Add AI Agent development guide for automated enterprise software generation

[Copilot]

Enables developers to build AI agents that automatically generate enterprise management software (CRM, ERP, HRM, etc.) from natural language requirements using ObjectStack specifications.

Documentation Suite (62.8KB, Chinese)

content/docs/ai/README.md - Navigation hub

• Quick start patterns (CLI, API, online IDE)
• Learning paths (beginner → advanced)
• Architecture overview

content/docs/ai/agent-development-guide.md - Core reference

• 61 metadata types cataloged across 8 layers:
  • Data (ObjectQL): Object, Field (28 types), Validation, Hook, Query, Filter, Driver
  • UI (ObjectUI): View (7 types), Dashboard, Report, Action, Chart (10 types)
  • Automation: Workflow, Flow, Approval, ETL, Trigger
  • AI: Agent, RAG Pipeline, Model Registry, NLQ
  • System (ObjectOS): Manifest, Plugin, Events, Metrics, Cache, Job
  • Auth: Permission, RLS, Role, Territory
  • API: REST, GraphQL, WebSocket, Batch
  • Integration: Connectors
• 5-module agent architecture (NLP → Architecture Design → Code Gen → Validation → Iteration)
• 16-week implementation plan

content/docs/ai/metadata-checklist.md - Quick reference

• Structured lookup by layer with usage scenarios, code snippets, file locations
• Completion tracking: 50/61 (82%) implemented

content/docs/ai/implementation-examples.md - Working examples

• Complete inventory/sales/purchase management system generation
• Three integration patterns with runnable code

Example Usage

const agent = new ObjectStackAgent({
  llm: { provider: 'openai', model: 'gpt-4' }
});

const result = await agent.generate({
  requirement: `进销存管理系统包括:
    - 商品管理 (Product, Category, Brand)
    - 库存管理 (Warehouse, Inventory)
    - 采购/销售管理 (Purchase/Sales Orders)
    - 仪表板 (销售趋势, 库存预警)`
});

// Generates: 9 Objects, 3 Views, 1 Dashboard, TypeScript code

Design Decisions

• Zod-first metadata: Runtime validation + type inference for all 61 metadata types
• DDD alignment: Business entities map directly to ObjectStack Object schema
• Template-based generation: Ensures consistency with ObjectStack conventions (camelCase config, snake_case data)
• Multi-stage validation: Zod schema → TypeScript compilation → ESLint → business logic checks

Original prompt

我想开发一个agent，让他基于 @objectstack/spec 中规范去自我迭代开发企业管理软件。帮我整理跟业务开发相关的元素据清单和介绍。编写完整的设计方案和开发计划。

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
spec	Ready	Preview, Comment	Jan 31, 2026 5:45pm