Help Center/ CodeArts Agent/ User Guide (IDE)/ Agents/ Built-in Agents
Updated on 2026-07-29 GMT+08:00
View PDF
Share

Built-in Agents

Agents are programming assistants tailored for diverse development scenarios. CodeArts Agent provides multiple out-of-the-box agents that require no manual creation, aiming to provide efficient programming assistance for various development scenarios.

Built-in Primary Agents

Table 1 Built-in primary agents

Name

Creator

Description

Agent

System built-in

The core functional module of CodeArts Agent, supporting the understanding of complex developer requirements and autonomous execution path planning. It applies to end-to-end development scenarios (such as requirement analysis → code generation → test verification), achieving full-process automation to significantly boost R&D efficiency.

AgentTeam

System built-in

A multi-agent collaboration engine featuring the architecture of "Leader intelligent orchestration + Teammate autonomous execution." It decomposes complex tasks into parallel subtasks, allowing multiple specialized agents to collaborate like a real team. The entire process is event-driven and visualized, without manual intervention.

AgentTeam can only be used in CodeArts Agent Space. For details, see Multi-task Parallelism.

Intelligent Q&A

System built-in

An interactive knowledge assistant based on large language models, supporting questions in natural languages and accurate answers.

Built-in Sub-agents

In CodeArts Agent, a sub-agent acts as an AI assistant to the primary agent, designed to handle specific types of tasks and return the results to the primary agent. Sub-agents are primarily used in the following scenarios:

  • Long-chain complex tasks: Sub-agents are applicable to lengthy processes and complex logic, where they autonomously break down the task, plan steps, and implement them, such as full-chain code development.
  • Context Isolation: Sub-agents isolate sub-task contexts to prevent the primary agent from context redundancy, token bloating, and context overflow, such as in batch data parsing.
  • Vertical domain expertise: Each sub-agent can be configured with exclusive role prompts and private knowledge bases to develop domain expertise, enhancing inference precision and task accuracy, such as a security audit expert sub-agent.
  • Permission isolation: Sub-agents are configured with independent access permissions to achieve isolation for sensitive data, private knowledge bases, high- and low-risk interfaces, file resources, and project data, ensuring data and system security, such as permission isolation for high-risk operations like data deletion and system commands.
Table 2 Built-in sub-agents

Name

Description

general

A general-purpose agent used to research complex problems and execute multi-step tasks.

explore

A code repository exploration agent dedicated to quickly locating files, searching code keywords, and answering related questions.

spec-task-agent

An agent used to generate implementation tasks based on requirements and design.

spec-requirement-agent

An agent used to generate requirements for the Easy Approach to Requirements Syntax (EARS) format based on the project description and context.

spec-design-agent

An agent used to generate comprehensive technical designs, converting requirements (what to do) into architectures (how to do it).

developer-test-agent

An agent dedicated to unit test generation, repair, coverage optimization, and review.

Prompt-based calling: Enter a command directly into the chat box to invoke it. For example, use developer-test-agent to generate a unit test for a specific method in a specific file.

bug-fix-agent

An agent used to parse software issues from user bug reports and related artifacts, focusing on issue analysis, fault localization, and patch implementation.

rule-generator

Rule generator.

Modifying the Basic Model of a System Built-in Agent

Only enterprise administrators of the CodeArts Agent basic or professional edition can modify the basic model of a system built-in agent.

On the Agents tab of the CodeArts Agent console, you can view details about the current agent and flexibly adjust the underlying basic model based on actual service requirements to achieve optimal performance.

  1. Log in to the CodeArts Agent console.
  2. In the navigation pane on the left, choose Agent Settings > Agents.
  3. In the agent list, click Edit next to a built-in agent. The model editing page is displayed.
  4. In the basic model drop-down list, select the basic model of the agent and click OK.

    Figure 1 Modifying a basic model

  5. Return to the chat panel of CodeArts Agent IDE and select the corresponding built-in agent. Verify that the model has been modified.

    Figure 2 Modified basic models of the agent

Parent Topic: Agents