The agent initially does not have any private keys. Keys are added using ssh-add1. Multiple identities may be stored in ssh-agent concurrently and ssh(1) will automatically use them if present. ssh-add1 is also used to remove keys from ssh-agent and to query the keys that are held in one.
The options are as follows:
If a commandline is given, this is executed as a subprocess of the agent. When the command dies, so does the agent.
The idea is that the agent is run in the user's local PC, laptop, or terminal. Authentication data need not be stored on any other machine, and authentication passphrases never go over the network. However, the connection to the agent is forwarded over SSH remote logins, and the user can thus use the privileges given by the identities anywhere in the network in a secure way.
There are two main ways to get an agent set up: The first is that the agent starts a new subcommand into which some environment variables are exported, eg ssh-agent xterm & The second is that the agent prints the needed shell commands (either sh(1) or csh(1) syntax can be generated) which can be evaluated in the calling shell, eg eval `ssh-agent -s` for Bourne-type shells such as sh(1) or ksh(1) and eval `ssh-agent -c` for csh(1) and derivatives.
Later ssh(1) looks at these variables and uses them to establish a connection to the agent.
The agent will never send a private key over its request channel. Instead, operations that require a private key will be performed by the agent, and the result will be returned to the requester. This way, private keys are not exposed to clients using the agent.
A UNIX socket is created and the name of this socket is stored in the SSH_AUTH_SOCK environment variable. The socket is made accessible only to the current user. This method is easily abused by root or another instance of the same user.
The SSH_AGENT_PID environment variable holds the agent's process ID.
The agent exits automatically when the command given on the command line terminates.
In Debian, ssh-agent is installed with the set-group-id bit set, to prevent ptrace(2) attacks retrieving private key material. This has the side-effect of causing the run-time linker to remove certain environment variables which might have security implications for set-id programs, including LD_PRELOAD LD_LIBRARY_PATH and TMPDIR If you need to set any of these environment variables, you will need to do so in the program executed by ssh-agent.