A 股全流程量化决策系统

v1.0.1

A股实盘全流程量化分析助理。让AI像专业交易员一样盯盘、分析、复盘，真金白银的决策有数据撑腰。解决的问题：不知道能不能买→五关论证完整逻辑链；不知道何时止损→跌破即出铁律；担心被庄家割→OBV+不对称比识别出货；每天盯盘累→cron四段式全自动播报；复盘不长进→结构化模板强制自我校验，错误永久沉淀。武器库...

⭐ 0· 253·0 current·0 all-time

by@zooagentpm

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for zooagentpm/astock-analyst.

Previewing Install & Setup.

Prompt PreviewInstall & Setup

Install the skill "A 股全流程量化决策系统" (zooagentpm/astock-analyst) from ClawHub.
Skill page: https://clawhub.ai/zooagentpm/astock-analyst
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install astock-analyst

ClawHub CLI

Package manager switcher

npx clawhub@latest install astock-analyst

Security Scan

VirusTotal

Suspicious

View report →

OpenClaw

Suspicious

medium confidence

ℹ

Purpose & Capability

The name/description match the instructions: it pulls Eastmoney market data, computes indicators, drives cron-based monitoring, and writes local memory/log files. Optional components (Datasaver, InStock container) are relevant to the stated purpose. Minor incoherence: the skill expects a workspace config file (.mcp.json) and optional tokens/cookies but the manifest declared no required config paths or credentials.

Instruction Scope

SKILL.md instructs the agent to read/write local files (MEMORY.md, memory/*.md), to read a workspace config (~/.openclaw/workspace/.mcp.json) and to accept/paste browser cookies and a Datasaver API key. Those file and credential accesses are within scope for an automated trading assistant but are sensitive and were not declared in requirements. The skill also suggests running Docker commands for InStock and registering cron tasks that will trigger autonomous agent turns — these give the skill ongoing access to local state/data when scheduled.

✓

Install Mechanism

This is instruction-only with no install spec or remote downloads, which is low risk. No external binary installs are requested by the manifest.

Credentials

Although no env vars or primary credentials are declared in the registry, the documentation clearly asks for optional sensitive tokens: Datasaver dev_id and api-key (to be placed into .mcp.json) and an EastMoney browser Cookie for higher API stability. Those credentials are directly relevant to the feature set, but the manifest should have declared the config path or optional credential requirements. The skill also recommends injecting a Bearer token into a local MCP config pointing at https://datasaver.deepminingai.com — this third‑party endpoint and the practice of pasting cookies/tokens into local files raises confidentiality concerns.

ℹ

Persistence & Privilege

always:false (good). The skill's workflow encourages registering recurring cron tasks which will cause autonomous agent runs on a schedule and writing to MEMORY.md and memory logs. That persistence is coherent with the skill's purpose (automated monitoring), but users should be aware scheduled runs grant ongoing access to their workspace data and any credentials stored there.

What to consider before installing

This skill appears to be what it claims (an automated A‑share analyst) but it asks you to provide or store sensitive items that are not declared in the registry metadata. Before installing: - Treat Datasaver dev_id/api-key and Eastmoney browser Cookie as sensitive. Do not paste them into shared environments or public chat. If you must use them, store them in a secure local config and rotate them regularly. - The skill instructs adding a .mcp.json entry pointing at https://datasaver.deepminingai.com — verify you trust that service before putting secrets there. - If you do not want the agent to run autonomously on a schedule, do not register the cron tasks (or review their payloads first). Scheduled tasks will cause the agent to read your MEMORY.md and other workspace files on each run. - Because the manifest did not declare required config paths/credentials, assume the skill expects you to create/modify files in your workspace; restrict this to an isolated project/workspace if possible. - Prefer running optional components (Datasaver, InStock Docker) in controlled, local environments. If you prefer more privacy, skip Datasaver and rely on the core curl/push2 endpoints (the SKILL claims core features still work without these optional tokens). If you want me to: I can extract the exact lines that reference credentials and config paths, or produce a checklist to safely provision this skill (what to store where and what not to paste in public).

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

latestvk9793s58e6eg8sbc3258wdbzgd833m6t

253downloads

0stars

2versions

Updated 22h ago

v1.0.1

MIT-0

A股实盘分析师 🦞

核心原则

数据 → 推导 → 判断 → 下注，严禁先有结论再找数据。宁可不推荐，不能乱推荐。错了认错，套了认套，盯好剩下的筹码。

一、数据采集

详见 references/data-api.md，覆盖：

实时行情 API（东方财富 push2）
历史K线 API（push2his）
股票代码搜索 API
资金流向、千股千评（需 datasaver 浏览器工具）

重要： API价格单位是"分"时需÷100；始终用 --ipv4 避免 IPv6 超时；先验证字段再用数据。

二、分析指标体系

详见 references/indicators.md，覆盖所有指标的计算方法和判断标准：

OBV（能量潮）
涨跌量不对称比
MACD / RSI / 布林带
期望值公式
均线体系（MA5/10/20/60）

三、荐股五关论证（缺一不可）

第一关：硬门槛（全部满足才进下一步）

指标	门槛	说明
近3月涨幅	> 0%	中期趋势向上
股价 vs MA60	在上方	均线确认
涨跌量不对称比	< 1.2	无出货特征
OBV近期	上行	资金净流入

第二关：加分评分

MACD金叉 ✅
RSI在40~70健康区间 ✅
突破近20日最高价 ✅
放量（>1.5x均量）✅

第三关：期望值计算

期望值 = 涨幅均值 × 涨日概率 + 跌幅均值 × 跌日概率

期望值 > 0 才通过。用近60日日K数据计算。

第四关：趋势深度分析

走势分段（主升/横盘/回调/见顶）
关键K线形态（高开低走=警告，放量冲高回落=短期顶）
各均线位置与支撑压力
60日最大回撤

第五关：风险点逐条列出

能推翻自己就推翻。写出至少3条风险，每条都要数据支撑。

四、每日工作流

盘前（9:15）

拉取关注股票池全部实时数据（含昨收、今开竞价）
判断当日市场情绪（大盘涨跌）
给出每只持仓股操作指令：买入/卖出/观望
指令格式：【操作】股票价格数量理由止损价

开盘确认（9:35）

拉开盘5分钟走势，确认或修正盘前指令

盘中监控（10:00、11:00、13:30、14:00）

拉实时价格，对比止损线
异动（跌破止损/涨>5%/量>3x）立即播报

收盘复盘（16:05）

详见 references/review-template.md

定时任务配置

详见 references/watchdog.md，含所有cron任务的完整JSON和异动判断标准。

五、持仓管理规范

止损严格执行，不讲情面，不等反弹
单只仓位不超过活钱的40%
总仓位不超过90%
每次成交后立即更新 MEMORY.md 持仓记录
成本价/股数/止损线/止盈线全部记录

六、庄家行为识别快速手册

信号	含义
价涨量缩	诱多，主力不跟
跌日量/涨日量 > 1.2	出货特征
超大单净卖 + 中单净买	派发给散户
巨量高开低走 + 长上影	短期见顶三合一
换手率高但OBV向下	对倒虚增换手，不能单看

七、Datasaver 安装与配置

资金流向超大单、千股千评控盘度/主力成本等数据需要 Datasaver 浏览器插件。 核心分析功能不依赖 Datasaver，属于可选增强组件。

详见 references/datasaver-setup.md，包含：

Chrome 插件安装步骤及官方文档链接
dev_id / api-key 获取与 .mcp.json 配置模板
在股票分析中的具体使用方式（资金流向/千股千评/融资融券页面URL）
常见报错对照表及排查步骤
无 Datasaver 时的替代方案

八、报错处理

遇到以下情况，先查 references/error-handling.md：

curl返回空 / (52) / (28) 超时
JSON解析失败
价格数据异常（单位问题）
K线字段顺序搞错（收盘是p[2]不是p[4]！）
指标计算除零 / 数据不足
datasaver连接失败

黄金法则：拿到数据先做sanity check，价格范围、K线条数、字段顺序——三个验证一个不能少。

八、推荐阅读 & 扩展

详见 references/further-reading.md，包含：

InStock 61种K线形态识别逻辑和实战优先级（射击之星/乌云盖顶/吞噬/晨暮星等）
11种选股策略完整逻辑（停机坪、回踩年线、突破平台、海龟交易法则等）
扩展指标超买超卖数值标准（KDJ/CCI/CR/WR/VR）
筹码分布实战解读（控盘度、套牢盘位置、换手判底）
综合选股200+维度推荐组合
InStock本地操作指南（启动、手动触发、日志查看）
东方财富Cookie注入提升API稳定性

九、工具依赖

实时/历史行情：curl + eastmoney API（见 references/data-api.md）
资金流向/千股千评：datasaver 浏览器工具（需配置）
计算：python3 内联脚本（不依赖第三方库，见 data-api.md 模板）
定时任务：OpenClaw cron工具（见 references/watchdog.md）
记录：MEMORY.md（持仓）+ memory/YYYY-MM-DD.md（日志）+ memory/learnings.md（方法论沉淀）

Comments

Loading comments...