What is the Claude.md file and what does it do?	Claude.md is a markdown file added to the root of your project that gives Claude code persistent memory about your project. Its contents are automatically appended to your prompt at the start of each session, acting as an onboarding script for your codebase.
What problem does Claude.md solve?	Without Claude.md, Claude code has to start fresh each session, re-explore the codebase, understand dependencies, and make assumptions. Claude.md prevents this by providing persistent context about your project.
How can you generate a Claude.md file automatically?	You can run the slash init command, which will make Claude generate a Claude.md file based on your codebase.
What are the three levels of Claude.md files in the hierarchy?	Project level (in root directory, shared with team), user level (in configuration folder, personal only), and file-level references (using @ symbol with file paths).
Should you share Claude.md in version control?	Yes, you should share the project-level Claude.md in your version control so your team can use it.
What should a Claude.md file typically include?	It should include your tech stack, coding preferences, commands, code style guidelines, and project structure conventions.
What does the user-level Claude.md file contain?	Personal preferences that apply across all your projects, such as how you write code comments and personal coding conventions.
What should you do if you have to correct Claude about a practice?	Explicitly ask Claude to save the correction to memory so it will remember that preference every time you return to the project.
What is the recommended approach for starting a new project with Claude.md?	Start without a Claude.md file so you can observe where you need to constantly course correct the model, then build the file compactly with only necessary information.
How does Claude.md impact the quality of Claude code sessions?	The difference between a frustrating and productive Claude code session comes down to context, and Claude.md is how you provide that context to the model.