Corv: An SSH client for AI agents (and humans) (github.com)

Corv has introduced a specialized SSH client designed for both AI agents and human users, streamlining the connection and command execution processes. Unlike traditional SSH, which requires constant re-authentication and exposes sensitive credentials, Corv allows agents to connect by name, execute commands without displaying passwords, and return structured JSON outputs. This design enhances security and efficiency, enabling long-running tasks to be detached and resumed seamlessly—beneficial for AI-driven infrastructures in need of reliable command execution. Significantly, Corv aims to optimize the experience for AI coding agents like Claude and Codex by providing an agent-friendly interface that minimizes the common pitfalls of traditional SSH, such as erroneous command parsing and output clutter. The client supports interactive connection management, encrypted storage of credentials, and an intuitive terminal UI, all while ensuring that critical secrets remain protected. Its innovative approach facilitates efficient command execution, robust auditing capabilities, and maintains compatibility across platforms (Linux, macOS, and Windows), establishing Corv as a valuable tool for the AI/ML community seeking to enhance operational workflows.