Functions

Project-Level Code Generation

• Powered by cross-file context, CodeArts Agent auto-generates code in multiple popular programming languages.
• By reading intent from line and function comments, it delivers real-time completions, creating a smoother coding experience.

For more information, see Agent Chats.

Code Completion

• CodeArts Agent anticipates your programming goals by analyzing the surrounding code context within the editor and delivers intelligent and tailored code recommendations instantly.
• Code completion can be triggered either automatically or using shortcuts.

R&D Intelligent Q&A

• CodeArts Agent delivers end-to-end (E2E) support for coding, testing, and release. It acts as a conversational sidekick that keeps your team in flow by bringing the right project knowledge when you need it. As the conversation continues, it maintains multi-turn context, cutting time-to-answer.
• It allows connection to your organization's technical documentations and code repositories, along with individual knowledge bases, to deliver answers tailored to your business needs.

For more information, see Intelligent Q&A.

Unit Test Case Generation

The Unit Test is an end-to-end test case generation agent that provides core capabilities throughout the entire process, including test design, test case generation, and intelligent test case repair.

For more information, see Unit Test (Agent).

Code Repo Indexing

CodeArts Agent grasps code repo intricacies with precision, meticulously parsing and indexing them to enhance contextual comprehension and dramatically boost code retrieval precision.

For more information, see Repository Index.

Code Explanation

CodeArts Agent walks you through complex code, breaking it down by function, purpose, application scenario, and main logic. The clear explanations help you quickly familiarize yourself with the codebase, enhancing information flows, and keeping the whole team in sync.

For more information, see Intelligent Q&A.

Code Commenting

• CodeArts Agent analyzes the code's structure and logic, and then auto-generates meaningful comments to improve readability and reduce the effort needed to understand the code.
• With less time spent in manually commenting on the code, you have more time to focus on core logic. Down the line, maintenance and bug fixes become smoother and faster, boosting R&D efficiency across the board.

Code Optimization

CodeArts Agent optimizes code across multiple major languages. It intelligently identifies potential code defects and suggests fixes, significantly enhancing code quality and security.

Skills

Skills reshape an agent by eliminating traditional limitations. They allow an agent to move beyond just "thought processing" and engage in actual conversation. CodeArts Agent understands the given skills and executes tasks autonomously. It seamlessly translates interactive commands or service demands into tangible outcomes, bridging the gap between linguistic input and visual action for comprehensive solutions.

CodeArts Agent has a skill market, offering a rich variety of out-of-the-box standardized skills. You can invoke these skills without any manual configuration required, significantly reducing integration costs and improving code development efficiency.

For more details, see Skills.

Hooks

With hooks, you can insert custom logic into key lifecycle nodes of the CodeArts Agent IDE and plugins to extend functionality without modifying existing code. Unlike prompts which rely on model understanding, hooks yield deterministic outcomes. Once an event is triggered, the designated script executes reliably and consistently, without being affected by model understanding discrepancies.

For more details, see Hooks.

Slash Commands

Slash commands are an efficient way to interact with your agent. They allow you to quickly give the Agent instructions for common tasks within a conversation. CodeArts Agent comes preconfigured with some common commands. You can also create commands as needed.

For more details, see Slash Commands.

Memory

Memory allows CodeArts Agent to adapt to how you work. The Agent will record your habits and preferences, such as your preferred answer style and language. Once enabled, the agent will learn your preferences based on your historical sessions and automatically apply these settings to subsequent conversations, providing you with a more tailored service experience.

For more details, see Memory.

Checkpoints

CodeArts Agent checkpoints are automatic code snapshots that allow for rollbacks. The system automatically tracks code edits when the Agent is running. Every time the Agent prepares to modify code, the system will first automatically save the current code and dialog status, which enables quick cancellation, rollback, and historical version recovery. This function is useful for AI-assisted development scenarios that require quick trial-and-error, secure editing, and one-click rollback.

For more details, see Checkpoints

Multi-Session Management

In traditional development, project sessions must be executed one by one. Only one session can be executed at a time, which is inefficient. With the multi-session management function, the system can execute multiple sessions at the same time, improving the overall development efficiency.

For more details, see Multi-Chat Management.

Sandbox

A sandbox is a clean runtime environment used to isolate high-risk operations and prevent unauthorized resource access. In a sandbox environment, risky commands require user confirmation before execution. In addition, sensitive directories can be isolated to prevent data leaks. Sandboxes help you comply with security requirements of industries such as finance, healthcare, and government.

For more details, see Sandbox.