#Sub-Agents · Workflow Orchestration

Sub-agents are AI assistants specialized in generating data for specific modules. Each sub-agent is responsible for a distinct segment of the project — they each have their own responsibilities, linked together in a chain, collaborating to complete a full system build.

#What Are Sub-Agents

Sub-agents are specialized, clearly divided AI assistants:

• Functional Description Agent — Responsible for analyzing requirements and outlining the functional landscape
• Data Model Agent — Responsible for designing the database structure
• Function Menu Agent — Responsible for generating interface navigation
• Code Definition Agent — Responsible for establishing coding standards
• Backend Code Agent — Responsible for generating server-side code
• Frontend Code Agent — Responsible for generating the user interface

#Workflow Orchestration

Sub-agents are orchestrated through workflows. The platform provides a complete workflow engine — you can build a process for execution simply through drag-and-drop.

#Node Types

Process nodes are divided into three categories:

#General Control Nodes

#General Task Nodes

#Data Processing Nodes

Used to pass and transform data between nodes, supporting formats such as JSON, Markdown, HTML, etc.

#Sub-Agent Collaboration Mechanism

Sub-agents pass data to each other via Expression References, for example:

${Project Info.Functional Description}   → Gets the output of the functional description node
${Project Info.Data Model}   → Gets the output of the data model node

Benefits of this mechanism:

• Flexible Orchestration — Each sub-agent can reference the output of any preceding node as needed
• Multi-format Output — Supports multiple data formats such as JSON, Markdown, HTML
• Single Responsibility — Each sub-agent does only one thing, making optimization and replacement easy
• Constraints Upfront — The code definition stage establishes standards in advance, ensuring code quality