← Back to Skills Marketplace
kernelh

Alpha Pulse 1.0.0

by kernelh · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
106
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install alpha-pulse-1-0-0
Description
Generates top 30 high-potential A-share stocks for next-day short-term trading using multi-factor signals and automatic daily market scanning.
README (SKILL.md)

alpha-pulse — A股次日短线收益最大化信号引擎

🤖 由 Jarvis 构建 | 专为 T+1 短线交易设计 | 每日收盘后自动生成 30 只高潜力标的

✅ 核心能力

命令 作用
alpha-pulse scan 扫描全市场(~5000只股票),计算10+短线动因因子
alpha-pulse predict 输出次日涨幅概率 Top 30 股票(含信号分、动因摘要)
alpha-pulse report 生成 Markdown 报告 + CSV + 图表(可自动打开)
alpha-pulse notify 推送信号至 Windows 弹窗/剪贴板/语音

🔬 短线动因因子(T+1 专属)

  1. 资金面:龙虎榜净买入强度、北向尾盘30分钟流入占比
  2. 量价面:量比 > 3.0 + 涨停封单/成交额 > 0.5
  3. 技术面:5日线上穿10日线 + RSI(6) 从30以下拐头向上
  4. 消息面:当日公告关键词(重组/订单/新品)+ 股吧热度突增
  5. 情绪面:融资余额环比增长 > 2% + 融券余额下降

🛡️ 风控规则(自动启用)

  • 排除 ST/*ST、*退市风险警示
  • 流通市值 \x3C 50亿元 → 过滤
  • 单行业持仓 ≤ 3 只(防行业黑天鹅)
  • 信号分 \x3C 70 → 不入选

📁 目录结构

skills/alpha-pulse/
├── SKILL.md
├── lib/
│   ├── __init__.py
│   ├── scanner.py      # 数据获取(akshare 优先)
│   ├── factors.py      # 因子计算(向量化,高效)
│   ├── predictor.py    # 概率模型(含 demo 训练脚本)
│   └── filter.py       # 熔断逻辑
├── config.yaml         # 日期、阈值、token
└── examples/
    └── run_tomorrow.py # 主入口:今日收盘后运行,输出明日信号

⚙️ 首次使用

  1. 安装依赖:pip install akshare pandas numpy xgboost
  2. (可选)配置 tushare token(提升龙虎榜数据质量)
  3. 每日 15:30 后运行:python skills/alpha-pulse/examples/run_tomorrow.py

💡 提示:你只需说 alpha-pulse predict,我就会调用此技能生成信号。


下一步:我将立即创建 config.yamllib/scanner.py 骨架。
你无需操作——除非你想调整某条风控规则或因子权重

继续?
✅ 回复“继续” 或 “custom [需求]”

Usage Guidance
This package appears to be a scaffold rather than a complete feature: SKILL.md promises predictor, reporting and notification components that are not included. Before installing or running: (1) treat it as incomplete — expect to need additional code or scripts; (2) inspect any additional files the author supplies (predictor, factors, CLI wrappers) before running them; (3) run in an isolated environment if you must execute it (virtualenv/container); (4) do not provide credentials/tokens unless you confirm how they are used; and (5) ask the publisher for the missing modules or a clear README showing how the full pipeline is implemented.
Capability Assessment
Purpose & Capability
The name/description claim a complete T+1 signal engine that produces Top‑30 predictions, reports and notifications. The repository only provides a scanner skeleton (lib/scanner.py) and config.yaml; SKILL.md refers to factors.py, predictor.py, filter.py, examples/run_tomorrow.py and CLI commands (alpha-pulse predict/report/notify) that are not present. This is an important mismatch between claimed capability and actual included artifacts.
Instruction Scope
SKILL.md instructs users to run commands (alpha-pulse predict, alpha-pulse scan, python examples/run_tomorrow.py) and to install model packages (xgboost) and optionally configure a tushare token. The code only implements market list and daily-kline retrieval via akshare. The instructions therefore overreach the actual code and could mislead a user or an agent into invoking non-existent CLI entrypoints or assuming model/reporting behavior that isn't delivered.
Install Mechanism
There is no install spec (instruction-only skill). The README suggests installing common Python packages via pip (akshare, pandas, numpy, xgboost). That is proportionate and expected for the stated domain; no downloads from untrusted URLs or archive extraction are present.
Credentials
The skill declares no required environment variables or secrets. SKILL.md mentions an optional tushare token to 'improve' certain data, but config.yaml does not include a token field. No credentials are requested, which is proportionate; be aware that optional tokens (tushare) are mentioned but not integrated in the included code.
Persistence & Privilege
The skill is not always-enabled and does not request elevated persistence. It does not modify other skills or system configuration in the provided files.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install alpha-pulse-1-0-0
  3. After installation, invoke the skill by name or use /alpha-pulse-1-0-0
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of alpha-pulse — an A-share T+1 short-term signal engine. - Scans the entire market (~5000 stocks) daily using 10+ short-term drivers. - Ranks and predicts next-day top 30 stocks by potential return, providing signal scores and driver summaries. - Generates markdown reports, CSVs, and charts, with automated Windows notifications, clipboard copy, or voice alerts. - Built-in risk controls: industry concentration, market cap, ST/under-special-treatment/exit warning exclusions, and signal score filtering. - Modular directory with scanner, factor calculation, prediction model, and filter logic for flexible extension. - Easy setup—run a single script post-close to output tomorrow’s signals.
Metadata
Slug alpha-pulse-1-0-0
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Alpha Pulse 1.0.0?

Generates top 30 high-potential A-share stocks for next-day short-term trading using multi-factor signals and automatic daily market scanning. It is an AI Agent Skill for Claude Code / OpenClaw, with 106 downloads so far.

How do I install Alpha Pulse 1.0.0?

Run "/install alpha-pulse-1-0-0" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Alpha Pulse 1.0.0 free?

Yes, Alpha Pulse 1.0.0 is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Alpha Pulse 1.0.0 support?

Alpha Pulse 1.0.0 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Alpha Pulse 1.0.0?

It is built and maintained by kernelh (@kernelh); the current version is v1.0.0.

💬 Comments