关于
Command Guard acts as a proactive safety layer for Claude Code, ensuring that every terminal command is vetted against your local .agentpolicy and .agentignore files before execution. It is designed to protect developers from accidental data loss, unintended deletions, or risky system modifications by automatically blocking unauthorized commands and requiring explicit user confirmation for sensitive operations. By maintaining a detailed Action Log of every decision, it provides both a safety net for automated workflows and an audit trail for security-conscious environments.
