🌐 中文文档：README.zh.md

A reusable DeepSeek Harness (DSH) skill that turns any DSH agent into a global-markets, multi-asset investment research expert. It covers equities, fixed income, commodities, FX, crypto, funds/ETFs and derivatives, and ships with macro / fundamental / technical / quantitative analysis frameworks plus a trading-strategy library. For research & education only — not investment advice.

---

📖 Introduction

Financial Expert Assistant is a reusable DSH skill that upgrades any DSH agent into an investment research expert across global markets and multiple asset classes. It packages an "analyst" capability into an on-demand instruction set (SKILL.md + a reference library), so the model, when answering finance/investment questions:

• Follows a top-down (macro → asset → industry → security) + bottom-up (micro fundamentals) cross-validation framework;
• Produces systematic analysis across equities, bonds, commodities, FX, crypto, funds/ETFs, derivatives and more;
• Covers fundamental, technical, quantitative, macro, event-driven, options and arbitrage strategies;
• Upholds a compliance baseline: separating facts from opinions, addressing risk before return, no return promises, and never constituting investment advice.

⚠️ This project is for research and education only and does not constitute investment advice. Past performance does not guarantee future results. Markets are risky — please exercise independent judgment and assume your own risk.

---

✨ Features

• 🧭 Full asset-class coverage: equities / fixed income / commodities / FX / crypto / funds·ETF·REITs / derivatives / alternatives.
• 🏗️ Four-layer analysis framework: macro (top-down), micro fundamentals (bottom-up), technical timing, and reproducible quantitative research.
• 📚 Trading-strategy library: value / growth / GARP / dividend, trend / momentum / mean reversion, multi-factor / statistical arbitrage, global macro allocation, event-driven, options, fixed-income and arbitrage strategies.
• 🛡️ Built-in risk & compliance: position sizing, stop-loss, drawdown, scenario stress-testing, plus disclaimers and no-stock-tipping guardrails.
• 🔌 Tool orchestration: web_search (live quotes/macro data), choice-quantapi-skill (quant data & backtesting), and a local quant-bot.html (visual backtesting demo).
• 🧩 Zero-dependency, drop-in: pure Markdown, organized per the DSH skill spec and auto-discovered by skill-filesystem.

---

🚀 Quick Start

Install

The skill is auto-discovered by DSH from several roots (directory-bundle form: <name>/SKILL.md):

Scope	Path	Priority
Project (DSH native)	<projectRoot>/.dsh/skills/financial-expert-assistant/	High
Project (compatible)	<projectRoot>/.agents/skills/financial-expert-assistant/	Medium
User global (DSH)	~/.dsh/skills/financial-expert-assistant/	Medium
User global (compatible)	~/.agents/skills/financial-expert-assistant/	Low

# Option 1: global install (recommended — available to all projects)
mkdir -p ~/.agents/skills
git clone https://github.com/Ricky-Sunny/financial-expert-assistant.git ~/.agents/skills/financial-expert-assistant

# Option 2: project-level install
mkdir -p .dsh/skills
cp -R <path-to-this-skill> .dsh/skills/financial-expert-assistant

DSH auto-detects new skills through filesystem watching (no restart needed); open a new session to see financial-expert-assistant in the skill catalog.

Usage

Ask any finance/investment question in a DSH session, or invoke the skill explicitly. For example:

• "Compare current valuations of A-shares vs. US equities."
• "Break down this company's ROE with DuPont analysis."
• "What are the key drivers of the gold price, and where are we now?"
• "Design a dual moving-average + stop-loss backtest and verify it on the local demo."

When the request involves quantitative data or backtesting, the skill guides loading choice-quantapi-skill to produce reproducible scripts; for visual demos it points to quant-bot.html.

---

🧠 Coverage

Asset class	Key analysis points
Equities	Financial quality, valuation (P/E·P/B·EV/EBITDA·PEG·DCF), sectors & styles
Fixed income	Yield curve, duration, credit spreads, convertible-bond terms
Commodities	Supply/demand balance, inventory cycle, basis/term structure, USD & real rates
FX	Rate differentials, interest-rate parity, central-bank policy, risk sentiment
Crypto	Halving cycle, on-chain data, ETF flows, regulation
Funds / ETF / REITs	Fees, tracking error, premium/discount, distributions & NAV
Derivatives	Futures margin, option Greeks, implied volatility

Strategy library: value / growth / GARP / dividend / indexing, trend following / momentum / mean reversion, multi-factor / statistical arbitrage / market-neutral, global macro / risk parity / all-weather, event-driven, options, fixed-income and arbitrage strategies.

---

📁 Project Structure

.
├── .dsh/skills/financial-expert-assistant/   # DSH skill (directory bundle)
│   ├── SKILL.md                              # Core: role / capabilities / framework / workflow / output / compliance
│   └── references/
│       ├── asset-classes.md                  # Knowledge base for 8 asset classes
│       ├── strategies.md                     # Trading & investment strategy library
│       └── risk-and-compliance.md            # Risk management framework + compliance guardrails
├── quant-bot.html                            # Optional: visual backtesting demo (CN/HK/US equities)
├── README.md                                 # This file (English)
├── README.zh.md                              # 中文文档
└── LICENSE

quant-bot.html is a companion educational backtesting demo (dual MA / RSI / Bollinger, powered by Eastmoney public quotes with synthetic-data fallback). It connects to no broker and involves no real funds; open it directly in a browser.

---

⚠️ Disclaimer

1. Output from this project is research, education and information sharing, and does not constitute investment advice, financial advice, a securities recommendation, or a solicitation to buy or sell.
2. No return promises — no prediction of certain gains, and no "guaranteed/risk-free/bottom/top" language.
3. Markets are risky; past performance does not indicate future results; leverage and derivatives can magnify losses.
4. Data and facts are sourced and dated where possible; judgments and forecasts are "opinions" — please verify independently.
5. We do not assist with insider trading, market manipulation, money laundering, regulatory evasion, or any illegal activity.

---

📄 License

MIT © financial-expert-assistant contributors

---

🙏 Related

• DeepSeek Harness (DSH) — the agent runtime that hosts this skill.
• choice-quantapi-skill — the companion quantitative data/backtesting skill (if installed).