅用于资讯分析

【强制】报告中所有"现价/涨跌幅/涨跌额"必须使用脚本返回的数据，禁止使用搜索结果中的价格。

技术指标（必须包含）

每只股票：现价、涨跌幅、MA均线多空、MACD状态、KDJ状态、RSI数值、BOLL位置、OBV背离、DMI趋势、WR威廉、综合信号

综合信号评级

评分差	信号
buy > sell + 3	⭐⭐⭐ 强烈买入
buy > sell	⭐⭐ 建议买入
abs差 ≤ 3	⚪ 观望
sell > buy	⭐ 建议卖出
sell > buy + 3	⭐⭐ 强烈卖出

买卖点位计算规则

场景	买入点位	买入区间	卖出点位	卖出区间	止损位	目标位
回踩支撑	支撑位	支撑×0.98~1.02	-	-	支撑下方3%	阻力位/BOLL上轨
突破阻力	突破后确认	阻力~阻力×1.02	-	-	阻力下方3%	上一阻力位
RSI超卖	BOLL下轨/支撑	下轨×0.98~1.02	-	-	前低下方3%	BOLL上轨/阻力位
RSI超买	-	-	BOLL上轨/阻力位	上轨×0.97~1.01	-	-
MACD死叉	-	-	短线阻力位	阻力×0.97~1.01	-	-

计算依据：

• 支撑位：近30日低点、BOLL下轨、60日均线
• 阻力位：近30日高点、BOLL上轨、近期反弹高点
• 止损：支撑/均线下方3%（保守可设2%）
• 目标位：阻力位、BOLL上轨、量度涨幅（前一高低点差）
• 买入区间：在点位附近±2%范围分批建仓
• 卖出区间：在点位附近±3%范围分批止盈

---

报告模板

详见 reports/templates.md

报告生成流程

1. python stock_monitor.py index 获取大盘指数
2. 读取 ~/.openclaw/stock-pool.json 获取股票列表
3. python stock_monitor.py monitor-a/hk 获取技术面数据
4. web_fetch 获取资讯（优先东方财富/新浪/雪球）
5. 生成报告（价格数据必须用步骤3返回的数据）
6. 推送至飞书群聊

---

配置文件

文件	路径
股票池	~/.openclaw/stock-pool.json
持仓记录	~/.openclaw/stock-positions.json
交易记录	~/.openclaw/stock-trades.json
预警配置	~/.openclaw/stock-alerts.json
Cron任务	~/.openclaw/cron/jobs.json

---

使用提示

• 技术指标有滞后性，用于确认趋势而非预测
• 多条件共振更可靠，单一指标容易假信号
• 动态止盈：回撤5%减仓、10%清仓（建议，根据市场灵活调整）
• 核心原则：预警系统目标是"不错过大机会，不犯大错误"