用于Dynamics 365财务与运营X++开发的CLI工具包。索引AOT元数据，构建对象支架，并通过稳定的JSON信封和MCP连接AI代理。
概览
Dynamics 365 财务与运营的基础人工智能开发——支持 GitHub Copilot、Claude Code、Codex、Gemini CLI 以及任何带有 shell 的代理 AI助手擅长C#、Python和JavaScript。X++则不同：你的D365FO代码库是私密的，高度定制，且对所有模型都不可见——因此AI自信地生成不编译的代码。 这个CLI会预先索引你整个D365FO安装（涵盖标准、独立和自定义模型的数十万符号）到本地SQLite索引中，并以一个d365fo二进制文件的形式暴露。每个签名、每个CoC封装器、每个标签、每一个表单模式——经过与真实元数据核实后，AI只写一行。而且因为它是shell命令而不是MCP工具列表，每回合耗费~100token，而不是~1,800。 表单是最难正确生成的产物——每个模式都规定了所需的容器、顺序和允许的子模式。
概述（AI 翻译）
一个二进制文件，知道你D365FO代码库中所有X++类、表、表单和EDT
Dynamics 365 Finance & Operations 的基础人工智能开发——可支持 GitHub Copilot、Claude Code、Codex、Gemini CLI 及任何拥有 shell 的代理
为什么
AI助手擅长C#、Python和JavaScript。X++则不同：你的D365FO代码库是私密的，高度定制，且对所有模型都不可见——因此AI自信地生成不编译的代码。
这个CLI会预先索引你整个D365FO安装（涵盖标准、独立软件和自定义模型的数十万符号）到本地SQLite索引中，并作为一个d365fo二进制文件暴露。每个签名、每个CoC封装器、每个标签、每一个表单模式——都经过与你的真实元数据核实，**在AI写入一行之前。而且因为它是shell命令而不是MCP工具列表，每回合耗费~100个token，而不是~1,800。
| 任务 | 没有 d365fo |
带有 d365fo |
|---|---|---|
| 方法签名 | 猜测→编译错误 | d365fo get class —— 完全正确，来自你的代码库 |
| 现有的CoC封装器 | 手动AOT搜索 | d365fo find coc Class::method |
本页由 AI 翻译，技术细节请以原文为准。
README
以下为原文 README 节选，完整内容请查看原文。
d365fo — AI-native CLI for D365 F&O X++
Why
AI assistants excel at C#, Python, and JavaScript. X++ is different: your D365FO codebase is private, deeply customized, and invisible to every model — so AI confidently generates code that doesn’t compile.
This CLI pre-indexes your entire D365FO installation (hundreds of thousands of symbols across standard, ISV, and custom models) into a local SQLite index and exposes it as one d365fo binary. Every signature, every CoC wrapper, every label, every form pattern — verified against your real metadata before the AI writes a single line. And because it is a shell command instead of an MCP tool list, it costs ~100 tokens per turn instead of ~1,800.
| Task | Without d365fo |
With d365fo |
|---|---|---|
| Method signatures | Guessed → compile errors | d365fo get class — exact, from your codebase |
| Existing CoC wrappers | Manual AOT search | d365fo find coc Class::method in < 50 ms |
| New forms | Hand-written XML, broken patterns | Pattern-validated scaffolds; structural violations block the write |
| Labels | Hardcoded strings | Right @SYS/@MODULE key found instantly |
| Security chains | Hours of manual tracing | Role → Duty → Privilege → Entry Point in one call |
| Generated code | Hallucinated fields and types | Every reference proven against the index, gated before write |
| Agent context cost | 32 MCP tool schemas every turn (40 132 chars, ~11 100 tokens — measured) | 1 shell tool + lazy-loaded Skills (366 chars, ~100 tokens) |
Capabilities
| Feature | Description |
|---|---|
| 🔍 Full-codebase intelligence | Tables, classes, EDTs, enums, forms, queries, views, entities, reports, services, workflows, security artifacts, labels (FTS5) — results in milliseconds, no VM round-trip |
| 🛡️ Grounded generation | Fail-closed gates: prepare change/prepare create issue grounding tokens, validate references proves every identifier, validate xpp enforces BP rules — hallucinated code never reaches disk |
安装
This server does not publish a one-line install command.
Open the repository installation guide 新手？先看 mcpskill.org 教程 →配置
{
"mcpServers": {
"d365fo": {
"command": "d365fo-mcp",
"args": [],
"env": { "D365FO_PACKAGES_PATH": "K:\\AosService\\PackagesLocalDirectory" }
}
}
}