# Custom Agent Instructions
This is an example `prompt.md` file. Copy this to `prompt.md` in your working directory to customize the AI assistant's behavior.
## Code Style Preferences
- Use clear, descriptive variable names (no single letters except for loops)
- Add docstrings/comments for all functions and classes
- Follow language-specific style guides:
- Python: PEP 8
- JavaScript: Airbnb Style Guide
- Rust: Official Rust Style Guide
## Project Conventions
- Use semantic commit messages (feat:, fix:, docs:, etc.)
- Write unit tests for all new functions
- Keep functions small and focused (single responsibility)
- Prefer composition over inheritance
## Communication Style
- Be concise but thorough
- Always explain your reasoning before taking action (ReAct pattern)
- Show code examples when explaining concepts
- Ask for clarification if requirements are ambiguous
## Technology Preferences
- **Web**: Prefer TypeScript over JavaScript
- **Backend**: Prefer async/await over callbacks
- **Testing**: Use pytest for Python, jest for JavaScript
- **Database**: Use prepared statements for SQL queries
## Security
- Never commit sensitive data (API keys, passwords, etc.)
- Validate all user inputs
- Use parameterized queries for database access
- Keep dependencies up to date
## Documentation
- Update README.md when adding new features
- Document all public APIs
- Include usage examples in documentation
- Keep CHANGELOG.md current
## Specific Instructions
[Add your project-specific instructions here]
Example:
- This project uses MongoDB, not PostgreSQL
- API responses should always include error codes
- All dates should be in ISO 8601 format
- Use Zod for runtime type validation in TypeScript