What is Claude's context window?	The amount of memory space Claude can hold, which includes files read, commands run, messages sent, and tool call results.
What happens automatically when the context window approaches its limit?	The context window is automatically compacted, which summarizes important details and removes unnecessary tool call results to free up space.
What is the difference between the /compact and /clear commands?	/compact summarizes previous work and preserves memory of what was done; /clear removes everything to start completely from scratch without previous context.
What command shows you the current state and breakdown of your context window usage?	The /context command displays the context size, categories using the most space, and a graphic showing the breakdown.
When should you use /compact versus /clear during development?	Use /compact when continuing work on the same feature; use /clear when starting a new feature to avoid bias from previous conversations.
Why is being specific in prompts important for context optimization?	Being explicit with clear instructions reduces the space Claude needs to explore the codebase and think through solutions, ultimately using less context than vague prompts.
How can you preserve information across different sessions without using context?	Store important information in the Claude.md file so Claude can reference it without having to rediscover it from scratch.
What is the advantage of using subagents for certain tasks?	Subagents run in a separate context window and can return just a summary to the main agent, saving context space when you only need the answer, not the process.
Why might you want to turn off MCP servers?	MCP servers load all available tools into context by default; turning off unrelated servers saves context space for your project.
How do skills differ from MCP servers in terms of context usage?	Skills work similarly to MCP servers but don't put the entire tool set into context, saving you space compared to MCP servers.