Cursor File and Terminal Shortcuts

Saving a file (Cmd/Ctrl+S), creating a new file (Cmd/Ctrl+N), and toggling the integrated terminal (Cmd/Ctrl+`) all function exactly as they do in VS Code, since this base editor and workspace functionality is inherited wholesale from the underlying fork rather than being an area where Cursor has made changes. The integrated terminal specifically is worth noting as a point of interaction with Cursor's AI features beyond just being a standard shell — Agent mode in particular can execute terminal commands as part of a multi-step task, running a test suite or a build script and reading the output to inform its next move, though the terminal itself remains a standard, unmodified terminal emulator rather than having its own distinct AI-specific shortcut layer. For teams and individuals evaluating a move from plain VS Code to Cursor specifically for its AI capabilities, understanding that file management, terminal usage, and general editor navigation require zero adjustment is a meaningful part of the value proposition — the switching cost is concentrated entirely in learning the new AI-specific shortcuts (Cmd/Ctrl+K, Cmd/Ctrl+L, Cmd/Ctrl+I for Agent mode, and the Tab-based multi-suggestion flow) covered in the AI Editing and AI Chat categories, rather than needing to relearn foundational editor usage from scratch. One detail worth flagging for anyone managing a larger project: because Agent mode can create, modify, and save multiple files as part of a single autonomous task, it's worth keeping version control commits granular during an active Cursor session, so that if an autonomous multi-file change needs to be partially reverted, the git history stays clean enough to do that precisely rather than needing to hand-pick individual lines back out of one large sprawling commit. A point worth flagging specifically for Cursor users rather than plain VS Code users: because Agent mode can create new files as part of an autonomous multi-step task, it is worth periodically checking the file explorer sidebar during and after a longer Agent session, since new files can appear without the same explicit New File action a human would normally take, and it is easy to lose track of exactly what changed if you were not watching closely during a long autonomous run. Reviewing the full diff before committing remains the most reliable safety net regardless of how the files were created. For solo developers working without a formal code review step, it's worth treating frequent, small commits as a lightweight substitute for that review — committing right before starting a new Cursor-assisted change, and again right after, creates natural checkpoints that make it trivial to see exactly what the AI touched in isolation, rather than trying to reconstruct it later from a single sprawling end-of-day commit. Because the terminal panel is unmodified from VS Code, any existing terminal-based workflow — a custom shell profile, aliases, a particular multiplexer setup — carries over without adjustment, and Agent mode's ability to run commands in that same terminal means custom aliases and scripts a developer already relies on are available to the AI as well, for better or worse, which is worth keeping in mind when reviewing what an autonomous Agent run actually executed. Cursor's settings sync feature, when enabled, carries a user's keybinding customizations, extensions, and preferences across multiple machines the same way VS Code's own Settings Sync does, meaning a developer moving between a desktop and a laptop doesn't need to manually reconfigure file-management or terminal preferences twice — a small convenience but one that meaningfully reduces the friction of maintaining a consistent setup across more than one machine. Finally, for anyone maintaining a `.gitignore` or `.cursorignore` file, it's worth knowing that Cursor respects an ignore file specifically for its own AI indexing, letting you exclude directories like build output, vendored dependencies, or generated files from what the AI considers when answering questions or proposing changes — keeping that ignore list current both speeds up indexing on a large repository and improves the relevance of the AI's context by keeping irrelevant generated code out of its retrieval entirely. This is a low-effort, high-value housekeeping task worth revisiting whenever a project adds a new build tool or generated-output directory, since a stale ignore list quietly degrades both indexing speed and answer quality over time without any obvious error message pointing back to the cause. Teams standardizing on Cursor across a codebase sometimes go a step further and commit a shared `.cursorignore` alongside the repository's existing `.gitignore`, ensuring every contributor benefits from the same indexing scope without each person needing to configure it individually on their own machine.