## Brief overview
This document outlines the core guidelines for developing and maintaining the grocery-agent project. These rules cover documentation practices, code style, testing standards, and architectural preferences to ensure consistency, maintainability, and alignment with the project’s goals.
## Documentation Requirements
- Update `retail_agent/docs` directory whenever new features are implemented or existing ones are modified.
- Keep `retail_agent/README.md` synchronized with the latest capabilities, including usage examples and setup instructions.
- Maintain `CHANGELOG.md` by adding entries for every new feature, bug fix, or breaking change using the [Keep a Changelog](https://keepachangelog.com/en/1.0.0/) format.
## Testing Standards
- Write **unit tests** for all business logic, using `pytest`.
- All tests must be located in a dedicated `tests/` directory at the root.
- Any feature should have unit-test first, that describe all scenarios, then implementation should pass all those tests.
## Project Context
- This project is a grocery shopping assistant powered by an MCP server, integrating multiple grocery providers (Keshet, Rami Levy, Shufersal).
- The agent logic resides in `retail_agent/src/agent`.
- Tools developed as modular MCP servers and located under `/retail_agent/src/server`.
- Use `uv` as the Python package manager; ensure `uv.lock` is updated after any dependency changes.
## Other Guidelines
- Avoid long-running or complex prompts in agent logic; break down reasoning into small, composable steps.
- Use meaningful variable and function names; avoid abbreviations unless widely recognized.
- Do not commit to `main` directly; always use feature branches and open a PR for review.
- Prioritize clarity and maintainability over brevity in code and comments.
- In case of file creation, verify folder exists first, if not, create it
