Harvis is designed to automate your C2 Infrastructure.

Related tags

Text Data & NLPHarvis
Overview

Harvis

Harvis is designed to automate your C2 Infrastructure, currently using Mythic C2.

📌 What is it?

Harvis is a python tool to help you create multiple hauls for a specific operation during a red team engagement. It can automatically create your C2 machine, redirector machine, setup SSL, .htaccess configuration, firewall rules and more. Harvis also has the purpose of automating the job of the operator of identifying burned domains/redirectors that may be caught during the operation. After identifying the burned domains it provides the possibility of rotating the infrastructure, setting up different redirectors and atributing a different domain.

📌 How?

🔨 Harvis uses the Digital Ocean API to automate creation and deletion of droplets.

🔨 The Namecheap API is used to set DNS records to redirectors.

🔨 The API Void is used to constantly verify the state of the redirectors and check if anything is blacklisted.

Features

Harvis has several features to help you organize your available domains and redirector/C2 machines.

  • Namecheap Interaction - Harvis enables you to buy domains directly through the command line, as long as you have credits in your account.
  • Multiple Hauls - It is possible to create as many hauls as possible, each one having specific configurations.
  • Multiple C2 Profiles by Haul - Harvis allows you to create multiple C2 profiles by hauls: you could two HTTP listeners in one of the C2's and 3 in the other one, each of them listening on different ports.
  • Multiple Redirector Configuration - Each redirector can have a different .htaccess configuration, defined by the operator.
  • Customizable Firewall Rules - Harvis comes with default firewall rules for the redirectors and C2's, however, this feature is easily customizable.
  • Priority System - Each haul has a queue system, in a way that the new redirector will replace the blacklisted one with the domains in the queue.
  • Priority System - Automatically replace your droplets: Harvis identifies any blacklisted redirector and print the results to the operator. It allows the operator to create a temporary droplet to replace the blacklisted one. It does not configure the migration of any active agents, since the way the agent will be migrated/spawned to connect to the new domain might be a very personal decision in an engagement. After creating the temporary droplet, it allows you to migrate any active agents and kill the older redirector.
  • Priority System - If somehow the script crashes, all the information will be saved in the backup.py file. Restarting the script will recover all your infrastructure as it was.

Installation

git clone https://github.com/thiagomayllart/Harvis/
cd Harvis
pip3 install -r requirements

Running

python3 harvis.py

Configuring API Keys

Harvis can only be used with the proper API Keys from Digital Ocean, Namecheap, APIVoid.

These api keys should be added to the config.py file in the respective lines. Don't forget to setup your namecheap username in namecheap_username variable:

digital_ocean_token = ""
...
namecheap_key = ""
...
apivoid_key = ""
...
namecheap_username = ""
...

For more information regarding these API Keys, visit:

https://app.apivoid.com/ https://ap.www.namecheap.com/ https://cloud.digitalocean.com/

First Run

Before running the script, you should apply some modifications to the config file, which describes the configuration of your infrastructure:

  1. Modify the "names" variable.

This variable holds the names of each haul you want your infrastructure you have. Theses names should be applied in the next variables. You can have as many hauls you want. You can also specify which domain should already be configure to each haul. If you don't specify the domains, you will be asked to move domains to each haul in the first run. If you don't have any available domains in your namecheap API, you can buy it directly though Harvis. It is also possible to have more than one domain in each haul: the first one will be used in the redirector and the others will already be in the backup list for further infrastructure rotations. Example:

names = {"short":[],"long":[],"exploitation":[],"testing":[]}
names = {"short":["domain1.com"],"long":["domain2.com","domain3.com"],"exploitation":["domain4.com"],"testing":["domain5.com"]}
  1. Modify the "config_htaccess_dic" variable:

You can customize your htaccess rules for each haul in this variable. This variable is dictionary, so remember to add an htacces for each haul you added previously in the "names" variable like:

config_htaccess_dic = \
    {"short":"""
RewriteEngine On
RewriteCond %{REQUEST_URI} ^/({1})/?$ [NC]
RewriteRule ^.*$ https://{2}%{REQUEST_URI} [P]
RewriteRule ^.*$ http://{3}? [L,R=302]
""","long:"""
RewriteEngine On
RewriteCond %{REQUEST_URI} ^/({1})/?$ [NC]
RewriteRule ^.*$ https://{2}%{REQUEST_URI} [P]
RewriteRule ^.*$ http://{3}? [L,R=302]
""","exploitation":"""
RewriteEngine On
RewriteCond %{REQUEST_URI} ^/({1})/?$ [NC]
RewriteRule ^.*$ https://{2}%{REQUEST_URI} [P]
RewriteRule ^.*$ http://{3}? [L,R=302]
""","testing":"""
RewriteEngine On
RewriteCond %{REQUEST_URI} ^/({1})/?$ [NC]
RewriteRule ^.*$ https://{2}%{REQUEST_URI} [P]
RewriteRule ^.*$ http://{3}? [L,R=302]
"""
                       }

You can notice the presence of the fields: {1}, {2}, {3}. If you customize this variable, do not remove them. {1} are the parameters your Mythic agent will use to communicate to the C2 (it also allows customization). {2} is the IP address of your C2. {3} is the location the redirector will be redirecting (it also allows customization).

  1. Modify the "agent_profiles" variable:

This variable describes the HTTP parameters that your agent will use to communicate to your C2. Mythic allows setting these parameters during the creation of the agent, so, these values should match the ones you will be configurating the agent. The first one is the GET parameter and other one is the POST parameter. Also, remeber to once again add a configuration to each haul you created previously:

agent_profiles = {"short":{"URI":"data|index"},"long":{"URI":"q|id"}... ...
  1. Modify "domain_front_redirector" variable:

This variable holds the domain your redirectors will be redirecting anyone that tries to access. Add a configuration to each haul you created previously:

domain_front_redirector = {"short":"www.example.com","long":"www.example2.com"... ... ...

  1. Modify "c2_profiles" variable:

With the "c2_profiles" variable you can create different listener profiles for each Haul you created. The format is exactly the same as the JSON you may find when accessing Configuring a C2 Profile in Mythic. You can also have other profiles than HTTP, however, depending on the protocol used, it may be necessary to change firewall rules in the C2 or the redirector (further explained).

Example:


c2_profiles = {"short":[{"name":"HTTP","config":"""{
  "instances": [
  {
    "ServerHeaders": {
      "Server": "NetDNA-cache/2.2",
      "Cache-Control": "max-age=0, no-cache",
      "Pragma": "no-cache",
      "Connection": "keep-alive",
      "Content-Type": "application/javascript; charset=utf-8"
    },
    "port": 443,
    "key_path": "privkey.pem",
    "cert_path": "cert.pem",
    "debug": true
    }
  ]
}"""}],"long":[{"name":"HTTP","config":"""{
  "instances": [
  {
    "ServerHeaders": {
      "Server": "NetDNA-cache/2.2",
      "Cache-Control": "max-age=0, no-cache",
      "Pragma": "no-cache",
      "Connection": "keep-alive",
      "Content-Type": "application/javascript; charset=utf-8"
    },
    "port": 443,
    "key_path": "privkey.pem",
    "cert_path": "cert.pem",
    "debug": true
    }
  ]
}"""}] }
  1. Modify the "check_infra_state" variable:

This variable holds the value (in seconds) that will be used as the interval between each verification of blacklisted domains by APIVoid.

  1. Modify "ip_allowed_to_connect_c2":

Replace it with the IP you will be using as the proxy to connect to your Mythic C2 panel. You can use your public IP, but it is not recommended.

  1. Replace "username":

Replace it with the username you will be using during the engagement. This variable is used to tag each droplet created in digital ocean, making the distinction between the droplets of each operator easier. It also avoids that the tool erases the droplet of another user (in case you guys are using the same API Keys).

  1. Modify Firewall Rules (OPTIONAL):

If you want to modify the firewall rules for the C2/redirector, you will find them respectively at:

C2: C2_setup.py: function firewall_rules Redirector: redirect_setup.py: firewall_rules

Important

Harvis whitelists the IP of the machine you are deploying it as being able to access your Mythic C2 panel. It is highly recommended to use a VPC to deploy Harvis.

Owner
Thiago Mayllart
- Information Security Researcher / RedTeam
Thiago Mayllart
[WWW 2021 GLB] New Benchmarks for Learning on Non-Homophilous Graphs

New Benchmarks for Learning on Non-Homophilous Graphs Here are the codes and datasets accompanying the paper: New Benchmarks for Learning on Non-Homop

94 Dec 21, 2022
CMeEE 数据集医学实体抽取

医学实体抽取_GlobalPointer_torch 介绍 思想来自于苏神 GlobalPointer,原始版本是基于keras实现的,模型结构实现参考现有 pytorch 复现代码【感谢!】,基于torch百分百复现苏神原始效果。 数据集 中文医学命名实体数据集 点这里申请,很简单,共包含九类医学

85 Dec 28, 2022
Making text a first-class citizen in TensorFlow.

TensorFlow Text - Text processing in Tensorflow IMPORTANT: When installing TF Text with pip install, please note the version of TensorFlow you are run

1k Dec 26, 2022
This repository contains the codes for LipGAN. LipGAN was published as a part of the paper titled "Towards Automatic Face-to-Face Translation".

LipGAN Generate realistic talking faces for any human speech and face identity. [Paper] | [Project Page] | [Demonstration Video] Important Update: A n

Rudrabha Mukhopadhyay 438 Dec 31, 2022
A simple version of DeTR

DeTR-Lite A simple version of DeTR Before you enjoy this DeTR-Lite The purpose of this project is to allow you to learn the basic knowledge of DeTR. P

Jianhua Yang 11 Jun 13, 2022
Global Rhythm Style Transfer Without Text Transcriptions

Global Prosody Style Transfer Without Text Transcriptions This repository provides a PyTorch implementation of AutoPST, which enables unsupervised glo

Kaizhi Qian 193 Dec 30, 2022
Code release for NeX: Real-time View Synthesis with Neural Basis Expansion

NeX: Real-time View Synthesis with Neural Basis Expansion Project Page | Video | Paper | COLAB | Shiny Dataset We present NeX, a new approach to novel

537 Jan 05, 2023
Finetune gpt-2 in google colab

gpt-2-colab finetune gpt-2 in google colab sample result (117M) from retraining on A Tale of Two Cities by Charles Di

212 Jan 02, 2023
News-Articles-and-Essays - NLP (Topic Modeling and Clustering)

NLP T5 Project proposal Topic Modeling and Clustering of News-Articles-and-Essays Students: Nasser Alshehri Abdullah Bushnag Abdulrhman Alqurashi OVER

2 Jan 18, 2022
Code and data accompanying Natural Language Processing with PyTorch

Natural Language Processing with PyTorch Build Intelligent Language Applications Using Deep Learning By Delip Rao and Brian McMahan Welcome. This is a

Joostware 1.8k Jan 01, 2023
An open collection of annotated voices in Japanese language

声庭 (Koniwa): オープンな日本語音声とアノテーションのコレクション Koniwa (声庭): An open collection of annotated voices in Japanese language 概要 Koniwa(声庭)は利用・修正・再配布が自由でオープンな音声とアノテ

Koniwa project 32 Dec 14, 2022
Question and answer retrieval in Turkish with BERT

trfaq Google supported this work by providing Google Cloud credit. Thank you Google for supporting the open source! 🎉 What is this? At this repo, I'm

M. Yusuf Sarıgöz 13 Oct 10, 2022
Perform sentiment analysis on textual data that people generally post on websites like social networks and movie review sites.

Sentiment Analyzer The goal of this project is to perform sentiment analysis on textual data that people generally post on websites like social networ

Madhusudan.C.S 53 Mar 01, 2022
FedNLP: A Benchmarking Framework for Federated Learning in Natural Language Processing

FedNLP is a research-oriented benchmarking framework for advancing federated learning (FL) in natural language processing (NLP). It uses FedML repository as the git submodule. In other words, FedNLP

FedML-AI 216 Nov 27, 2022
SHAS: Approaching optimal Segmentation for End-to-End Speech Translation

SHAS: Approaching optimal Segmentation for End-to-End Speech Translation In this repo you can find the code of the Supervised Hybrid Audio Segmentatio

Machine Translation @ UPC 21 Dec 20, 2022
Code for ACL 2021 main conference paper "Conversations are not Flat: Modeling the Intrinsic Information Flow between Dialogue Utterances".

Conversations are not Flat: Modeling the Intrinsic Information Flow between Dialogue Utterances This repository contains the code and pre-trained mode

ICTNLP 90 Dec 27, 2022
Bidirectional LSTM-CRF and ELMo for Named-Entity Recognition, Part-of-Speech Tagging and so on.

anaGo anaGo is a Python library for sequence labeling(NER, PoS Tagging,...), implemented in Keras. anaGo can solve sequence labeling tasks such as nam

Hiroki Nakayama 1.5k Dec 05, 2022
Sequence-to-Sequence Framework in PyTorch

nmtpytorch allows training of various end-to-end neural architectures including but not limited to neural machine translation, image captioning and au

LIUM 395 Nov 21, 2022
Train BPE with fastBPE, and load to Huggingface Tokenizer.

BPEer Train BPE with fastBPE, and load to Huggingface Tokenizer. Description The BPETrainer of Huggingface consumes a lot of memory when I am training

Lizhuo 1 Dec 23, 2021
Chinese named entity recognization (bert/roberta/macbert/bert_wwm with Keras)

Chinese named entity recognization (bert/roberta/macbert/bert_wwm with Keras)

2 Jul 05, 2022