The Better Agent Structure ensures industry best practices, making your agent ready for production:
- Scenario agent tests written for every feature to ensure agent behaviour
- Versioning of the prompts for collaboration
- Evaluation notebooks for measuring specific prompts performance
- Already instrumented for full observability
- Standardization of structure for better project maintainability
my-agent-project/
├── app/ (or src/) # The actual agent code, structured according to the chosen framework
├── tests/
│ ├── evaluations/ # Jupyter notebooks for evaluations
│ │ └── example_eval.ipynb
│ └── scenarios/ # End-to-end scenario tests
│ └── example_scenario.test.{py,ts}
├── prompts/ # Versioned prompt files for team collaboration
│ └── sample_prompt.yaml
├── prompts.json # Prompt registry
├── .mcp.json # MCP server configuration (universal)
├── .cursor/
│ └── mcp.json # Symlink to ../.mcp.json for Cursor IDE
├── AGENTS.md # Development guidelines
├── CLAUDE.md # References AGENTS.md for Claude Code
├── .env # Environment variables
└── .gitignore
Development guidelines that ensure every new feature is properly tested, evaluated, and that prompts are versioned.
References AGENTS.md for Claude Code compatibility. Automatically created so the project works with any AI coding editor.
MCP server configuration that makes your coding assistant an expert in your chosen framework and provides access to additional tools. The .cursor/mcp.json is a symlink to this file for Cursor IDE compatibility.
End-to-end tests that simulate conversations with your agent to ensure it behaves as expected.
Jupyter notebooks for evaluating pieces of your agent pipeline such as RAG or classification tasks.
Versioned prompt files in YAML format, managed through prompts.json for team collaboration and playground integration.