Claude Code is Anthropic's coding agent for the terminal, IDEs, desktop app, and web. Its controls include allow, ask, and deny permission rules, permission modes, an OS-level sandbox for shell commands, hooks that can block tool calls, and centrally enforced settings. Know these and you can set a safe baseline for a team.
Read in this order
- OfficialSecurity (Claude Code Docs)Start here: the security model and what each safeguard does and does not cover.
- OfficialConfigure permissions (Claude Code Docs)How to write allow, ask and deny rules for tools, files and commands.
- OfficialConfigure the sandboxed Bash tool (Claude Code Docs)How the sandboxed Bash tool restricts files and network, and how to configure it.