ssh-agent

Start ssh-agent and set environment variables

$ eval $(ssh-agent)

$ ssh-agent -s

$ ssh-agent -c

$ ssh-agent -k

$ ssh-agent [command]

$ ssh-agent -t [1h]

ssh-agent [-c | -s] [-Dd] [-a SOCKET] [-t LIFE] [COMMAND [ARGS...]]

ssh-agent is a program that holds private keys used for SSH public key authentication. The agent runs in the background and provides keys to SSH clients without requiring passphrase re-entry.
Keys are added to the agent using ssh-add. Once added, the agent responds to authentication requests from SSH connections, including forwarded agent connections from remote hosts.
The agent outputs shell commands to set environment variables (SSHAUTHSOCK and SSHAGENTPID). These must be evaluated in the shell using eval $(ssh-agent) to enable communication between SSH clients and the agent.

-s

Generate Bourne shell commands on stdout.

Generate C-shell commands on stdout.

Kill the current agent (using SSHAGENTPID).

Debug mode (don't fork).

Foreground mode without debug output.

Bind to specified Unix socket instead of random.

Default maximum lifetime for added keys (e.g., 1h, 30m).

The agent keeps decrypted private keys in memory. If compromised, all loaded keys are exposed. Use -t to limit key lifetime. Agent forwarding (-A in ssh) should be used carefully as remote hosts can use your agent. Kill the agent when done with sensitive sessions.

ssh-agent was developed as part of OpenSSH, originally created by Tatu Ylönen and later maintained by the OpenBSD project. The agent model allows secure key storage while enabling passwordless SSH connections, and has become the standard approach for managing SSH keys across all major platforms.

ssh-add(1), ssh(1), ssh-keygen(1)