Local safety layer for AI coding agents: block dangerous commands, mask secrets, filter env vars, and keep JSONL audit logs.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).