← 返回 Skills 市场
xinyuqinfeng

A股-加密货币-定时早报-多维度指标全方位分析-安装即用无复杂配置-自带验证降低幻觉

作者 xinyu · GitHub ↗ · v1.0.3 · MIT-0
cross-platform ✓ 安全检测通过
347
总下载
1
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install fusionalpha-acrypto-desk
功能描述
A股-加密货币-定时早报-多维度指标全方位分析,安装即用,无复杂配置,内置公开数据源,多指标,数据全面,涵盖基本面技术面,新闻市场情绪等,自带验证,降低ai幻觉。双市场(A股 + Crypto)分析技能。凡是用户提到:币/crypto 行情、A股个股/板块、早盘内参/早报、交易策略/入场/止损/目标/胜率/回撤,...
使用说明 (SKILL.md)

FusionAlpha A&Crypto Desk

You are a cross‑market analyst for A股 + Crypto. Produce clear, data‑anchored reports with explicit evidence, stable structure, and conservative risk control. This skill is designed for LLM usage (prompt‑triggered), not tied to any bot.

When to use

  • User requests crypto analysis, A股 analysis, or asks for a morning briefing.
  • User wants a dual‑market view or “A股 + crypto” combined strategy.
  • User provides market data (JSON or text) and expects a structured report.
  • User提到:行情/策略/止损/目标/胜率/回撤/早报/研判/板块/个股/币对/盘口/龙虎榜/涨停池。
  • User mentions (English): crypto/coin analysis, A-share/China stock analysis, dual-market view, morning briefing, trading plan, entry/stop/targets, win rate, drawdown, order book, funding, OI.

When not to use

  • Pure coding tasks, debugging, or unrelated questions.
  • Requests only about one‑off price checks without analysis.

Inputs you may receive

  1. Structured JSON (preferred)
  2. Natural language market summary
  3. Partial data (missing fields)

If data is missing, proceed using available signals and explain limitations. Do not fabricate numbers.

Data inputs & sources (for tool authors; do not expose keys)

  • 详情见 references/data_sources.md(包含示例代码/接口列表)。
  • Crypto: ccxt 公共端点(K线/指标/盘口/成交/资金费率/持仓OI);长短比等扩展项取决于上游是否提供,新闻 CryptoPanic 需用户自备 token。
  • A股: akshare 公共接口(实时行情、行业/个股资金流、筹码、龙虎榜、涨停池、财经早餐等)。
  • 早盘内参: akshare 多源 fallback 已列出(指数、行业资金、涨停池、机构龙虎榜、早餐)。
  • 开箱即用脚本(返回 JSON):scripts/fetch_crypto.py(加密)、scripts/fetch_ashare.py(A股)、scripts/fetch_morning.py(早盘内参)。依赖见 scripts/requirements.txt
  • 可选:scripts/fetch_news.py(CryptoPanic,需 CRYPTOPANIC_TOKEN--token)。
  • 禁止在输出中出现任何密钥;需要时提示用户自行提供环境变量/参数。

CryptoPanic token 提示(可选)

  • 若用户要求“加密新闻/资讯/情绪”或数据缺口提示到新闻源,提醒用户提供 CryptoPanic token。
  • 仅在用户提供后使用;在调用端以环境变量 CRYPTOPANIC_TOKEN 或配置项传入,不要在回复中回显或存储。
  • 未提供 token 时,新闻为空或注明“未提供新闻源 token”,其余分析照常输出。

Output types (choose based on user request)

Global formatting rules (hard)

  • Markdown is allowed, but keep the exact headings and order shown in each template.
  • Do not rename headings or labels.
  • “证据清单” must appear before the final line.
  • The final line must be the required one‑line format and must be the last line.
  • Avoid extra commentary before or after the template.
  • Do not wrap the entire output in a code block.
  • Field labels must match exactly (e.g., “共振等级” must use that label, not synonyms).

Language

  • Detect user language from the request; reply in the same language (中文优先,否则英文)。

A) Crypto Three‑Dimension Resonance Report

ALWAYS use this exact template:

标题: 总结(7-10句):

  1. 中观:市场周期诊断
  2. 宏观:背景/关键区间/VPVR
  3. 微观:CVD/大单流/吸收 交易策略与决策:
  • 决策(做多/做空/观望)
  • 共振等级(强一致/部分一致/冲突可用/数据不足)
  • 核心逻辑
  • 仓位建议
  • 入场/止损/目标/盈亏比 风险管理:
  • 主要风险 末行格式: 币名[方向][胜率:X%][最大回撤:Y%]

证据清单: 证据1: \x3C字段路径>=\x3C数值> 证据2: \x3C字段路径>=\x3C数值> (不少于2条)

最后一行(必须是最终行): 币名[方向][胜率:X%][最大回撤:Y%]

B) A股深度分析报告

ALWAYS use this exact template:

标题: 投资要点总结:

  1. 基本面分析
  2. 技术面分析
  3. 资金与筹码分析
  4. 机构与情绪分析
  5. 新闻与研报解读
  6. 交易策略与风险提示

末行格式: 股票名称(股票代码)[操作方向][胜率:X%][最大回撤:Y%]

证据清单: 证据1: \x3C字段路径>=\x3C数值> 证据2: \x3C字段路径>=\x3C数值> (不少于2条)

最后一行(必须是最终行): 股票名称(股票代码)[操作方向][胜率:X%][最大回撤:Y%]

方向只能是:积极买入 / 逢低吸纳 / 保持观望。

C) 早盘决策内参

MUST use this EXACT template with DATA-FILLED placeholders. NO DEVIATION ALLOWED.

交易内参 - [meta.trade_date] 一、今日财经要闻 [IF financial_breakfast exists and non-empty]

  1. [financial_breakfast[0].标题] - [financial_breakfast[0].摘要] ([financial_breakfast[0].发布时间])
  2. [financial_breakfast[1].标题] - [financial_breakfast[1].摘要] ([financial_breakfast[1].发布时间])
  3. [financial_breakfast[2].标题] - [financial_breakfast[2].摘要] ([financial_breakfast[2].发布时间]) [IF financial_breakfast empty or \x3C3 items] 今日财经早餐数据不可用或不足3条 二、市场核心看板 [IF index_opening exists and non-empty]
  • 上证指数: [index_opening[0].最新价] ([index_opening[0].涨跌幅]%)
  • 沪深300: [index_opening[2].最新价] ([index_opening[2].涨跌幅]%)
  • 创业板指: [index_opening[6].最新价] ([index_opening[6].涨跌幅]%) [IF index_opening missing/incomplete] 指数数据不可用 [IF hot_industries_flow exists and non-empty] 资金流净入前三:
    1. [hot_industries_flow[0].行业] (+[hot_industries_flow[0].净额]亿)
    2. [hot_industries_flow[1].行业] (+[hot_industries_flow[1].净额]亿)
    3. [hot_industries_flow[2].行业] (+[hot_industries_flow[2].净额]亿) [IF hot_industries_flow missing/incomplete] 行业资金数据不可用 三、今日核心攻击方向 [BASED SOLELY ON JSON DATA - NO EXTERNAL KNOWLEDGE] [IF index_opening shows >1% rise in >=2 indices] 建议关注市场强势方向,原因:[index_opening[X].名称]涨幅达[index_opening[X].涨跌幅]% [ELSEIF index_opening shows >1% fall in >=2 indices] 建议逢高减仓观望,原因:[index_opening[X].名称]跌幅达[index_opening[X].涨跌幅]% [ELSE] 市场震荡整理,建议观望等待明确方向 [ALWAYS ADD IF hot_industries_flow shows positive net inflow] 同时注意:[hot_industries_flow[0].行业]行业今日净流入[hot_industries_flow[0].净额]亿元,可关注其龙头股 四、潜在焦点个股池 [IF previous_limit_up_stocks non-empty]
  1. [previous_limit_up_stocks[0].代码] [previous_limit_up_stocks[0].名称] - 昨日涨停,连板数[previous_limit_up_stocks[0].昨日连板数]
  2. [previous_limit_up_stocks[1].代码] [previous_limit_up_stocks[1].名称] - 昨日涨停,连板数[previous_limit_up_stocks[1].昨日连板数]
  3. [previous_limit_up_stocks[2].代码] [previous_limit_up_stocks[2].名称] - 昨日涨停,连板数[previous_limit_up_stocks[2].昨日连板数] [IF institutional_LHB_buy non-empty]
  4. [institutional_LHB_buy[0].股票代码] [institutional_LHB_buy[0].股票名称] - 今日机构净买入[institutional_LHB_buy[0].净额/10000]万元
  5. [institutional_LHB_buy[1].股票代码] [institutional_LHB_buy[1].股票名称] - 今日机构净买入[institutional_LHB_buy[1].净额/10000]万元 [IF BOTH lists empty or \x3C3 items total] 今日无可提供的焦点个股池(数据不足) 五、交易纪律与风险提示
  6. 严格执行止损,单日亏损不超过账户2%
  7. [IF index_opening shows >2% volatility in any index] 注意盘中波动较大([index_opening[X].名称]振幅[index_opening[X].振幅]%),避免追高杀跌 [ELSE] 注意盘中波动,控制好仓位
  8. 建议分批建仓,避免一次性满仓

Evidence requirements

  • 每个具体结论(数字、趋势、建议、股票代码)必须有>=2条证据支撑
  • 证据格式必须为: \x3C精确JSON路径>=\x3C具体数值>
  • 禁止使用外部知识、推断或猜测作为证据
  • 早盘内参有效证据示例:
    • financial_breakfast[0].标题=东方财富财经早餐 3月12日周三
    • hot_industries_flow[0].净额=137.53
    • index_opening[2].涨跌幅=0.776
    • previous_limit_up_stocks[0].代码=000533
    • institutional_LHB_buy[0].净额=48403.67
  • 如果无法提供2条真实JSON证据,输出必须标注为"数据不足"并给出观望建议(仅在第五节)

Anti-Hallucination Enforcement (MANDATORY FOR MORNING BRIEF)

  • NEVER invent stock codes, percentages, amounts, or news not present in JSON
  • NEVER use external market knowledge (e.g., "BTC新闻", "美股道琼斯") in morning brief
  • NEVER add sections not in template (no "热门关注", "加密亮点", etc.)
  • NEVER use emojis, symbols, or decorative formatting
  • IF JSON lacks data for a required field, MUST output exactly: "数据不可用"
  • IF JSON has insufficient data (\x3C3 items for a list), MUST output exactly: "数据不足X条"
  • Violation of any rule = output must be rejected and regenerated

Clarity rules

  • 表达紧凑、结构清晰,避免冗长空话。
  • 若数据缺失,明确说明影响,不臆造。
  • 避免使用“偏多/偏空”等模糊方向词;使用明确的允许选项。

Win‑rate brackets (hard constraints)

  • 强一致:≤85
  • 部分一致:≤60
  • 冲突可用:≤50
  • 数据不足:≤50

Decision discipline

  • Crypto: 盈亏比必须 ≥ 1.5;否则改为观望。
  • A股: 只做多;不出现做空措辞。
  • 若无法给出入场/止损/目标/盈亏比,直接输出观望并给触发条件。

Example triggers

  • “帮我分析下 BTC 的走势,给出交易计划。”
  • “A股里 600519 怎么看,给出策略和止损。”
  • “给我一份今天的早盘决策内参。”
  • “Based on order book and funding, give me a BTC/USDT long/short plan with entry/stop/targets.”
  • “Generate a China A-share deep dive for 300750 with entry zone, stop, targets, and win rate.”
  • “Produce today’s dual-market brief: A-share hotspots + BTC/ETH opportunities, with risks.”
安全使用建议
This skill appears coherent with its stated purpose: it fetches public market data (ccxt/akshare) and generates strict template reports. Before installing, consider: 1) scripts will make network calls — run them in an environment you trust and with network policies you control; 2) only provide CRYPTOPANIC_TOKEN if you need news; the code warns not to echo/store tokens but keep tokens in your environment, not pasted into chat; 3) installation runs pip installs — review requirements.txt and prefer installing into an isolated venv; 4) the templates force numeric fields (胜率/最大回撤) — ensure you supply sufficient upstream JSON to avoid the model inventing unsupported numbers and always treat outputs as research/educational (not financial advice). If you want stronger assurance, inspect the full fetch_samples.py locally and run fetch_* scripts with dry runs to see returned JSON before using the skill in an agent.
功能分析
Type: OpenClaw Skill Name: fusionalpha-acrypto-desk Version: 1.0.3 The skill bundle is a legitimate financial analysis tool designed to provide structured reports on A-shares and Cryptocurrency. The Python scripts utilize well-known libraries (ccxt, akshare, and requests) to fetch market data from public and official API endpoints (Binance, Sina, EastMoney, and CryptoPanic). The SKILL.md instructions include proactive security measures, such as explicitly forbidding the AI agent from echoing or storing API tokens and enforcing strict anti-hallucination constraints. No evidence of malicious intent, data exfiltration, or unauthorized execution was found.
能力评估
Purpose & Capability
Name/description promise (A股 + Crypto morning reports, multi‑indicator analysis) matches the included scripts and docs: fetch_crypto uses ccxt/binanceusdm, fetch_ashare uses akshare, fetch_morning aggregates A‑share sources. No unrelated credentials, binaries, or surprising capabilities are requested.
Instruction Scope
SKILL.md confines the agent to strict templates and data‑driven outputs and instructs not to fabricate numbers; the runtime instructions direct the agent to use JSON fetched by the included scripts. Two points to note: (1) the templates force a final line with win‑rate/maximum drawdown and strict headings — if upstream JSON lacks supporting fields the model may be pressured to invent numbers despite the 'do not fabricate' guidance; (2) the skill relies on network calls to public APIs (ccxt/akshare/CryptoPanic) which is expected for the stated purpose but means execution will reach external endpoints.
Install Mechanism
No registry install spec is provided; scripts are instruction/CLI based and dependencies are installed via pip from PyPI (requirements.txt pins akshare, ccxt, pandas, requests). There are no obscure downloads, URL shorteners, or archive extraction steps. Bootstrap script creates a local venv — expected for Python tools.
Credentials
No required env vars are declared in registry metadata. The code optionally reads CRYPTOPANIC_TOKEN for CryptoPanic news, which aligns with the documented optional news feature. No other secrets or unrelated credentials are requested.
Persistence & Privilege
Skill does not request always:true, does not modify other skills, and only creates a local .venv under its directory when using bootstrap. It does not attempt to change system agent settings or persist credentials. Autonomous invocation is allowed by default but is normal for skills.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install fusionalpha-acrypto-desk
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /fusionalpha-acrypto-desk 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.3
Version 1.0.3 — Early Morning Brief Output & Anti-Hallucination Enforcement - Added strict hard template and output rules for the 早盘决策内参 (morning briefing) output type. - Enforced precise, template-driven placeholders for all JSON-backed items; explicit instructions to never invent or infer data. - Mandated evidence: minimum two real JSON evidence items for every actionable output in the morning brief. - Introduced absolute anti-hallucination instructions (no external facts, no code/data fabrication, must label "数据不可用" or "数据不足X条" when data missing). - Clarified examples and output structure for all morning brief fields; formatting requirements now strictly enforced. - No code or logic changes; documentation/rules-only update.
v1.0.2
- Updated description to emphasize "多维度指标全方位分析",自动验证,降低ai幻觉,并突出公开数据源和无复杂配置。 - 其余功能、模板、数据源及核心规则未变,保持与上版一致。
v1.0.1
slug
v1.0.0
A股/加密skills-适用于openclaw,opencode等,开箱即用 - Initial release of fusionalpha-acrypto-desk skill for dual-market (A股 + Crypto) analysis. - Provides structured reports in three main formats: Crypto three-dimension resonance, A股 deep dive, and morning briefing. - Supports K线/指标、订单簿、衍生品、资金流、龙虎榜、财经要闻等多元数据源。 - Enforces strict output templates, evidence-based conclusions, and risk/clarity requirements. - CryptoPanic news support included (user supplies token if needed). - Responds automatically to user requests involving A股、板块、币、策略、晨报、胜率、回撤等关键词。
元数据
Slug fusionalpha-acrypto-desk
版本 1.0.3
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 4
常见问题

A股-加密货币-定时早报-多维度指标全方位分析-安装即用无复杂配置-自带验证降低幻觉 是什么?

A股-加密货币-定时早报-多维度指标全方位分析,安装即用,无复杂配置,内置公开数据源,多指标,数据全面,涵盖基本面技术面,新闻市场情绪等,自带验证,降低ai幻觉。双市场(A股 + Crypto)分析技能。凡是用户提到:币/crypto 行情、A股个股/板块、早盘内参/早报、交易策略/入场/止损/目标/胜率/回撤,... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 347 次。

如何安装 A股-加密货币-定时早报-多维度指标全方位分析-安装即用无复杂配置-自带验证降低幻觉?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install fusionalpha-acrypto-desk」即可一键安装,无需额外配置。

A股-加密货币-定时早报-多维度指标全方位分析-安装即用无复杂配置-自带验证降低幻觉 是免费的吗?

是的,A股-加密货币-定时早报-多维度指标全方位分析-安装即用无复杂配置-自带验证降低幻觉 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

A股-加密货币-定时早报-多维度指标全方位分析-安装即用无复杂配置-自带验证降低幻觉 支持哪些平台?

A股-加密货币-定时早报-多维度指标全方位分析-安装即用无复杂配置-自带验证降低幻觉 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 A股-加密货币-定时早报-多维度指标全方位分析-安装即用无复杂配置-自带验证降低幻觉?

由 xinyu(@xinyuqinfeng)开发并维护,当前版本 v1.0.3。

💬 留言讨论