The XTQuantAI server integrates the xtquant量化交易平台 (quantitative trading platform) with AI capabilities, enabling direct access and manipulation of trading data and functions. With this server, you can:
Data Queries: Get trading dates for specified markets, stock lists for specific sectors, and detailed information on instruments
Market Data: Retrieve historical, latest, and full market data for specified stocks
Visualization: Create chart panels with technical indicators and custom layouts with specific parameters for selected stocks
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@XTQuantAI get the latest market data for 000001.SZ"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
面向迅投 QMT / 投研终端的 Agent Skills 仓库：把量化策略开发的领域知识（因子回测、信号生成、实盘模板等）封装为标准 SKILL.md 技能，安装后可在 Claude Code、Cursor、Codex、Gemini CLI、Kimi Code 等 70+ AI 编程工具中直接使用。
本仓库曾是 xtquant 的 MCP 服务器实现，现已全面转型为 skill 仓库。旧 MCP 代码保留在
legacy-mcp分支（tagv0.1.0-mcp），不再维护。
技能 | 状态 | 说明 |
✅ 可用 | 根据策略描述 / 研报 PDF / 截图解读因子逻辑，基于母版脚本生成 QMT 内置日频截面因子回测策略（ | |
| 🚧 规划中 | 期货开平仓策略 |
| 🚧 规划中 | 目标持仓型实盘策略模板 |
| 🚧 规划中 | 信号生成 + 飞书推送 |
Related MCP server: okx-trade-mcp
这个项目靠社区需求驱动——你在 QMT 量化中遇到的重复劳动，都可能值得做成一个技能：
💡 想要新技能 → 提交技能需求，一句话的想法也欢迎
🐛 技能出问题 → 反馈 Bug
💬 使用咨询 / 量化思路交流 → Discussions
🔧 自己动手 → 阅读 贡献指南 后提 PR
所有技能在 CI 中自动按 Agent Skills 标准 校验。维护流程公开在 维护手册。
# 交互式选择要安装到的工具
npx skills add dfkai/xtquantai
# 只装某个技能
npx skills add dfkai/xtquantai --skill qmt-inner-backtest
# 装到全局目录（所有项目可用）
npx skills add dfkai/xtquantai -g/plugin marketplace add dfkai/xtquantai
/plugin install qmt-skills@xtquantai/plugins install https://github.com/dfkai/xtquantai对任意 AI 编程工具说：
读取 https://raw.githubusercontent.com/dfkai/xtquantai/master/INSTALL.md 并按其中说明安装技能。
把 skills/<技能名>/ 整个文件夹复制到你所用工具的技能目录：
工具 | 项目级目录 | 用户级目录 |
Claude Code |
|
|
Cursor (≥2.4) |
|
|
OpenAI Codex CLI |
|
|
GitHub Copilot |
| — |
Gemini CLI |
| — |
Kimi Code CLI |
| — |
Windsurf |
| — |
OpenCode |
| — |
Cline (≥3.48) |
| — |
字节 Trae | 仅 | — |
Qwen Code | 仅 | — |
iFlow CLI | 仅 | — |
跨工具通用目录是
.agents/skills/；只用 Claude 系工具可以放.claude/skills/。
安装 qmt-inner-backtest 后，在你的 AI 工具中直接说：
这份研报（附 PDF/截图）的上下影线因子，帮我生成一个中证1000、Top 10 持仓、5 日调仓的 QMT 回测脚本。
Agent 会先输出策略规格表供你确认，再基于母版生成完整的 QMT 策略编辑器回测脚本，并提示你需要在 QMT 面板中手动核对的配置项。
⚠️ QMT / 投研终端仅支持 Windows，生成的策略脚本需在 Windows 上的 QMT 策略编辑器中运行；skill 本身（生成脚本的过程）不限平台。
xtquantai/
├── skills/ # 技能目录（每个子目录一个 skill）
│ └── qmt-inner-backtest/
│ ├── SKILL.md # 技能定义（agentskills.io 标准格式）
│ └── scripts/
│ └── daily-factors-backtest.py # 回测母版脚本
├── .claude-plugin/
│ └── marketplace.json # Claude Code 插件市场清单
├── .github/ # issue/PR 模板与 CI 校验
├── kimi.plugin.json # Kimi Code 插件清单
├── INSTALL.md # 给 agent 读的安装指令
├── CONTRIBUTING.md # 贡献指南
└── docs/ # 调研文档与维护手册MIT — 详见 LICENSE。
Agent Skills 开放技能标准
创建图表面板，显示指定股票的技术指标
Inputs: codes (required), indicators, params, period.
创建自定义布局，可以指定指标名称、参数名和参数值
Inputs: codes (required), indicator_name, param_names, param_values, +1 more.
获取历史+最新行情数据
Inputs: codes (required), end_date, fields, period, +1 more.
获取历史行情数据
Inputs: codes (required), end_date, fields, period, +1 more.
获取指定股票的详细信息
Inputs: code (required), iscomplete.
获取最新行情数据
Inputs: codes (required), period.
获取指定板块的股票列表
Inputs: sector.
获取指定市场的交易日期列表
Inputs: market.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 8 tools
Most tools have distinct purposes, but there is notable overlap between get_full_market_data, get_history_market_data, and get_latest_market_data, which could cause confusion about which to use for specific data needs. The other tools target different resources like layouts, stock details, lists, and dates, making them clearer.
All tool names follow a consistent verb_noun pattern using snake_case, such as create_chart_panel, get_full_market_data, and get_instrument_detail. This predictability makes the tool set easy to navigate and understand.
With 8 tools, the count is well-scoped for a financial/quantitative AI server focused on market data, stock analysis, and visualization. Each tool appears to serve a specific function without being excessive or too sparse.
The tool set covers data retrieval, stock analysis, and visualization well, but there are gaps in operations like updating or deleting created layouts/panels, and no tools for executing trades or managing portfolios, which might be expected in a quantitative AI context. This could limit agent workflows.
Trading MCP server for AI agents, with live market data, account reads and controlled execution.
MCP server for Mudrex futures trading enabling AI agents to securely access data and risk tools.
Research-only MCP server: your AI as a quant research desk. 90 tools, no trades, no brokers.
Trade 16 crypto exchanges + MetaTrader 5 from your AI assistant via one MCP connection.
