nocobase-ai-employee

Configure AI employees, tools, skills, MCP servers, and LLM providers in NocoBase applications.

9|3|Updated Aug 14, 2026
One-click install
npx skills add https://github.com/nocobase/nocobase3 --skill nocobase-ai-employee-nocobase
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: nocobase-ai-employee
Source: https://github.com/nocobase/nocobase3/tree/main/packages/libs/ai-employee/.agents/skills/nocobase-ai-employee
Command: npx skills add https://github.com/nocobase/nocobase3 --skill nocobase-ai-employee-nocobase

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @nocobase/ai-employee, @nocobase/app-plugin-ai-employee, zod, @langchain/openai, and includes references (resource) components.

What problem does it solve? Developers building NocoBase applications with pnpm create @nocobase/app need a correct, contract-safe way to configure AI employees, tools, skills, MCP servers, and LLM services through @nocobase/ai-employee without modifying dependency internals or misusing manager APIs. ## Core Features & Use Cases - Declarative AI Resources: Place employees, tools, skills, and MCP definitions in the App's ai/ directory and LLM services in config.yml under ai.llmServices. - Programmatic Registration: Register dynamic tools, computed resources, and custom LLM providers through the shared deps.ai manager aggregate in an enabled plugin's server/bootstrap.ts. - Exact Contract References: Self-contained references define every manager signature, parameter type, enum casing, and error so coding agents never guess API shapes. - Use Case: Add a support-agent AI employee with a policy skill and a ticket-lookup tool, configure an OpenAI-compatible LLM service via environment variables, and verify everything with the App's lint, typecheck, and test scripts. ## Quick Start Ask the agent to add a new AI employee with a skill and tool to your NocoBase App created with pnpm create @nocobase/app, using the ai/ directory and config.yml for the LLM service.

Frequently Asked Questions about nocobase-ai-employee

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add an AI employee to a NocoBase application?▼

Create a TypeScript file under the App's server/ai/employees directory that default-exports defineAIEmployee with a unique username, skills, tools, and an inline systemPrompt. The enabled AI employee plugin loads it automatically at startup.

How do I configure LLM services in a NocoBase App?▼

Declare services in config.yml under ai.llmServices with name, provider, options, and enabledModels. The configuration is authoritative and live-reloadable, so invoke application config reload after editing without restarting or rebuilding.

Can I register a custom LLM provider without modifying @nocobase/ai-employee?▼

Yes. Implement LLMProvider and optionally EmbeddingProvider subclasses in an enabled App plugin, then register them synchronously in server/bootstrap.ts via deps.ai.llmProviderManager.registerLLMProvider. Reference the registry key in config.yml ai.llmServices.

When should I use registerDynamicTools instead of static tool files?▼

Use registerDynamicTools from plugin bootstrap when tools depend on request, session, role, or runtime context. Static files under server/ai/tools are better for fixed backend tools, since dynamic tools are built per query and never persisted.

Why does my MCP server fail to connect in NocoBase?▼

Common causes are omitting command for stdio transport or url for http/sse transport, or passing registration as a single options object instead of a name-keyed map. After changes, call rebuildClient() to refresh live connections and tools.

What are the limitations of this NocoBase AI employee approach?▼

It does not cover modifying the @nocobase/ai-employee dependency itself, building chat UI, or creating a second AIManager instance. Credentials must come from environment variables, and internal src imports are prohibited.