Tag
Discussion on the need for local safety boundaries in AI coding agents to prevent unauthorized file access or command execution.