DeepExploit: No Bingos when training on Metasploitable3

13o-bbr-bbq / machine_learning_security

Source code about machine learning and security.

1.96k stars 646 forks source link

DeepExploit: No Bingos when training on Metasploitable3 #45

Open hamidb opened 4 years ago

hamidb commented 4 years ago

Hi, I was trying to train the model targeting Metasploitable 3 VM. However, after complete scan and training, not a single Bingo (award > 0) was found. I tried it on both versions of metasploitable 3 (ubuntu and win) with no success. I also tried it on another linux VM and still could not find any bingo. I was wondering if that's a known issue or there might be a solution for it. For reference, I have attached nmap output as well as the log file from DeepExploit.py. train_met3_win.log

nmap_result_192.168.57.5.txt

13o-bbr-bbq commented 4 years ago

@hamidb I think two solutions.

increase training number.
By modify the train_max_num in config.ini , you can increase the training number.
modify the exploit's module options.
Usually, the DeepExploit uses default options of exploit's module.
By you modify the exploit's module option, exploitation accuracy of DeepExploit is improved.
Please see the Tips: 1. How to change "Exploit module's option".