stock-rt-subscribe
Subscribe to real-time stock market data via WebSocket long connection from receiver nodes (A-shares, HK stocks, ETFs). Use this skill when the user wants to monitor live stock prices, set up real-time alerts, or stream minute-level K-line data. The skill provides a WebSocket server that connects to receiver nodes and pushes real-time price updates to all connected clients.
How do I install this agent skill?
npx skills add https://github.com/yourdaylight/stock_datasource --skill stock-rt-subscribeIs this agent skill safe to install?
- Gen Agent Trust Hubwarn
This skill monitors real-time stock data but performs high-privilege operations including runtime package installation and shell command execution. It also processes external stock data and inserts it into user-facing messages without sanitization, creating a potential surface for indirect prompt injection.
- Socketwarn
2 alerts: gptAnomaly
- Snykwarn
Risk: MEDIUM · 2 issues
- Runlayerpass
1 file scanned · No issues
- ZeroLeakspass
Score: 93/100 · 2 sections analyzed
What does this agent skill do?
实时股票数据订阅(WebSocket)
通过 WebSocket 长连接 实时接收 A 股、港股、ETF 行情数据,每 3~5 秒更新一次。
架构
Receiver Node (HTTP API) → scripts/subscribe_client.py → ws://localhost:8765 → 客户端
环境准备
pip install requests websockets
export STOCK_RT_NODE_URL="http://your-node:9100"
export STOCK_RT_TOKEN="your-token" # 可选,有鉴权时设置
快速开始
一次性查询(验证连通性)
python3 scripts/subscribe_client.py --symbols 00700.HK 600519.SH --once
启动 WebSocket 推送服务
# 基础订阅
python3 scripts/subscribe_client.py --symbols 00700.HK 09988.HK 600519.SH
# 指定节点和端口
python3 scripts/subscribe_client.py --node-url http://your-node:9100 --ws-port 8765 --symbols 00700.HK
# 涨跌幅告警(±2% 触发)
python3 scripts/subscribe_client.py --symbols 00700.HK --alert-pct 2.0
# 输出到文件
python3 scripts/subscribe_client.py --symbols 00700.HK --output ticks.jsonl
编程方式启动
from scripts.subscribe_client import StockWSServer
server = StockWSServer(
node_url="http://your-node:9100",
symbols=["00700.HK", "09988.HK"],
poll_interval=3.0,
)
server.add_callback(lambda sym, tick: print(f"{tick.name}: {tick.close}"))
server.run(port=8765)
AI Agent 集成
from scripts.ai_agent_integration import StockDataAgent
agent = StockDataAgent()
agent.install() # 检查依赖、验证节点连通性
# 完整工作流(订阅 + 告警监控 60 秒)
import asyncio
asyncio.run(agent.complete_workflow(['00700.HK', '600519.SH'], duration=60))
告警策略
告警规则配置见 references/strategy_config.json,内置策略:
limit_up_down:涨停/跌停监控big_move:大幅波动(≥ ±5%)vol_spike:成交量异动(≥ 近5日均量3倍)price_breakout:价格突破近20日高点
参考文档
- references/api.md:WebSocket 消息格式、数据字段说明、客户端连接示例(Python/JS/wscat)、动态订阅管理指令、排错指南
- references/strategy_config.json:告警策略完整配置
How can the creator link this skill?
Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.
<a href="https://skillzs.dev/skills/yourdaylight/stock_datasource/stock-rt-subscribe">View stock-rt-subscribe on skillZs</a>