RandomSelectRedAgent

Red Agent that random selects an action (with parameters) to execute.

Attributes:

Name	Type	Description
`step`	`int`	number of steps the agent has taken in the environment
`last_action`	`str or Action`	the previous action that was executed
`print_output`	`bool`	option as to if action info is printed to the terminal
`disable_withdraw`	`bool`	when true, the Withdraw action is removed from the list of possible actions, so that the agent will no longer perform Withdraw.

Other Parameters:

Name	Type	Description
`name`	`str`	name of the agent
`np_random`	`numpy.random._generator.Generator`	numpy random generator initialised on creation of scenario. This allows for the seed to be consistent with the CybORG() seed parameter

Functions

get_action(observation, action_space)

Chooses a valid action randomly from the action space, along with corresponding parameters - picked randomly when given options.

Parameters:

Name	Type	Description	Default
`observation`	`dict`	agent observation from last action	required
`action_space`	`dict`	agent action space	required

Raises:

Type	Description
`ValueError`	There are no valid actions for the agent to take. Sleep should always be a valid action, so will only occur in error.

last_turn_summary(observation: dict)

Prints action name, parameters and success