Skip to content

trae-community/trae-agents

TRAE Agents

TRAE Agents Banner

License: MIT PRs Welcome

English | 中文

TRAE Agents is a community-maintained collection of custom agent configurations for the TRAE editor.

Here, developers share high-quality agent configuration schemes they've created, including agent names, prompts, tool configurations, and more. You can directly reference these configurations to quickly create your own agents in TRAE.

💡 Note: This project focuses on custom agent configuration schemes. If you're looking for MCP Servers (Tools), please check out TRAE MCPs.


📖 What is a Custom Agent?

A Custom Agent in TRAE is an AI assistant that you configure by setting specific prompts and tools, allowing the AI to specialize in tasks within a particular domain.

Imagine creating agents like these:

  • 👨‍💻 Code Review Expert: Reviews your code and identifies potential issues
  • 📝 Documentation Assistant: Helps you write technical documentation and comments
  • 🐛 Debug Expert: Systematically helps you locate and fix bugs
  • 🔧 Git Assistant: Generates standardized commit messages and manages branches

Each agent includes the following configurations:

  • 📋 Name: The agent's identifier
  • 💬 Prompt: Defines the agent's role, objectives, and workflow
  • 🛠️ Tools: Callable MCP services and other capabilities
  • 🤝 Collaboration: Other agents that can be invoked

🚀 Features

  • Ready to Use: Provides validated agent configuration schemes
  • Community Driven: Gathers wisdom from developers worldwide
  • Fully Transparent: Publishes all configuration details, including prompts and tool settings
  • Easy to Reuse: Clear import guide for quick creation in TRAE

📦 Agent List

Agent Name Description Status Documentation
git-commit-generator Generate Git commit messages following Conventional Commits specification ✅ Stable Configuration Details
(To be added) More agents under development... 🚧 -

(The repository is currently initializing. PRs to contribute your agent configurations are welcome!)

🛠️ How to Use

Step 1: Browse Agents

Browse the agent list above to find the agent you need.

Step 2: View Configuration Details

Click on the agent name to view complete configuration information, including:

  • Agent name
  • Full prompt text
  • Tool configuration
  • Usage examples

Step 3: Create in TRAE

3.1 Open TRAE

Launch the TRAE editor.

3.2 Access Agent Management

  • Click the agent icon (🤖) in the left sidebar
  • Select "Manage Agents" or "Create New Agent"

3.3 Fill in Configuration

Name: Enter the agent name, e.g., Git Commit Generator

Prompt: Copy the provided complete prompt content and paste it into the Prompt input box.

Tools: Check the recommended built-in tools based on recommendations:

  • File editing
  • Terminal commands
  • Web search
  • Browser automation

MCP Servers (Optional): If the agent requires MCP services, add them according to the provided JSON configuration.

Other Agents (Optional): If you need to invoke other agents, add their names.

3.4 Save and Test

  • Click "Save" to save
  • Return to the chat interface
  • Try testing the agent with inputs from the examples

Step 4: Adjust and Optimize

Based on actual usage experience, you can:

  • Adjust the wording of prompts
  • Add or remove tools
  • Modify constraints
  • Optimize workflows

🤝 How to Contribute

We welcome and appreciate community contributions! If you've configured a useful agent, please follow these steps to share:

  1. Read the Contribution Guidelines
  2. Fork this repository and create a new branch
  3. Create your agent configuration directory under the agents/ directory
  4. Provide complete configuration information following the template format
  5. Update the Agent List section
  6. Submit a Pull Request

📚 Agent Development Resources

📄 License

This project is licensed under the MIT License.


Star History

Star History Chart

Disclaimer

The agent configurations in this repository are provided for community exchange and learning purposes only. Please fully test and adjust them before using them in production or security-sensitive scenarios.

Links

About

AI coding agent practices | AI 编程 Agent 实践

Resources

License

Code of conduct

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors