Add a new rule called “Shell Commands”:

A ‘’‘Shell Command’‘’ is a type of Virtual Action that the Mainframe performs as a Client on behalf of an Agent. For the purposes of resolving Shell Commands, a Client is considered an Agent; The Client serves as a proxy for the Agent who initiated the command.

An Agent must be connected to a Client to perform Shell Commands as that Client. ‘‘To perform a Shell Command, you must always specify exactly one Client that you are connected to, or the action fails.’‘

The definition of a Shell Command should be contained within the “Shell command” flair; If one is not, the Mainframe may edit the ruleset to make it so.

{{Flair top|Shell command}}
As a Shell Command, a Client can ‘’‘ping’‘’ the Mainframe. The Mainframe will respond to confirm that it is online.
{{Flair bottom}}

Add a subrule to “Shell Commands” called “Logging”:

The Mainframe shall maintain a ‘’‘log’‘’ of successful Shell Commands. Each new log entry shall include:
* the time/date it occurred
* the IP of the initiating Agent
* the name of the Client
* the type of command
* any other data the command specifies to be logged

{{Flair top|Shell command}}
As a shell command, a Client may ‘’‘request a copy of logs’‘’ within a particular scope (e.g. a certain timeframe, logs associated with a certain IP, etc). Each log entry will be accompanied by an MD5 hash of the entry by which it can be identified.

If the Client does not have access to all of the logs in the desired scope, the action will fail. By default, a Client only has access to the logs of actions performed by that particular Client.
{{Flair bottom}}