HuaHuaDailyMCP

v1.2.0

花花日记助手 - 查询用户的基金持仓、今日收益、累计收益、市场行情等数据,并可向 App 发送交易请求信号

1· 280·1 current·1 all-time
bybaiye@baiye1997

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for baiye1997/huahua-daily.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "HuaHuaDailyMCP" (baiye1997/huahua-daily) from ClawHub.
Skill page: https://clawhub.ai/baiye1997/huahua-daily
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Required env vars: BAIYE_AGENT_TOKEN
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 huahua-daily

ClawHub CLI

Package manager switcher

npx clawhub@latest install huahua-daily
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
medium confidence
Purpose & Capability
Name/description match the actual behavior: the skill reads cloud-synced fund data and posts transaction request signals to the Huahua backend. It requires a single Agent Token (BAIYE_AGENT_TOKEN), which is appropriate. Note: SKILL.md/README refer to huahuadaily.preview.aliyun-zeabur.cn while server.py sets _OFFICIAL_API to https://huahua.preview.aliyun-zeabur.cn — a hostname mismatch that should be verified (could be an innocuous preview vs. alias discrepancy or a configuration error).
Instruction Scope
SKILL.md instructs only network calls to the service APIs, use of BAIYE_AGENT_TOKEN, and that request_transaction only sends a request for the user to confirm in the app. The instructions do not ask the agent to read unrelated files or sweep other environment variables. They do instruct the user to place a token in an environment variable or call set_token.
Install Mechanism
The registry entry has no formal install spec, but the package includes server.py, pyproject.toml, and README with installation instructions (uvx/git clone or python). The code depends on public packages (mcp, httpx) and references a GitHub repo for uvx installs. This is a common pattern but means running the skill will execute Python code that performs network requests — verify the referenced GitHub source before running.
Credentials
Only one required environment variable (BAIYE_AGENT_TOKEN) is declared and used as the primary credential; this aligns with the skill's need to call the backend API. No unrelated secrets or broad credential access is requested.
Persistence & Privilege
always is false; the skill does not request elevated persistence. It runs as an MCP server process and does not attempt to modify other skills or system-wide settings. Autonomous invocation by the agent is the platform default (disable-model-invocation is false) and is not by itself a red flag here.
Assessment
This skill appears to do what it says: query your HuahuaDaily cloud data and send transaction request signals using a PRO Agent Token. Before installing or running: 1) verify the upstream code (the README points to a GitHub repo) to ensure the server.py code matches what you expect; 2) confirm the API hostname mismatch (huahuadaily.preview... vs huahua.preview...)—ask the author or inspect the repo to ensure requests go to the official service; 3) treat BAIYE_AGENT_TOKEN like a secret: only set it for this service and don't paste it into unknown sites; 4) run the MCP server in an isolated environment if you have security concerns (it will make network requests to the service using your token); 5) remember that request_transaction only sends a request that requires confirmation in the official App—transactions are not executed directly by the skill.

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

Runtime requirements

🌸 Clawdis
OSmacOS · Linux · Windows
EnvBAIYE_AGENT_TOKEN
Primary envBAIYE_AGENT_TOKEN
latestvk9750ngt92pr7ctg6wy2nf0g5585cpv0
280downloads
1stars
12versions
Updated 5d ago
v1.2.0
MIT-0
macOS, Linux, Windows

花花日记助手 (HuahuaDaily)

1. Description

花花日记助手让 AI 直接读取用户在花花日记 App 中的基金持仓数据,提供持仓收益查询、市场行情查看、基金搜索等功能。所有工具均需 Agent Token(花花日记 PRO 会员专属),数据来自云端同步。

交易操作(买入/卖出)不直接执行,AI 通过 request_transaction 发送信号,由用户在 App 中一键确认后生效。

2. When to use

  • 用户说:"我今天收益多少?"、"帮我看一下我的持仓"、"我的总资产是多少?"
  • 用户说:"易方达蓝筹现在涨了多少?"、"帮我查一下 110022 的净值"
  • 用户说:"今天大盘怎么样?"、"沪深300涨跌如何?"
  • 用户说:"帮我搜一下华夏基金"、"查一下这只基金的费率"
  • 用户说:"帮我看看这只基金近一年排名"、"今天走势怎么样?"
  • 用户说:"帮我看看夜盘基金"、"夜盘现在怎么样"、"我夜盘列表里哪些在涨"
  • 用户说:"我想买 1000 元易方达消费"、"帮我发一个卖出请求"
  • 用户说:"我的基金跑赢大盘了吗?"、"帮我对比一下沪深300"
  • 用户说:"今天下午买入,什么时候确认份额?"、"T+2 是哪天?"
  • 用户说:"中秋节后第一个交易日是哪天?"、"这个日期能买基金吗?"

3. How to use

  1. 首次使用:确认 BAIYE_AGENT_TOKEN 已通过环境变量配置;若未配置,引导用户前往 https://huahuadaily.preview.aliyun-zeabur.cn/ 注册账号并开通 PRO 会员,在「设置 → Agent Token」生成后配置到环境变量;

  2. 查询总览:优先调用 get_summary,返回总资产、今日收益、累计收益;如需明细再调用 get_records;返回结果中须将 dataUpdatedAt 同步时间告知用户(如"数据同步于 XX 时间"),让用户判断数据是否最新;

  3. 查询单只基金行情

    • 用户已提供代码(如"查 110022"):直接调用 get_item_estimate,无需先搜索;
    • 用户只给名称:先调用 search_item 找到 code,再调 get_item_estimate
    • 用户需要深度分析(历史、胜率、排名等):调用 get_item_detail(较慢,仅需深度信息时使用);
  4. 查询夜盘

    • 用户明确问夜盘估算:优先调用 get_night_est
    • 若用户未提供代码,则直接调用 get_night_est(),默认读取云同步中的夜盘关注列表;
    • 若需要先告诉用户当前夜盘列表有哪些基金,可先调用 get_night_watchlist
  5. 查询市场整体:调用 get_overviewget_indices

  6. 基准对比

    • 调用 get_benchmark_history("sh000300") 获取沪深300历史走势;
    • 结合 get_item_history(code) 的持仓基金历史,对比两者涨跌幅;
    • 常用指数代码:sh000300(沪深300)、sh000001(上证)、sz399001(深证成指);
  7. 买卖日期规划

    • 调用 get_next_trading_day(date) 确认某日是否为交易日及下一交易日;
    • 调用 calculate_trading_dates(date, time_mode, confirm_days) 推算净值日/确认日;
      • time_mode"PRE_MARKET"(收盘前买入)或 "POST_MARKET"(收盘后买入);
      • confirm_days:T+1(货币/债基),T+2(股票/混合),T+3(QDII);
  8. 发送交易信号

    • 若 code 来自搜索结果,须先向用户确认:「是【基金名 code:XXXXXX】吗?确认后我来发送请求。」
    • 用户确认后再调用 request_transaction
    • 调用完成后明确告知:「信号已发送,请打开 App 确认后才会生效,AI 不会直接执行交易。」
  9. 数据过时处理:若用户提到"刚刚刷新"或数据明显滞后,提示其在 App「设置 → 立即同步」手动上传后再查询;

  10. 结果复用:同一次对话中,一次查询结果可回答多个相关问题,不必重复调用同一工具。

4. 可用工具

认证

  • set_token — 运行时设置 Agent Token(推荐用环境变量,通常无需调用)

数据查询

  • search_item(query) — 按编号/名称搜索,返回最多 20 条(仅 code 未知时使用)
  • get_item_estimate(codes) — 批量获取今日实时估算净值(最多 50 个);日常查行情首选,轻量快速
  • get_night_watchlist() — 获取夜盘关注列表(来自最近一次云同步)
  • get_night_est(codes?) — 获取夜盘估算;不传代码时默认读取云同步中的夜盘关注列表
  • get_item_detail(code) — 基金深度信息(历史收益率、胜率分析、净值序列、费率);仅需深度分析时使用
  • get_item_history(code) — 历史净值走势
  • get_item_dividends(code) — 历史派息记录
  • get_fund_timeline(code) — 今日分时估值走势(非交易日返回空)
  • get_fund_fees(code) — 申购/赎回/管理/托管费率
  • get_fund_period_rank(code) — 近 1/3/6 月、1 年业绩及同类排名百分位
  • get_daily_rank() — 今日涨跌幅排行榜

概览数据

  • get_status() — 今日是否交易日
  • get_overview() — 市场整体概览(指数 + 板块 + 涨跌榜)
  • get_indices() — 主要指数实时数据(沪深、创业板、纳斯达克等)
  • get_benchmark_history(code?) — 指数/ETF 历史走势,默认沪深300(sh000300),用于基准对比
  • calculate_trading_dates(date, time_mode?, confirm_days?) — 推算 T+N 净值日/确认到账日
  • get_next_trading_day(date) — 获取指定日期起的下一个交易日(跳过周末和节假日)

记录管理(需 PRO 会员 + Agent Token)

  • get_current_user() — 当前账号信息
  • get_summary() — 持仓总览摘要(总市值/今日收益/累计收益/收益率),含 dataUpdatedAt
  • get_records() — 完整持仓明细(holdings 持仓 + watchlist 观察列),含实时收益计算和 dataUpdatedAt
  • request_transaction(item_code, item_name, record_type, amount, date?, note?, group_name?) — 向 App 发送买入/卖出请求信号

5. Edge cases

  • 未配置 Token:工具返回错误提示,引导用户前往 https://huahuadaily.preview.aliyun-zeabur.cn/ 注册/登录,开通 PRO 会员后在「设置 → Agent Token」生成并配置到环境变量 BAIYE_AGENT_TOKEN
  • Token 无效/过期:收到 401 错误,引导用户前往 https://huahuadaily.preview.aliyun-zeabur.cn/ 在「设置 → Agent Token」重新生成,并更新环境变量;
  • 数据过时(用户刚在 App 刷新):提醒用户在 App「设置 → 立即同步」手动上传最新数据,否则 MCP 仍读取旧的云同步数据;
  • 非交易日查询今日收益todayProfit 为 0,属正常现象,说明"今日未开市,显示的是官方净值下的持仓状态";
  • 基金估算为空(盘前/节假日):estimatedNav 为 null,todayProfit 为 0,属正常,无需报错;
  • 搜索无结果:引导用户确认基金编号或名称是否正确;
  • request_transaction 后:明确告知用户"信号已发送,请打开 App 确认后才会记录交易",不要误导用户认为已成功执行。

Comments

Loading comments...