Found a hidden gem of a GitHub repository: a complete Agent Skills toolkit built specifically for A-share quantitative trading!

The most tedious part of quant trading often isn’t writing the strategy—it’s gathering data, connecting to APIs, setting up a backtesting environment, and then stitching all those tools back together.

Recently discovered a highly practical open-source project: finance-quant-skills.

It organizes common capabilities for A-share quant trading into 13 installable Agent Skills:

• Data Acquisition: Multiple data sources including AKShare, BaoStock, Tushare, JoinQuant, and Wencai.
• Strategy Backtesting: Supports mainstream backtesting frameworks like Backtrader and RQAlpha.
• Strategy Development: Covers AKQuant, MiniQMT, and the quant strategy development workflow.
• Documentation Lookup: Built-in documentation for QMT, JoinQuant, etc., to reduce repetitive searching.
• Investment Research Output: Can organize company data and generate structured research reports.

After installation, there’s no need to memorize each API endpoint and field. You can directly describe tasks in natural language:

• Pull Kweichow Moutai’s financial data for the past three years and analyze changes in revenue, profit, and cash flow.
• Write a dual moving average strategy, backtest its performance over the last three years, and calculate the maximum drawdown and Sharpe ratio.
• Screen A-share companies with high ROE, low debt, and valuations at historical lows.

Repository address: https://github.com/lzwme/finance-quant-skills…

If interested, take a look—DYOR.

lzwme/finance-quant-skills

Source: https://github.com/lzwme/finance-quant-skills

Finance Quant Skills

A Claude Skills maintenance repository for the financial quantitative trading domain, built on the Agent Skills (https://agentskills.io) standard.

1. Directory Structure

This repository maintains Skills related to quantitative trading, covering scenarios such as quantitative strategy development, financial data analysis, and trading framework documentation lookup. Each Skill resides in its own folder, containing files like SKILL.md instructions.

finance-quant-skills/
├── skills/              # Collection of quantitative trading Skills
│   ├── akquant/         # AKQuant framework quantitative strategy development (event-driven, risk control & optimization)
│   ├── akshare/         # AKShare open-source financial data interface (stocks/futures/funds/crypto etc.)
│   ├── backtrader/      # Backtrader open-source quantitative backtesting framework
│   ├── baostock/        # BaoStock A-share data platform (free quotes, K-line, financial data)
│   ├── jqdatasdk/       # JoinQuant data interface (A-share quotes, financials, factor data)
│   ├── joinquant-docs/  # JoinQuant official strategy development documentation (backtesting/simulation/API/factors/technical indicators)
│   ├── miniqmt/         # MiniQMT XunTou quantitative trading interface (XtQuant, supports trade execution)
│   ├── pywencai/        # Tonghuashun Wencai data query (Chinese natural language query)
│   ├── qmt-docs/        # QMT strategy development guide and API reference documentation
│   ├── rqalpha/         # RQAlpha Mitrade open-source backtesting framework (A-shares/futures)
│   ├── equity-researcher/ # Institutional-grade research report generation (Official Kimi Skill)
│   ├── tdxquant/        # Tongdaxin quantitative data retrieval (quotes/K-line/financials/sector/formula execution)
│   └── tushare/         # Tushare data interface (A-share quotes, financials, macro data)
├── template/            # Skill template
└── .claude-plugin/      # Claude Code plugin configuration

2. Skill Installation

Please choose the appropriate installation method based on your use case:

2.1 Method 1: Let the AI Agent Help You Install

Tell the Agent assistant via conversation, such as OpenClaw:

Help me install all skills from this repository: https://github.com/lzwme/finance-quant-skills

Help me install the qmt-docs skill from this repository: https://github.com/lzwme/finance-quant-skills

2.2 Method 2: Install and Use in Claude Code

Register the repository as a Claude Code Plugin marketplace:

/plugin marketplace add lzwme/finance-quant-skills

Then install the specified Skills plugin:

/plugin install quant-skills@finance-quant-skills
/plugin install quant-docs@finance-quant-skills

You can also operate through the UI: Select Browse and install plugins → finance-quant-skills → Choose plugin → Install now.

After installation, simply describe your needs to trigger the corresponding Skill, for example: “Use the QMT Docs Skill to look up how to call the trading interface via Python.”

2.3 Method 3 (Recommended): Use npx skills Tool to Install Skills for the Assistant

# View help
npx skills --help

# View skills installed in the current project
npx skills list
# View globally installed skills
npx skills list -g
# Update installed skills
npx skills update

# Install skills from the current repository
npx skills add lzwme/finance-quant-skills

2.4 Installation Tip: Support Multiple Programming Assistant Tools via Symlinks

Different programming assistants have slightly different conventions for skills directories. When switching between multiple coding tools, you need to configure multiple skills directories.

Using symlinks and .gitignore configuration, you can link the skills directory to the skills directories of multiple programming tools, allowing you to maintain and update files only in the source directory.

Example commands to create symlinks:

# Assuming skills are maintained in agents/skills
# Create skills symlinks for .claude, .cursor

# On macOS/Linux
ln -s agents/skills .claude/skills
ln -s agents/skills .cursor/skills

# On Windows PowerShell (Junction can be replaced with SymbolicLink, but requires admin privileges)
New-Item -ItemType Junction -Path .claude/skills -Target <absolute_path>/agents/skills
New-Item -ItemType Junction -Path .cursor/skills -Target <absolute_path>/agents/skills

Example .gitignore configuration:

# Assuming skills are only maintained in .agents/skills
# Ignore skills symlinks in .claude, .cursor, etc. directories
.claude/skills
.cursor/skills
.codebuddy/skills

3. Skill Usage Examples

Detailed usage examples for the Skills can be found in USAGE_EXAMPLES.md, covering invocation scenarios for all 13 Skills across three main categories: data acquisition, strategy backtesting, strategy development & documentation.

3.1 Skills Overview

See USAGE_EXAMPLES.md for specific usage scenarios and parameter explanations.

4. Development

4.1 What are Skills?

Skills are folders composed of instructions, scripts, and resources that Claude dynamically loads to enhance its performance on specialized tasks. Skills teach Claude how to complete specific tasks in a repeatable way—such as writing trading programs following a quantitative strategy workflow, analyzing financial data, automating research report processing, etc.

• What are Skills? (https://support.claude.com/en/articles/12512176-what-are-skills)
• Using Skills in Claude (https://support.claude.com/en/articles/12512180-using-skills-in-claude)
• Creating Custom Skills (https://support.claude.com/en/articles/12512198-creating-custom-skills)
• Equipping agents for the real world with Agent Skills (https://anthropic.com/engineering/equipping-agents-for-the-real-world-with-agent-skills)

4.2 Creating a New Skill

Create a new folder under the skills/ directory and add a SKILL.md file along with related documentation or resource files. Template example:

---
name: my-quant-skill
description: Describe the function and applicable scenarios of this Skill
---

# Skill Name

[Write the instructions Claude should follow when activating this Skill here]

## Example
- Example usage 1
- Example usage 2

## Rules
- Rule 1
- Rule 2

Required Frontmatter fields:

• name — Unique identifier (lowercase, separated by hyphens)
• description — Functional description, explaining what the Skill does and when to use it.

5. Quantitative Finance Resource References

5.1 Quantitative Finance Related Skills

• East Money Miaoxiang:
  • https://ai.eastmoney.com/mxClaw
  • https://clawhub.ai/u/Financial-AI-Analyst
• Quantitative Related Skill Collection:
  • https://github.com/openclaw/skills/tree/main/skills/coderwpf
  • https://clawhub.ai/u/coderwpf
• JoinQuant Strategy Development Skill: https://clawhub.ai/daidaotian/joinquant-strategy
• Stock Quote Skill based on Futu OpenAPI: https://clawhub.ai/shuizhengqi1/futu-stock
• quantskills (https://github.com/quantskills) is an open quantitative community for the AI Agent era initiated by PandaAI (https://www.pandaaiquant.com/login?invite=1C91), focusing on two types of assets: Quant Skills and Agents.

High-Quality Skill Resources:

• anthropics/skills (https://github.com/anthropics/skills)
• The open agent skills tool - npx skills (https://github.com/vercel-labs/skills)
• https://skills.sh A Skills aggregation site launched by Vercel, containing a large number of open-source Skills, which can be quickly searched by 24-hour popularity, official certification, etc.
• http://clawhub.ai OpenClaw official Skills aggregation site.

5.2 Quantitative Trading Resource List

• Awesome Quant A curated list of high-quality quantitative trading resources for learning and use, for easy quick indexing and lookup.

License

This project is released under the MIT license.

This project is developed and maintained by Zhiwen Studio (https://lzw.me).

Alice (@BTCqzy1): Found an open-source A-share quant wonder tool—tickflow-stock-panel.

Stock screening, market watching, backtesting, AI review—all packed into one panel, without even needing to fuss with historical market data separately.

The project includes 18 built-in stock screening strategies, with common indicators like MA, MACD, RSI, KDJ, Bollinger Bands, and volume ratio ready to use directly to screen all A-shares.