Quant Tools 1.0.0

v1.0.0

学术导向量化研究工具集。包含7大核心库(因子分析、组合优化、AI增强、因果验证、衍生品定价、回测引擎、情感分析)和5大投研工具(VeighNa交易框架、Qlib AI投研、WTP高性能框架、AkShare数据接口、JupyterHub研究环境)。适用于策略研发、因子挖掘、论文复现、资产配置、API服务化等投研任务...

0· 148·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for jiadong0723/jiadong-quant-tools.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Quant Tools 1.0.0" (jiadong0723/jiadong-quant-tools) from ClawHub.
Skill page: https://clawhub.ai/jiadong0723/jiadong-quant-tools
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 jiadong-quant-tools

ClawHub CLI

Package manager switcher

npx clawhub@latest install jiadong-quant-tools
Security Scan
Capability signals
Crypto
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
The name/description (quant research toolbox) matches the content: listed libraries (AlphaLens, VectorBT, PyPortfolioOpt, FinRL, FinBERT, EconML, QuantLib, VeighNa, Qlib, WonderTrader, AkShare, JupyterHub) and recommended architectures are consistent with a research/development toolkit.
Instruction Scope
SKILL.md is high-level documentation and examples for using the named libraries; it does not instruct the agent to read unrelated files, exfiltrate data, or call external endpoints beyond referencing public GitHub repos and recommending typical deployment patterns.
Install Mechanism
No install spec and no code files are included (instruction-only). There is nothing downloaded or written by the skill itself, which minimizes on-disk risk.
Credentials
The skill declares no required environment variables or credentials, which is internally consistent. Note: many recommended tools (VeighNa for live trading, JupyterHub, cloud training for FinBERT/VectorBT, or data sources) will require credentials, GPU access, network exposure, and other secrets when you actually deploy them — but those are not requested by this skill itself.
Persistence & Privilege
always is false and the skill is user-invocable; it does not request permanent presence or attempt to modify other skills or system-wide settings. Autonomous invocation is allowed by default but is not combined with any other elevated privileges in this package.
Assessment
This skill is a documentation/architecture guide pointing you to many legitimate open-source quant projects — it appears coherent and not malicious. Before using the guidance in production, consider: (1) provenance: the skill has no homepage and an unknown source — ask the author for provenance or a trusted repository/packaged release; (2) dependency risk: following the guide will lead to installing many third-party packages — pin versions, audit dependencies, and prefer official GitHub releases or package registries; (3) credential handling: live-trading frameworks (VeighNa, broker integrations), cloud GPUs, and data APIs require secrets — store them securely (vault, env with least privilege), and never paste them into chat; (4) network exposure: JupyterHub, FastAPI services, or trading gateways must be behind TLS, auth, and network isolation; (5) sandbox first: test heavy-resource tools (VectorBT, FinBERT) in isolated, resource-limited environments (Docker, k8s namespaces) to avoid accidental data leaks or runaway compute costs; (6) licensing & compliance: check each project's license and ensure your use (especially anything public-facing or paid advice) complies with regulations. If you want a deeper safety check, provide any planned deployment scripts, Dockerfiles, or a list of exact package versions you intend to install so those can be reviewed for risky install sources or suspicious post-install actions.

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

analysisvk9767wsynda03kf265r1by809s84hg1pdatavk9767wsynda03kf265r1by809s84hg1pfinancevk9767wsynda03kf265r1by809s84hg1platestvk9767wsynda03kf265r1by809s84hg1pquantvk9767wsynda03kf265r1by809s84hg1presearchvk9767wsynda03kf265r1by809s84hg1p
148downloads
0stars
1versions
Updated 2w ago
v1.0.0
MIT-0

量化投研工具集 (Quant Tools)

概述

本 skill 提供完整的量化投研工具链,覆盖:

  • 学术研究库 (7个): 算法创新、数学模型验证、理论实现
  • 投研工具 (5个): 交易执行、AI研究、数据获取、团队协作

学术研究库

1. AlphaLens - 因子分析

  • 定位: Quantopian 开源,经典因子评估框架
  • 功能: 因子 IC、IR、分层回测指标
  • 用途: 因子有效性检验、多因子模型评估
  • GitHub: https://github.com/quantopian/alphalens
  • API化: ⭐⭐⭐⭐ 极易封装为报告生成 API

2. VectorBT - 高性能回测

  • 定位: 基于 NumPy/Pandas 向量化计算
  • 功能: 参数网格搜索、敏感性分析
  • 用途: 大规模因子测试、快速验证
  • GitHub: https://github.com/polakowo/vectorbt
  • API化: ⭐⭐⭐⭐ 可封装回测引擎 API
  • 注意: 内存占用大,需计算优化型实例

3. PyPortfolioOpt - 组合优化

  • 定位: 现代投资组合理论 (MPT)
  • 功能: Black-Litterman、风险平价、权重优化
  • 用途: 资产配置研究、权重优化服务
  • GitHub: https://github.com/robertmartin8/PyPortfolioOpt
  • API化: ⭐⭐⭐⭐ 极易封装为计算 API

4. FinRL - 强化学习交易

  • 定位: 斯坦福/CMU 等多校联合
  • 功能: PPO、A2C 等 RL 算法,环境丰富
  • 用途: AI策略生成、强化学习论文复现
  • GitHub: https://github.com/AI4Finance-Foundation/FinRL
  • API化: ⭐⭐⭐ 需封装训练/推理接口
  • 注意: 训练不稳定,超参敏感

5. FinBERT - 金融情感分析

  • 定位: 基于 BERT 的金融领域微调
  • 功能: 金融语境情感分析
  • 用途: 舆情监控、新闻因子挖掘
  • GitHub: https://github.com/ProsusAI/finBERT
  • API化: ⭐⭐⭐ 需封装模型推理 API
  • 注意: 需要 GPU,推理延迟较高

6. EconML - 因果推断

  • 定位: 微软研究院
  • 功能: 因果效应分析、策略归因
  • 用途: 验证策略是否真的有效、避免过拟合
  • GitHub: https://github.com/microsoft/EconML
  • API化: ⭐⭐⭐⭐ 适合封装为分析 API
  • 重要性: 必须使用以验证策略有效性

7. QuantLib - 衍生品定价

  • 定位: 经典量化金融数学
  • 功能: 期权、利率、信用定价
  • 用途: 复杂结构化产品研究、风险模型验证
  • GitHub: https://github.com/lballabio/QuantLib
  • API化: ⭐⭐⭐ PyQL 封装后可提供 API
  • 注意: 学习曲线极陡

投研工具

1. VeighNa (vn.py) - 综合交易框架

  • 定位: 社区最活跃的综合交易与投研框架
  • 功能: 期货/股票/crypto接口、插件丰富
  • 用途: 实盘交易、策略回测、云端交易网关
  • GitHub: https://github.com/vnpy/vnpy
  • 部署: ⭐⭐⭐⭐⭐ 官方 Docker/K8s
  • API: ⭐⭐⭐⭐ 内置 RpcService,可封装 REST

2. Microsoft Qlib - AI量化投研

  • 定位: 微软开源 AI 量化平台
  • 功能: Transformer/LSTM、因子挖掘、流程标准化
  • 用途: 机器学习模型训练、Alpha研究
  • GitHub: https://github.com/microsoft/qlib
  • 部署: ⭐⭐⭐⭐⭐ 支持分布式训练
  • API: ⭐⭐⭐ 库形式,需自行封装 FastAPI

3. WonderTrader (WTP) - 高性能框架

  • 定位: C++核心高性能量化框架
  • 功能: Tick级回测、低延迟
  • 用途: 高频策略、对性能要求高的回测
  • GitHub: https://github.com/wondertrader/wondertrader
  • 部署: ⭐⭐⭐⭐ Docker支持,资源占用低
  • API: ⭐⭐⭐⭐ HTTP监控,易扩展微服务

4. AkShare - 金融数据接口

  • 定位: 完全免费的金融数据接口库
  • 功能: 宏观/基金/股票数据源覆盖极广
  • 用途: 数据获取、基本面数据清洗、统一数据源
  • GitHub: https://github.com/akfamily/akshare
  • 部署: ⭐⭐⭐⭐ 无状态,极易容器化
  • API: ⭐⭐⭐ 库形式,建议封装为数据网关 API
  • 注意: 依赖第三方网站稳定性

5. JupyterHub - 交互式研究环境

  • 定位: 分析师零成本上手
  • 功能: Notebook、团队协作
  • 用途: 数据探索、研究报告生成、团队协作开发
  • GitHub: https://github.com/jupyterhub/jupyterhub
  • 部署: ⭐⭐⭐⭐⭐ 原生支持 K8s 多用户
  • API: ⭐⭐⭐⭐ 可通过 Papermill/Voila 转 API

推荐组合方案

需求场景推荐组合
实盘交易 + 基础回测VeighNa
AI选股/因子研究Qlib + FastAPI
资产配置优化PyPortfolioOpt
因果验证(防过拟合)EconML
高频策略WonderTrader
数据中台AkShare + ClickHouse
团队协作研究JupyterHub + VeighNa/Qlib
完整投研平台学术库组合 + FastAPI封装

完整投研架构

数据层: AkShare
    ↓
因子分析层: AlphaLens + VectorBT
    ↓
策略优化层: PyPortfolioOpt
    ↓
AI增强层: FinRL / FinBERT
    ↓
因果验证层: EconML (必须!)
    ↓
服务封装: FastAPI + Docker
    ↓
可选: VeighNa (实盘) / Qlib (AI研究)

使用示例

因子有效性分析

使用 AlphaLens 分析 [因子名称] 的 IC、IR 表现

组合优化

使用 PyPortfolioOpt 基于 [风险偏好] 优化 [资产列表] 的权重

因果验证

使用 EconML 验证 [策略] 是否真的有效,用因果推断排除过拟合

回测验证

使用 VectorBT 对 [策略] 进行大规模参数扫描和敏感性分析

舆情因子

使用 FinBERT 分析 [公司/行业] 的新闻情感作为因子

强化学习策略

使用 FinRL 训练 [市场环境] 下的强化学习交易策略

数据获取

使用 AkShare 获取 [股票/宏观/基金] 的 [数据类型]

注意事项

过拟合风险

  • 学术库易在历史数据上完美表现,实盘失效
  • 务必使用 EconML 或出样本测试验证

计算资源

  • VectorBT 和 FinBERT 对资源要求高
  • 建议计算优化型实例

依赖管理

  • QuantLib/PyQL 环境复杂
  • 建议用 Conda 管理环境并打包 Docker

合规风险

  • 仅用于内部投研
  • 避免涉及公开荐股或非法经营证券业务

API 封装

  • 大多数工具是 Library 而非 Service
  • 需要用 FastAPI 进行封装
  • 网络安全:Nginx反向代理 + HTTPS + API Key认证

Comments

Loading comments...