Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

stock-analysis-lianghua

分析任意股票的技术指标和趋势,或修改 TradingAgentsV2 中的分析师节点。当用户需要分析某只股票、查看技术指标、获取市场趋势判断时,使用独立分析脚本;当需要新增/修改分析师节点时,参考架构模板。

MIT-0 · Free to use, modify, and redistribute. No attribution required.
2 · 2k · 14 current installs · 14 all-time installs
MIT-0
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
The name/description (technical stock analysis and analyst-node templates) matches the delivered artifacts: analyze_stock.py (analysis script), an indicators reference, and SKILL.md with node templates. Declared dependencies (yfinance, stockstats, pandas, requests) and use of public data sources (Stooq, Yahoo, yfinance) are consistent with the stated purpose.
Instruction Scope
SKILL.md instructs running the included analyze_stock.py and describes modifying TradingAgentsV2 analyst node templates. The runtime instructions reference local project paths (tradingagents/dataflows/data_cache/, .cursor/skills/...) and direct the agent to read/write cached CSVs and potentially modify node code. The instructions do not ask for secrets or system credentials and do not send data to unknown external endpoints, but they do permit file writes within the project and network requests to public finance endpoints.
Install Mechanism
No install spec is provided and this is effectively an instruction+script package. That is low risk: nothing is downloaded or extracted during install. The only runtime network I/O comes from the analysis script contacting public finance APIs (Stooq, Yahoo) which is expected for this skill.
Credentials
The skill requests no environment variables, keys, or config paths. The code operates against public HTTP endpoints and local cache directories; the lack of credential requests is appropriate for the described data sources.
Persistence & Privilege
always is false and the skill does not request persistent platform-wide privileges. It may write cache files under local project paths and SKILL.md describes editing project node code, which is normal for a development helper but worth noting before allowing autonomous file modifications.
Assessment
This skill appears to do what it says: it fetches market data from public endpoints, computes indicators, and provides templates for LangGraph analyst nodes. Before enabling or running it: 1) review the included analyze_stock.py yourself (it runs HTTP requests and writes cache files to local project paths); 2) ensure you are comfortable the agent can modify files in your project (SKILL.md shows node-editing guidance); 3) install required Python packages in an isolated environment (virtualenv/container) to limit impact; 4) if you need to restrict network access, run the script in an environment without outbound HTTP or inspect logs; and 5) there are no secrets requested by the skill, so you don't need to provide API keys for normal operation.

Like a lobster shell, security has layers — review code before you run it.

Current versionv1.0.0
Download zip
latestvk97cwamrjy37rs442fkx4t5tnx81w10b

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

SKILL.md

股票分析 Skill

快速分析(独立脚本)

脚本位置:.cursor/skills/stock-analysis/analyze_stock.py(项目根目录也有一份副本 analyze_stock.py

当用户要求分析某只股票时,直接执行此脚本:

python .cursor/skills/stock-analysis/analyze_stock.py <股票代码> [选项]

参数

参数说明默认值
symbol (必填)股票代码,如 META, AAPL, 0700.HK, TSLA-
--date, -d分析日期,格式 YYYY-MM-DD今天
--days, -n回看天数90
--indicators, -i逗号分隔的指标列表,或 all8个核心指标

示例

python .cursor/skills/stock-analysis/analyze_stock.py META
python .cursor/skills/stock-analysis/analyze_stock.py AAPL --date 2025-02-20
python .cursor/skills/stock-analysis/analyze_stock.py 0700.HK --days 60
python .cursor/skills/stock-analysis/analyze_stock.py TSLA -i rsi,macd,atr,close_50_sma
python .cursor/skills/stock-analysis/analyze_stock.py NVDA -i all

数据源(多源容灾)

脚本按以下优先级获取数据,自动容灾切换:

  1. Stooq(免费、无需API key、不限速)
  2. Yahoo Chart API(直接 HTTP 请求)
  3. yfinance(Ticker.history)
  4. 本地缓存tradingagents/dataflows/data_cache/ 中已有的 CSV)

成功获取的数据会自动缓存到 data_cache/ 目录。

报告输出内容

  1. 近期行情 - 最近 15 个交易日 OHLCV + 涨跌统计
  2. 技术指标 - 每个指标的时间序列趋势
  3. 综合分析 - 趋势判断、动量分析、波动率分析、短期信号
  4. 指标汇总表 - 所有指标的当前值和信号判断

分析逻辑

脚本内置的分析逻辑对应 market_analyst.py 的 prompt:

  • 趋势判断: 基于价格与 50SMA/200SMA 的位置关系(多头/空头排列 + 金叉/死叉)
  • 动量分析: RSI 超买超卖(70/30 阈值)+ MACD 与信号线交叉 + 柱状图方向
  • 波动率: ATR 占股价比例 + 布林带位置
  • 短期信号: 价格与 10EMA 的关系

支持的全部指标

close_50_sma, close_200_sma, close_10_ema,
macd, macds, macdh,
rsi,
boll, boll_ub, boll_lb,
atr,
vwma

依赖

yfinance, stockstats, pandas, requests

项目架构(LangGraph 分析师节点)

tradingagents/
├── agents/analysts/           # 分析师节点
│   ├── market_analyst.py      # 市场/技术分析
│   ├── fundamentals_analyst.py # 基本面分析
│   ├── news_analyst.py        # 新闻分析
│   └── social_media_analyst.py # 社交媒体情绪分析
├── dataflows/
│   └── interface.py           # 数据接口(工具函数定义)
└── graph/
    └── trading_graph.py       # LangGraph 交易图

分析师节点结构

每个分析师遵循统一模式:create_xxx_analyst(llm, toolkit) -> node_function

核心模板

from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain_core.messages import HumanMessage
from ...dataflows.interface import get_market_type

def create_xxx_analyst(llm, toolkit):
    def xxx_analyst_node(state):
        current_date = state["trade_date"]
        ticker = state["company_of_interest"]
        market_type = get_market_type()  # "CN" 或 "US"

        # 1. 根据 market_type 设置 system_message 和 tools
        if market_type == "CN":
            system_message = "..."
            tools = [toolkit.cn_tool_1, toolkit.cn_tool_2]
        else:
            system_message = "..."
            if toolkit.config["online_tools"]:
                tools = [toolkit.online_tool]
            else:
                tools = [toolkit.offline_tool_1, toolkit.offline_tool_2]

        # 2. 构建 prompt(CN/US 可分别定义或共用)
        prompt = ChatPromptTemplate.from_messages([
            ("system",
             "你是一个有帮助的AI助手,与其他助手协作完成任务。"
             "使用提供的工具来推进回答问题。如果你无法完全回答,没关系;"
             "其他拥有不同工具的助手会在你停下的地方继续帮忙。"
             "你可以使用以下工具:{tool_names}\n{system_message}"
             "供参考,当前日期是 {current_date}。我们正在分析的公司是 {ticker}"),
            MessagesPlaceholder(variable_name="messages"),
        ])
        prompt = prompt.partial(
            system_message=system_message,
            tool_names=", ".join([t.name for t in tools]),
            current_date=current_date,
            ticker=ticker,
        )

        # 3. 调用 LLM
        chain = prompt | llm.bind_tools(tools)
        messages = state["messages"].copy()

        # 工具调用次数限制(防止无限循环)
        tool_call_count = sum(
            1 for msg in messages
            if hasattr(msg, 'tool_calls') and msg.tool_calls
        )
        if tool_call_count >= 3:
            final_prompt = ChatPromptTemplate.from_messages([
                ("system", system_message + "\n\n重要提醒:请基于已获取的信息生成最终报告,不要再调用任何工具。"),
                MessagesPlaceholder(variable_name="messages"),
            ])
            result = (final_prompt | llm).invoke(messages)
        else:
            if not (messages and getattr(messages[-1], "role", None) == "user"):
                messages.append(HumanMessage(content=f"请分析{ticker}的相关信息,并调用相关工具获取数据。"))
            result = chain.invoke(messages)

        # 4. 返回结果(key 与 state schema 对应)
        return {
            "messages": [result],
            "xxx_report": result.content,
        }

    return xxx_analyst_node

State 返回字段映射

分析师返回 key说明
market_analystmarket_report技术指标与趋势分析
fundamentals_analystfundamentals_report财务与基本面分析
news_analystnews_report新闻与公告分析
social_media_analystsentiment_report社交媒体情绪分析

可用技术指标(市场分析师)

指标名称必须与以下精确匹配,否则工具调用会失败:

类别指标说明
移动平均线close_50_sma50日简单移动平均线
close_200_sma200日简单移动平均线
close_10_ema10日指数移动平均线
MACDmacdMACD 值
macdsMACD 信号线
macdhMACD 柱状图
动量rsi相对强弱指数
波动率boll / boll_ub / boll_lb布林带(中/上/下轨)
atr平均真实波幅
成交量vwma成交量加权移动平均线

选择指标时最多 8 个,避免冗余(如不要同时选 rsi 和 stochrsi)。

数据工具对照表

分析师A股(CN)工具美股(US)在线工具美股离线工具
市场分析get_akshare_data / get_akshare_data_onlineget_YFin_data_onlineget_YFin_data
get_stockstats_indicators_report / _onlineget_stockstats_indicators_report_onlineget_stockstats_indicators_report
基本面get_akshare_balance_sheetget_fundamentals_openaiget_simfin_* / get_finnhub_*
get_akshare_cashflow / income_stmt / finance_analysis
get_akshare_special_data
新闻get_company_news / get_market_newsget_global_news_openai / get_google_newsget_finnhub_news / get_reddit_news / get_google_news
社交媒体get_xueqiu_stock_infoget_stock_news_openaiget_reddit_stock_info / get_finnhub_news

工具函数定义在 tradingagents/dataflows/interface.py

市场类型配置

通过 get_market_type() 获取,返回 "CN""US"。配置来源于 tradingagents/dataflows/config.py

修改指南

新增技术指标

  1. interface.py 的 stockstats 工具中添加指标定义
  2. market_analyst.pysystem_message 中添加指标描述
  3. 指标名需与 stockstats 库一致

新增分析师类型

  1. tradingagents/agents/analysts/ 创建新文件
  2. 遵循上方核心模板
  3. trading_graph.py 中注册新节点
  4. 返回值 key 需在 state schema 中定义

修改报告格式

所有分析师 system_message 末尾已要求附加 Markdown 表格总结。如需修改格式,调整 system_message 的指令即可。

注意事项

  • 工具调用上限默认 3 次,超出后强制生成报告(防死循环)
  • online_tools 配置决定使用在线/离线数据源
  • 所有分析师输出中文,system_message 统一用中文编写
  • 报告末尾需附 Markdown 表格,方便前端展示

Files

3 total
Select a file
Select a file to preview.

Comments

Loading comments…