Refactor

Component(s) to be refactored

• HadoopExploiter

Explanation

The time has come! It's time to start migrating exploiters from being integrated with the agent monolith to plugins that are loaded on-demand.

Tasks

• Generate a config schema for the hadoop exploiter (0d) @ilija-lazoroski @shreyamalviya
  • Read through the existing exploiter and identify any timeouts, ports, or other options that should be configurable.
• Generate a plugin manifest for the hadoop exploiter (0d) @shreyamalviya
• Rewrite the hadoop exploiter as a plugin (0d) @mssalvatore
  • Put this in monkey/agent_plugins/exploiters/hadoop
  • If ports in TargetHost are open, run exploiter. If closed, return. If unknown, run exploiter. Try all configured ports
  • Write a runner to speed up tests/debugging
  • Unit tests! - @mssalvatore
  • Don't inherit from WebRCE or HostExploiter
  • Don't forget to add MITRE attack tags
  • Tag the ExploitationEvent with relevant CVEs (AFAIK there are no CVEs, that's just how Hadoop is by default :/).
  • Write a build script that builds the plugin as a tar archive (Do we need one for each supported OS?) @ilija-lazoroski
  • Test ETE @ilija-lazoroski
  • Add more logging @ilija-lazoroski
  • Fix agent download race condition @mssalvatore
• Remove hardcoded Hadoop (0d) @ilija-lazoroski
• Create a wrapper to wrap plugins and protect them from trying to access the same handle simultaneously (0d) @mssalvatore
• Make TCPPortSelector multiprocessing safe (0d) - @ilija-lazoroski @cakekoa
• Pass TCPPortSelector to exploiter plugins at construction (0d) - @mssalvatore